JP2006040181A - Data analysis apparatus, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data analysis apparatus that can extract factors in a predetermined output attribute in simple form and rank a plurality of factors by importance according to a definite index. <P>SOLUTION: The data analysis apparatus comprises: a data classification part 4 for classifying a basic data group as conforming and nonconforming according to the value of an output attribute; an analysis data group extraction part 6 for extracting an analysis data group from the classified basic data group; a cumulative frequency difference computation part 10/input attribute condition decision part 11 for deciding, for each input attribute of the analysis data group, an input attribute condition satisfying first correlation rules that "the input attribute, if satisfying the input attribute condition, indicates data belonging to a nonconforming unit in the analysis data group and, if not, indicates data belonging to a conforming unit in the analysis data group"; and a rank decision part 13 for ranking the plurality of decided input attribute conditions by importance as a factor in an output attribute condition corresponding to a nonconforming unit according to an index that is at least either the number of nonconforming units or the degree of separation of nonconforming units. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an output attribute (object attribute) to be analyzed, such as characteristics of a product manufactured in a manufacturing process, and an input attribute (description attribute) that is an attribute affecting the output attribute, such as a manufacturing process condition. The present invention relates to a data analysis apparatus, a data analysis method, and a data analysis program for analyzing the causal relationship of the data.

  A decision tree technique is known as an effective technique for analyzing the causal relationship between output attributes and input attributes. This method creates a tree structure in which the values of the output attributes are well organized at the leaf portions that are sequentially separated by the values of the input attributes.

  FIG. 14 is an example of a decision tree created using the data group of Table 1 as an analysis target, and the Gini index and the degree of improvement described in p44 to p47 of Non-Patent Document 1 are used as evaluation indexes for branch conditions in the tree structure. ing.

  The data group of Table 1 is a sample prepared with reference to Table 1 of Patent Document 1, and the values of the four input attributes x1, x2, x3, and x4 and the value of the output attribute y for these input attributes. This is a set of 12 data sets. In the decision tree of FIG. 14, the values XZ and Y of the output attribute y are well separated by the values of the input attributes x2, x1, and x3.

  Next, a procedure of creating the decision tree of FIG. 14 (separating y = Y and y = XZ) using the Gini index and improvement degree described in p44 to p47 of Non-Patent Document 1 will be described.

  The Gini index is an index that represents the degree of grouping of the values of the output attribute y, and is represented by the following expression (1).

i (t) = 1−Σ {p (j | t)} ² (1)
Here, p (j | t) indicates the probability that the output attribute y is y = j at the node t, and that the Gini index i (t) is small indicates that the value of the output attribute y at the node t is It means that it is well organized.

  The degree of improvement is an index indicating “how much the output attribute is improved by branching a certain node t1 into two child nodes t2 and t3 according to a branching condition”. 2).

Δi (t1) = i (t1) − {pt2 · i (t2) + pt3 · i (t3)}
(2)
Here, i (t1), i (t2), and i (t3) indicate Gini indexes at the node t1, the child node t2, and the child node t3, respectively. Further, pt2 and pt3 indicate branching ratios when the node t1 is branched into the child node t2 and the child node t3.

  In the data group of Table 1, the Gini index i (t1) at the root node t1 is 0.444 from the equation (1). The degree of improvement when this root node t1 is branched into two child nodes according to various branch conditions is shown in FIGS. Here, the number of branch condition patterns for calculating the degree of improvement (calculation of equation (2)) is a total of 12 conditions (each input) for all the branch condition patterns that the input attributes x1, x2, x3, and x4 can take. 3 conditions for attributes). The condition having the maximum improvement in each input attribute is the optimum branching condition for separating y = Y and y = XZ in that input attribute. 15 to 18, “branch between x1 = A, B and x1 = C, D” for the input attribute x1, and “x2 = a, b, x2 = c, for the input attribute x2”. “branch with d” is “branch between x3 = 1, 2 and x3 = 3, 4” for the input attribute x3, and “branch with d4” is “x4 = 10, x4 = 20, 30, It can be seen that the “branch with 40” is appropriate.

Then, among the above-mentioned optimum branch conditions for each input attribute, a condition having the maximum improvement degree, that is, “a branch between x2 = a, b and x2 = c, d” is determined from the root node t1. Is selected as the final branch condition. Thereafter, the same operation as described above is repeated, and finally the decision tree of FIG. 14 is created. In the process of creating the decision tree of FIG. 14, the degree of improvement is maximized when node t5 is branched into “x3 = 1, 2 and x3 = 3, 4”. Degree = 0.5.
JP-A-8-314725 (Publication date: November 29, 1996) Atsushi Otaki, Yuji Horie, Dan Steinberg, "Applied binary tree analysis method-by CART-", Nikka Giren, July 6, 1998, P44-P47

  Applying the conventional decision tree (Fig. 14) based on the Gini index method (Gini index and improvement degree) to factor analysis of product characteristic defects in the manufacturing process of products such as devices, the problem of the conventional technology explain.

  Now, the input attributes x1, x2, x3, and x4 in Table 1 are various process data and in-line inspection data in the product manufacturing process, and the output attribute y is the product characteristic data. The output attribute y = Y is the product characteristic. It shall correspond to a defect. Then, the process engineer uses the conventional decision tree (FIG. 14) for the product characteristic defect y = Y, and causes the product characteristic defect (“product characteristics are bad because which input attribute is in which value range”). ")") Shall be investigated.

In the case of the above material, the conventional decision tree (FIG. 14) has the following problems.
That is, the condition of product characteristic failure where y = Y is
・ "X2 = c, d"
Or
“X2 = a, b” and “x1 = C, D” and “x3 = 3,4”
It appears in a plurality of places (two places in this example) in the form of a hierarchical structure. For this reason, specifically, it is difficult for the process engineer to determine the cause of the product characteristic failure, such as “Which range of each input attribute is the product characteristic bad?”. For example, for the input attribute x2, it is not known which of “x2 = c, d” and “x2 = a, b” is the product characteristic defect condition of y = Y. In this example, since y = Y appears only in two places in the decision tree, the process engineer can somehow determine the cause of the product characteristic failure. However, in manufacturing sites of products such as actual devices (especially semiconductor devices), there are process data and in-line inspection data of about 10 to 100 attributes per process, and the values are multivalued and in a very wide range. Distributed. In such a case, a product characteristic defect with y = Y appears throughout the decision tree, and branching under different conditions in the same input attribute has occurred many times, and the process engineer is no longer As a result, it becomes impossible to properly identify the cause of the failure.

  Furthermore, in the conventional decision tree (FIG. 14), only the accuracy of carving at each node and the goodness of grouping after carving are evaluated based on the Gini index and the degree of improvement, so y = Y that appears in various hierarchies. It is difficult to prioritize multiple causes of product characteristic defects. Specifically, four conditions “x2 = c, d”, “x2 = a, b”, “x1 = C, D”, “x3 = 3, 4” that cause a product characteristic defect y = Y. It is impossible to determine from which of these conditions the countermeasures should be taken in order. If importance is attached to the degree of improvement, “x3 = 3, 4” where the degree of improvement is the maximum (0.5) is ranked higher, which is “x2 = a, b” and “x1 = C, D”. It is for the population of "and has little meaning.

  That is, in the conventional decision tree, the first problem is that it is difficult for the process engineer to determine the cause of the product characteristic failure, such as “Which range of each input attribute is the product characteristic bad?”, And The second problem is that it is difficult to understand what priority should be taken for each condition (defect factor) appearing on various levels.

  The present invention has been made in view of the above-described conventional problems, and a first object of the present invention is to make it easier for a user to determine an input attribute condition that is a factor of a predetermined output attribute (a factor of product characteristic failure). To provide a data analysis apparatus, a data analysis method, a data analysis program, and a recording medium that can be extracted in a simple form. In addition, the second purpose is that for a plurality of input attribute conditions extracted as a factor of a predetermined output attribute (factor of product characteristic failure), a factor of the predetermined output attribute (factor of product characteristic failure) based on a clear index A data analysis apparatus, a data analysis method, a data analysis program, and a recording medium that can be classified according to the importance as the input attribute conditions, thereby clarifying which input attribute condition should be taken in order It is to provide.

  In order to solve the above-described problem, the data analysis apparatus of the present invention provides a basic data group that is a set of data composed of a plurality of input attributes and output attributes. A data analysis apparatus for analyzing causal relationships and extracting information indicating the causal relationships, wherein the basic data group is classified into a first data group and a second data group according to the value of the output attribute. Each of each of the analysis data group, a classification means for assigning a classification flag, an analysis data group extraction means for extracting an analysis data group to be analyzed from the basic data group after classification by the classification means, and For each of the input attribute conditions related to the input attribute, “the first data group and the second data group are grouped by dividing the analysis data group by the input attribute condition and its exclusive condition, respectively. If the input attribute satisfies the input attribute condition, the data belongs to the second data group in the analysis data group, and if the input attribute does not satisfy the input attribute condition, the second data group in the analysis data group. The first evaluation means for calculating the input attribute condition evaluation index, which represents the probability of the first association rule that the data belongs to one data group, and the input attribute of each of the analysis data group, An input attribute condition determining means for determining an input attribute condition having the largest input attribute condition evaluation index as an input attribute condition satisfying the first correlation rule; and the plurality of inputs determined by the input attribute condition determining means A second value representing the number of data satisfying the input attribute condition in the basic data group and corresponding to the second data group for at least part of the attribute condition Index of at least one of the number of corresponding data groups and the second data group separation degree representing the ratio of the number of data corresponding to the second data group in the data satisfying the input attribute condition in the basic data group And ranking determining means for ranking by importance as a factor of the output attribute condition corresponding to the second data group.

  In order to solve the above problem, the data analysis method of the present invention uses the data analysis apparatus of the present invention to perform a basic data group that is a set of data composed of a plurality of input attributes and output attributes. A data analysis method for analyzing a causal relationship between the input attribute and the output attribute and extracting information indicating the causal relationship, wherein the basic data group is determined by the classification means according to the value of the output attribute. Then, the data is classified into a first data group and a second data group, a classification flag is assigned, and the analysis data group extraction means performs analysis of the basic data group after classification by the classification means. For each of the input attribute conditions related to each input attribute of the analysis data group by the analysis data group extraction step for extracting the target analysis data group and the first evaluation unit, “the input attribute The first data group and the second data group are grouped by dividing the analysis data group according to a condition and an exclusive condition, and if the input attribute satisfies the input attribute condition, the analysis data group in the analysis data group Probability of the first correlation rule that is data belonging to the second data group and is data belonging to the first data group in the analysis data group if the input attribute does not satisfy the input attribute condition. A first evaluation step of calculating an input attribute condition evaluation index, and an input having the maximum input attribute condition evaluation index for each of the input attributes of the analysis data group by the input attribute condition determination means An input attribute condition determining step for determining an attribute condition as an input attribute condition satisfying the first correlation rule, and the rank determining means, The number of data satisfying the input attribute condition in the basic data group and corresponding to the second data group for at least some of the plurality of input attribute conditions determined by the input attribute condition determining means And at least a second data group separation degree representing a ratio of the number of data corresponding to the second data group in data satisfying the input attribute condition in the basic data group And a rank determining step for ranking according to the importance as a factor of the output attribute condition corresponding to the second data group using one as an index.

  In order to solve the above problem, the data analysis program of the present invention provides a basic data group, which is a set of data composed of a plurality of input attributes and output attributes, for the input attributes and the output attributes. A data analysis program for analyzing a causal relationship and extracting information indicating the causal relationship, wherein the computer includes the basic data group, the first data group and the second data group, depending on the value of the output attribute. Each of the analysis data group, the analysis data group extraction means for extracting the analysis data group to be analyzed from the basic data group after classification by the classification means, For each of the input attribute conditions related to the input attribute, “the first data group and the second data are separated by dividing the analysis data group by the input attribute condition and its exclusive condition. If the input attribute satisfies the input attribute condition, the data belongs to the second data group in the analysis data group. If the input attribute does not satisfy the input attribute condition, the analysis is performed. First evaluation means for calculating an input attribute condition evaluation index, which represents the probability of the first association rule that the data belongs to the first data group in the data group, and the input of each of the analysis data groups For each attribute, the input attribute condition having the largest input attribute condition evaluation index is determined by the input attribute condition determining unit and the input attribute condition determining unit that determine the input attribute condition satisfying the first correlation rule. For at least some of the plurality of input attribute conditions, the basic data group satisfies the input attribute condition and corresponds to the second data group. And a second data group separation representing a ratio of the number of data corresponding to the second data group in the data satisfying the input attribute condition in the basic data group It is characterized by functioning as rank determining means for ranking according to importance as a factor of the output attribute condition corresponding to the second data group using at least one of the degrees as an index.

  The recording medium of the present invention is a computer-readable recording medium on which the data analysis program of the present invention is recorded.

  According to the apparatus, the method, the program, and the recording medium of the present invention, the input attribute condition determining unit is not limited to the branch condition in the decision tree, and for all input attributes, the output attribute condition corresponding to the second data group is set. Since the input attribute condition as a factor is extracted (determined), even if a competing factor is present in the branch condition, it can be reliably captured without missing the factor.

  Then, the order determination means determines the importance as a factor of the output attribute condition corresponding to the second data group for at least a part of the plurality of input attribute conditions determined by the input attribute condition determination means. Based on at least one of the second data group hit count and the second data group separation degree, which is a clear index, ranking is performed according to importance as a factor of the output attribute condition corresponding to the second data group. As a result, the input attribute condition that causes the output attribute condition corresponding to the second data group can be recognized at a glance together with the order of importance. Therefore, a plurality of input attribute conditions that cause a predetermined output attribute can be extracted in a simple manner, and the first object of the present invention can be achieved.

  Further, in the conventional decision tree (FIG. 14), it is difficult to understand “what priority should be taken for each failure factor appearing in various layers”. On the other hand, in the apparatus, method, program, and recording medium of the present invention, at least a part of the plurality of input attribute conditions determined by the input attribute condition determining unit corresponds to the second data group. As a factor of the output attribute condition corresponding to the second data group, based on at least one of the second data group hit count and the second data group separation degree, which is a clear indicator of the importance as the factor of the output attribute condition Ranking by importance. Therefore, for a plurality of factors (input attribute conditions) that should be taken measures to eliminate the occurrence of the output attribute condition corresponding to the second data group, the priority order to take measures (from which factor to take measures in order) It is possible to clearly determine whether or not the generation of the output attribute condition corresponding to the second data group can be effectively eliminated). Therefore, the second object of the present invention can be achieved.

  The data analysis apparatus of the present invention is the data analysis apparatus having the above-described configuration, for each of the input attribute conditions determined by the input attribute condition determination means, “if the input attribute satisfies the input attribute condition, the analysis data A second evaluation means for calculating a division rule evaluation value representing the probability of the second association rule that the data is included in the second data group in the group, and the input attribute condition determination means Based on the input attribute condition having the largest division rule evaluation value among the input attribute conditions, the analysis data group is divided into a factor data group that satisfies the input attribute condition and other data that does not satisfy the input attribute condition. A dividing unit that divides into groups, and an input attribute condition determined by the input attribute condition determining unit, as a factor of an output attribute condition corresponding to the second data group. A factor determination unit that selects a high-level input attribute condition, and the analysis data group extraction unit extracts at least one of the data groups divided by the division unit as a new analysis data group, A series of processes consisting of processing by the analysis data group extraction means, processing by the first evaluation means, processing by the input attribute condition determination means, processing by the second evaluation means, and processing by the dividing means are repeated. The input attribute condition determining means is configured to determine a plurality of input attribute conditions for each of the input attributes by the repetitive processing, and the factor The determining means is, for each of the input attributes, the plurality of input attribute conditions determined by the input attribute condition determining means, An input attribute condition having high importance as a factor of an output attribute condition corresponding to two data groups is selected, and the rank determining unit is a part of the input attribute condition determined by the input attribute condition determining unit. It is preferable that the ranking is performed on the plurality of input attribute conditions selected by the factor determining means.

  According to the above configuration, a series of processes such as “divide the analysis data group based on the input attribute condition having the maximum division rule evaluation value and determine the input attribute condition for the divided data group”. Is repeatedly executed, it is possible to create a tree structure with the input attribute condition having the maximum division rule evaluation value as a node. Therefore, even if the analysis target is intertwined with a plurality of factors that are difficult to express with a single association rule, the factor of the output attribute condition corresponding to the second data group can be determined with sufficiently high accuracy.

  Further, according to the above configuration, the input attribute condition determining means repeatedly determines the input attribute condition for each of the input attributes and determines a plurality of input attribute conditions by the above-described repetitive processing. Then, the factor determining means selects an input attribute condition having a high importance as a factor of the output attribute condition corresponding to the second data group from the determined plurality of input attribute conditions. As a result, even if the above-described iterative process is performed to obtain a detailed factor analysis result, the result of the data analysis is not complicated, and it is very simple but highly accurate to the second data group. The cause of the corresponding output attribute condition can be determined.

  Further, according to the above configuration, the tree structure is created, and the plurality of input attribute conditions selected by the factor determining means (input attributes determined at various levels of the tree structure) by the rank determining means. (Condition) can be ranked according to importance as a factor of the output attribute condition corresponding to the second data group. As a result, it is possible to easily grasp “what priority should be taken with respect to each failure factor (input attribute condition) appearing in various layers?”.

  The “importance” of the input attribute condition is the extent to which the occurrence of the output attribute condition corresponding to the second data group is effectively eliminated when the occurrence of the input attribute condition is avoided. Represents. More specifically, the degree of importance of the input attribute condition is defined as an absolute value (number) or a relative value (basic data) that satisfies the input attribute condition in the basic data group and corresponds to the second data group. The ratio of the total number of data satisfying the input attribute condition in the group).

  In the data analysis device of the present invention, the plurality of input attribute conditions determined by the input attribute condition determination unit in the data analysis device having the above configuration include a first pattern that “input attribute is equal to or less than a threshold value”, “ A second pattern that the input attribute exceeds a threshold value, and the factor determination means includes the input attribute having the highest importance in the first pattern among a plurality of input attribute conditions in the same input attribute. The condition and the input attribute condition having the high importance in the second pattern may be selected.

  According to the above configuration, the plurality of input attribute conditions determined by the input attribute condition determining means include the first pattern that “the input attribute is equal to or less than the threshold” and the second pattern that “the input attribute exceeds the threshold”. Even in the case of including the pattern, it is possible to determine the factor of the output attribute condition corresponding to the second data group with high accuracy for both of the patterns in a very simple form.

  The data analysis apparatus according to the present invention is the data analysis apparatus having the above-described configuration, wherein the factor determination means has an input attribute that maximizes the second data group separation degree among a plurality of input attribute conditions to be selected. It is preferable that the condition is selected as the input attribute condition having a high importance.

  According to the above configuration, the input attribute condition that maximizes the second data group separation degree is selected as the input attribute condition having the high importance. Since the second data group separation degree of the input attribute condition represents the probability that the output attribute condition corresponding to the second data group can be cut out by the input attribute condition, in the above configuration, the output attribute corresponding to the second data group It is possible to select an input attribute condition with the highest degree of accuracy with which the condition can be extracted by the input attribute condition. Therefore, it is possible to extract an input attribute condition that is a factor of an output attribute condition corresponding to the second data group with high accuracy that is not significantly affected by other input attributes (disturbances).

  The plurality of input attribute conditions determined by the input attribute condition determining means include a first pattern that “the input attribute is equal to or less than the threshold value” and a second pattern that “the input attribute exceeds the threshold value”. If included, the factor determining means selects each of the first pattern and the second pattern from among a plurality of input attribute conditions in the same input attribute determined by the input attribute condition determining means, as a selection target. The input attribute condition that maximizes the second data group separation degree in each of the first pattern and the second pattern may be selected as the condition having the higher importance level. Of the plurality of input attribute conditions for the same input attribute determined by the determining means, all input attribute conditions including the first pattern and the second pattern are to be selected. And, an input attribute condition where the second data group separation among these input attribute condition is maximum, may be adapted to select a high condition of the importance.

  In the data analysis apparatus according to the present invention, in the data analysis apparatus having the above configuration, the factor determination unit has a higher score in descending order of the number of corresponding second data groups among the plurality of input attribute conditions to be selected. The input attribute condition that maximizes the total score of the first score given in this way and the second score given so that the score increases in descending order of the second data group separation degree is the important attribute It may be selected as a high-level input attribute condition.

  According to the above configuration, the input attribute condition that maximizes the total score of the first and second score points is selected as the input attribute condition having the high importance. Therefore, in consideration of both the second data group hit count and the second data group separation degree, an input attribute condition that can extract many output attribute conditions corresponding to the second data group with relatively high accuracy is extracted. Can do.

  The data analysis apparatus according to the present invention is the data analysis apparatus having the above-described configuration, wherein the rank determination unit is configured to output the second data with respect to at least some of the plurality of input attribute conditions determined by the input attribute condition determination unit. The configuration may be such that the ranking is performed in descending order of the number of hits in the group.

  According to the above configuration, since the ranking is performed in descending order of the number of hits in the second data group, the second attribute is generated with respect to a plurality of input attribute conditions that cause an output attribute condition corresponding to the second data group. Ranking can be performed in descending order of the number of output attribute conditions corresponding to the data group.

  Therefore, for example, when the first data group is a good data group and the second data group is a defective data group, the order of the input attribute conditions can be eliminated by avoiding the input attribute conditions. The higher the number of non-defective products, the higher the ranking can be made. As a result, when it is desired to eliminate as many defective products as possible regardless of the decrease in non-defective products, it is possible to clearly grasp in what priority order measures should be taken.

  The data analysis apparatus according to the present invention is the data analysis apparatus having the above-described configuration, wherein the rank determination unit is configured to output the second data with respect to at least some of the plurality of input attribute conditions determined by the input attribute condition determination unit. The configuration may be such that the ranking is performed in descending order of group separation.

  According to the above configuration, since the ranking is performed in descending order of the second data group separation degree, among the plurality of input attribute conditions that cause the output attribute condition corresponding to the second data group, the second data An input attribute condition having a high probability of generating an output attribute condition corresponding to a group and a low probability of generating an output attribute condition corresponding to the first data group can be given a higher rank.

  Therefore, for example, when the first data group is a good data group and the second data group is a defective data group, the order of the input attribute conditions can be eliminated by avoiding the input attribute conditions. The higher the ratio of the number of non-defective products to the number of non-defective products lost by avoiding the input attribute condition, the higher the ranking can be made. As a result, regardless of the number of defective products to be resolved, what priority should be taken when it is desired to minimize the decrease in non-defective products and eliminate the occurrence of defective products with high accuracy? Can be clearly understood.

  The data analysis apparatus according to the present invention is the data analysis apparatus having the above-described configuration, wherein the rank determination unit is configured to output the second data with respect to at least some of the plurality of input attribute conditions determined by the input attribute condition determination unit. The sum of the first score given so that the score becomes higher in descending order of the number of corresponding groups and the second score given so that the score becomes higher in descending order of the second data group separation degree is large The ordering may be performed in order.

  According to the above configuration, since the ranking is performed in descending order of the total score of the first and second score points, a plurality of input attributes that cause a generation of an output attribute condition corresponding to the second data group Among the conditions, the ranking may be performed in an order that takes into account the large number of output attribute conditions corresponding to the second data group and the high probability of generating the output attribute condition corresponding to the second data group. it can.

  Therefore, for example, when the first data group is a good data group and the second data group is a defective data group, the order of the input attribute conditions can be eliminated by avoiding the input attribute conditions. It is possible to make the order in consideration of the large number of non-defective products and the small number of non-defective products lost by avoiding the input attribute condition. As a result, when it is desired to eliminate as many defective products as possible while suppressing the decrease in non-defective products, it is possible to clearly grasp in what priority order measures should be taken.

  In the data analysis apparatus according to the present invention, in the data analysis apparatus having the above configuration, the analysis data group extraction unit extracts only the other data group from the data group divided by the division unit as a new analysis data group. It is preferable.

  According to the above configuration, since only the other data group among the data group divided by the dividing unit is repeatedly executed as a new analysis data group, the output attribute condition corresponding to the second data group A simple and sufficient factor analysis result can be obtained.

  Further, according to the above configuration, since only the other data group among the data group divided by the dividing unit is repeatedly executed as a new analysis data group, the series of processes before that is repeated. Thus, the influence of the factor (input attribute condition) extracted in the above process can be excluded, and a new factor of the output attribute condition corresponding to the second data group can be extracted with high accuracy.

  The data analysis apparatus of the present invention further includes an end condition determination unit that determines whether or not an end condition is satisfied in the data analysis apparatus having the above configuration, and the end condition determination unit determines that the end condition is satisfied. Then, the execution of the series of processes is ended, and the end condition determining means has zero data in the second data group in the analysis data group extracted by the analysis data group extraction means. It is preferable to determine whether there is an end condition.

  According to the above configuration, since the execution of the series of processes ends when the number of data in the second data group in the analysis data group becomes 0, the cause of the output attribute condition corresponding to the second data group is It is possible to avoid performing a series of processes on an analysis data group that does not include the second data group, which is an unnecessary process for determination. Therefore, the data analysis process can be promptly terminated, and the time required for the data analysis process can be shortened.

  In the data analysis apparatus according to the present invention, in the data analysis apparatus having the above-described configuration, all the input attributes in the analysis data group are numerical attributes, and the first evaluation unit is configured for all numerical values of the input attributes. In the first data group of the analysis data group, the ratio of the number of data whose input attribute is equal to or less than the numerical value is calculated as a first frequency, and in the second data group of the analysis data group, Frequency calculation means for calculating the ratio of the number of data whose input attribute is less than or equal to the value as a second frequency, and the difference between the first frequency and the second frequency for all the numerical values of each input attribute It is preferable to include a difference calculating means for calculating as an input attribute condition evaluation index.

  According to the above configuration, it is a very simple calculation process that only calculates the difference value between the first frequency and the second frequency for each of all the numerical values of each input attribute, and corresponds to the improvement degree of the Gini index method. An input attribute condition evaluation index to be obtained can be obtained. That is, in the above configuration, unlike the Gini index method, it is not necessary to perform enormous arithmetic processing such as calculating the Gini index and the improvement degree for every pattern of all branch conditions that the input attribute can take. It is only necessary to perform a calculation process for obtaining a difference value with respect to data corresponding to the number of values to be obtained. Therefore, the number of values that one input attribute can take is on the order of tens of thousands to hundreds of thousands, as in the case where the analysis data group is data of a manufacturing process of a product such as an actual device (especially a semiconductor device). Even in such a case, almost no calculation load is applied, and processing can be performed in a short time. That is, it is possible to determine an input attribute condition (optimal branch condition for each input attribute) that separates the first data group and the second data group in a short time without requiring a calculation load. Therefore, the efficiency of data analysis can be improved.

  The data analysis apparatus according to the present invention is the data analysis apparatus having the configuration described above, wherein the second evaluation unit is configured to perform the first analysis of the analysis data group for each of the input attribute conditions determined by the input attribute condition determination unit. The ratio of the ratio of the number of data satisfying the input attribute condition in the second data group of the analysis data group to the ratio of the number of data satisfying the input attribute condition in the data group is calculated as the division rule evaluation value. It is preferable.

  According to the above configuration, for each of the input attribute conditions determined by the input attribute condition determining means, the degree to which the second data group can be effectively detected by separating from the first data group by the input attribute condition, that is, “ If the input attribute satisfies the input attribute condition, it is possible to easily calculate a division rule evaluation value representing the probability of the second correlation rule that the data is included in the second data group in the analysis data group. it can.

  The data analysis apparatus according to the present invention further includes classification condition setting means for setting a classification condition in the data analysis apparatus having the above-described configuration, and the classification means includes the basic data based on the value of the output attribute and the classification condition. It is preferable that it is the structure which classify | categorizes a group.

  According to the above configuration, the user can appropriately set the output attribute condition corresponding to the second data group, and therefore the factor (input attribute condition) is determined corresponding to various output attribute conditions. be able to.

  In the data analysis method of the present invention, in the data analysis method, the input attribute is a manufacturing process condition and / or an inline inspection result in a product manufacturing process, and the output attribute is a product quality determination result, The second data group is preferably a data group with a poor quality determination result.

  According to the above method, the second data group includes the manufacturing process conditions in the manufacturing process and the characteristics (in-line inspection result) conditions during the manufacturing, which are factors that cause defective products (products whose quality judgment results are defective). Can be specified as an input attribute condition that causes an output attribute condition corresponding to. Furthermore, it is possible to easily grasp “what priority should be taken to prevent the occurrence of defective products” for a plurality of input attribute conditions.

  According to the apparatus, method, program, and recording medium of the present invention, a plurality of output attribute condition factors (particularly failure factors) corresponding to the second data group are determined, and the second data group corresponds to these factors. Based on a clear index such as the number and the second data group separation degree, ranking is performed according to the importance degree. Therefore, it is possible to determine the factor (especially the cause of failure) of the output attribute condition corresponding to the second data group with a high degree of accuracy in a very concise form, and to rank the importance of the factor (especially measures should be taken) The priority of factors) can be clarified.

[First Embodiment]
One embodiment of the present invention will be described below.

  First, the data analysis apparatus of this embodiment is demonstrated based on FIG. As shown in FIG. 1, the data analysis apparatus includes a basic data group storage unit 1, a character-numerical data conversion unit 2, a classification condition setting unit (classification condition setting unit) 3, a data classification unit (classification unit) 4, and after classification. Basic data group storage unit 5, analysis data group extraction unit (analysis data group extraction unit) 6, data row separation unit 7, data string extraction unit 8, frequency calculation unit (frequency calculation unit) 9, frequency cumulative difference calculation unit (difference) Calculation means) 10, input attribute condition determination section (input attribute condition determination means) 11, defective product data calculation section 12, frequency cumulative ratio calculation section (second evaluation means) 14, data division section (division means) 15, end Condition determining unit (end condition determining unit) 16, factor determining unit (factor determining unit) 17, rank determining unit (rank determining unit) 13, composite factor failure number calculating unit 18, numerical value-character data converting unit 19, analysis result data Storage unit 20 and output It is equipped with a 21. The data row separation unit 7, the data string extraction unit 8, the frequency calculation unit (frequency calculation unit) 9, and the frequency cumulative difference calculation unit (difference calculation unit) 10 include the first evaluation in the claims. Means.

  Next, the data analysis apparatus and the data analysis method of this embodiment will be described by taking as an example the case where the data group DA in Table 1 is a basic data group. The basic data group DA in Table 1 is stored in the basic data group storage 1 such as a hard disk.

  The basic data group DA in Table 1 is composed of 12 pieces of data having ids (identifiers) of 1 to 12. In Table 1, x1, x2, x3, and x4 are input attributes. The input attribute x1 is a character attribute that takes one of four characters A, B, C, and D. The input attribute x2 is a character attribute that takes one of the four characters a, b, c, and d. The input attribute x3 is a discrete attribute that takes one of four discrete values 1, 2, 3, and 4. The input attribute x4 is a discrete attribute taking any one of four discrete values 10, 20, 30, and 40. The input attribute may be any of a character attribute, a discrete numerical attribute, and a continuous numerical attribute.

  In Table 1, y is an output attribute. The output attribute may be any of a character attribute, a discrete numerical attribute, and a continuous numerical attribute. Here, the output attribute is a character attribute that takes one of the two characters XZ and Y.

  The data analysis apparatus and the data analysis method according to the present embodiment analyze a causal relationship between an input attribute and an output attribute, and extract information indicating the causal relationship. Here, the case where y = Y is considered as a problem event, and the cause of the output attribute y being Y is analyzed.

  As examples of the basic data group DA, for example, an input attribute is a manufacturing process condition and / or an inline inspection result (inspection result in the middle of a manufacturing line) in a product manufacturing process, an output attribute is a product quality determination result, Data in which the problem event y = Y is a bad quality determination result is exemplified. In this case, by analyzing the causal relationship between the input attribute and the output attribute by using the data analysis apparatus and the data analysis method of the present embodiment and deriving the cause of the problem event y = Y, the cause of the defect in the manufacturing process can be extracted. Therefore, it is possible to easily take measures to eliminate the occurrence of defective products. In addition, an efficient countermeasure can be executed by prioritizing a plurality of factors extracted as factors of defects. Therefore, it is possible to easily improve the manufacturing process such as improvement in yield.

  As a more specific example of the basic data group DA, for example, input attributes x1, x2, x3, and x4 are process data such as gas flow rate, gas pressure, input power, and film formation time of plasma CVD process, and output. Data in which the attribute y is the film thickness of the thin film to be formed can be mentioned. The values of the input attribute and the output attribute may be any of a continuous numerical attribute, a discrete numerical attribute, and a character attribute. In the case of the character attribute, for example, the output attribute is an example of the film thickness, and is expressed as “large”, “medium”, and “small”.

  Hereinafter, the operation of the data analysis apparatus of FIG. 1 will be described using the flowchart of the data analysis method shown in FIG.

[Step 0]
First, the character-numeric data conversion unit 2 performs numeric conversion processing on non-numeric data in the basic data group DA in Table 1 stored in the basic data group storage unit 1 which is a storage means such as a hard disk. (Hereinafter, the data of each table is stored in an appropriate storage means such as a hard disk or RAM, and is read out from these data and is subject to calculation by calculation means such as a CPU). Here, the character attributes in the basic data group DA are converted into numerical attributes (numerical data) according to the following conversion rules (S0). Note that this processing is omitted when the input attribute and output attribute of the basic data group DA are originally numeric attributes.
(X1) A → 1, B → 2, C → 3, D → 4
(X2) a → 1, b → 2, c → 3, d → 4
(X3) No conversion (x4) No conversion (y) XZ → 1, Y → 2
Through the above processing, each data is converted into numerical data. Then, the character-numeric data conversion unit 2 sends the converted data group DA0 to the data classification unit 4.

  Here, it is preferable that the above conversion rule is set so that the numerical value of the output attribute increases as the input attribute value after conversion becomes larger as much as possible, or vice versa. There is no limitation to the above example. The data group DA0 converted into numerical data by the conversion rule is as shown in Table 2.

  The data group DA0 obtained by this conversion is a set of data composed of a plurality of input attributes (description attributes) and output attributes (target attributes), each consisting of a numerical attribute. Hereinafter, the data group DA0 is also referred to as a basic data group.

[Step 1]
The classification condition setting unit 3 performs basic data group DA in accordance with predetermined setting information or in response to a user inputting a problem event attribute value y = Y from an input unit such as a keyboard or a mouse (not shown). A classification condition is set as a condition as to whether or not the condition (y = 2) of the output attribute y of the basic data group DA0 corresponding to the problem event y = Y, and is output to the data classification unit 4 (S1). In this example, the condition of the output attribute y of the basic data group DA0 corresponding to the problem event y = Y of the basic data group DA is y = 2.

[Step 2]
Next, based on the value of the output attribute y of the basic data group DA0 and the classification condition (the following comparison logic (1) (2)) output from the classification condition setting unit 3, the data classification unit 4 The group DA0 is classified into a first data group DA1 and a second data group DA2, and as shown in Table 3, classification flags (“DA1”, “DA2”) corresponding to each of the data groups DA1 and DA2 are basically used. It is given for each data of the data group DA0 (S2). Hereinafter, the data group in Table 3 is referred to as a post-classification basic data group DA00. The post-classification basic data group DA00 is stored in the post-classification basic data group storage 5 such as a hard disk. In this example, the classification condition includes the value of the output attribute y of the basic data group DA0 and the value of the output attribute y of the basic data group DA0 corresponding to the problem event y = Y of the basic data group DA (= 2). Comparison logic based on comparison.

(1) y ≠ 2 (y = 1) → DA1
(2) y = 2 → DA2

  Here, the second data group DA2 is a data group representing a problem event (for example, a device characteristic failure or the like). That is, the second data group DA2 is a data group in which the output attribute y is an attribute value (y = 2) representing a problem event, and the first data group DA1 is an attribute value (y) in which the output attribute y does not represent a problem event. = 1).

  The classification of the basic data group DA0 by the data classification unit 4 is not limited to the above logic, but is based on the logic shown in the following classification conditions based on the comparison between the value of the output attribute y of the basic data group DA0 and the threshold value yth. You may go.

(1 ') y> yth → DA1
(2 ′) y ≦ yth → DA2
Further, the basic data group DA0 may be classified by a logic based on a logical sum or a logical product of a plurality of conditions expressed as a comparison result with a plurality of thresholds, for example, a logic as shown in the following classification condition.

(1 ″) yth1 <y ≦ yth2 → DA1 (yth1, yth2: threshold)
(2 ″) y ≦ yth1 OR y> yth2 → DA2
Further, when there are a plurality of output attributes (y1, y2), the logical sum of the conditions for each output attribute expressed as a comparison result between each output attribute and each threshold value for each of these output attributes Or you may classify | categorize by the logic based on a logical product, for example, the logic as shown in the following classification conditions.

(1 ″ ′) y1 ≦ yth1 OR y2> yth2 → DA1
(2 ″ ′) y1> yth1 AND y2 ≦ yth2 → DA2
[Step 3]
The analysis data group extraction unit 6 extracts the analysis data group DA00 ′ to be analyzed from the classified basic data group DA00 and sends it to the data row separation unit 7.

  In this first process, the same data as the post-classification basic data group DA00 is extracted as the analysis data group DA00 ′. However, in the process of repeated processing described later, another data group output by the data dividing unit 15 is extracted. The

[Step 4]
Based on the classification flags of the first data group DA1 and the second data group DA2 in the analysis data group DA00 ′ (in the first processing, the basic data group DA00 after classification: Table 3). Then, the analysis data group DA00 ′ is divided into two, and each data group is extracted. Table 4 shows the first data group DA1 output from the data row separation unit 7, and Table 5 shows the second data group DA2.

  In the following description, the first data group DA1 is appropriately referred to as a non-defective product (OK product) data group, and the second data group DA2 is referred to as a defective product (NG product) data group.

[Step 5]
Next, the data string extraction unit 8 extracts each data string of the input attribute xj (1 ≦ j ≦ 4) from the good first data group DA1 (Table 4) (S5). This data string is called a 1-xj data group.

  Similarly, the data string extraction unit 8 extracts each data string of the input attribute xj (1 ≦ j ≦ 4) from the defective second data group DA2 (Table 5) (S5). This data string is called a 2-xj data group.

  The 1-xj data group is shown in Tables 6 to 9, and the 2-xj data group is shown in Tables 10 to 13.

[Step 6]
The frequency calculation unit 9 calculates each of the 1-xj data group extracted from the good first data group DA1 in step 5 and the 2-xj data group extracted from the defective second data group DA2 in step 5. Each of them is sorted in ascending order by the value of the input attribute xj (sorting process 1). Then, for each numerical value of the input attribute xj, 1-xj frequency cumulative% representing the ratio of the number of data less than that value in the first data group and 2 representing the ratio of the number of data less than that value in the second data group. -Xj Frequency cumulative% is calculated (S6).

  Here, Tables 14 to 17 in which Tables 6 to 9 are rearranged in ascending order by the value of the input attribute xj are used, and the number of data in each row (id) in the table is equal to or greater than the position of the data. The ratio of the first data group to the total number of data (= 8) is calculated as 1-xj frequency cumulative%. For example, in Table 14, eight x1 values are arranged from top to bottom in order from 1 to 4, and when there are a plurality of the same values of x1, they are arranged from top to bottom in the order of id. Focusing on the 4th x1 = 1 (id = 12) from the beginning, since there are four data (in this case, all x1 = 1) in the row including this row, 1-x1 frequency cumulative% is 4/8 = 50%. Similarly, using Table 18 to Table 21 in which Table 10 to Table 13 are rearranged in ascending order by the value of the input attribute xj, the number of data in each row (id) is equal to or greater than the position of the data in the table. The ratio of the second data group to the total number of data (= 4) is calculated as 2-xj frequency cumulative%.

  The values of 1-xj frequency cumulative% and 2-xj frequency cumulative% calculated here are shown in Tables 14 to 21.

  Further, the frequency calculation unit 9 is a table of 1-xj data groups that are non-defective product data groups for which 1-xj frequency cumulative% is calculated, and 2-items that are defective product data groups for which 2-xj frequency cumulative% is calculated. The table of the xj data group is joined (joining process). Specifically, for the input attribute x1, Table 14 and Table 18 are combined to combine the x1 frequency accumulation table of Table 22 (a), and for the input attribute x2, Table 15 and Table 19 are combined to The x2 frequency accumulation table is combined with Table 16 and Table 20 for the input attribute x3 to combine the x3 frequency accumulation table of Table 24, and the input attribute x4 is combined with Table 17 and Table 21 to generate the x4 frequency of Table 25. Each accumulation table is created (S6).

  Furthermore, the frequency calculation unit 9 sorts the frequency accumulation tables of Table 22 (a), Table 23, Table 24, and Table 25 in ascending order by the value of the input attribute xj (sorting process 2). Then, after performing the rearrangement process 2, values immediately above are substituted in the blanks of the 1-xj frequency accumulation% and the 2-xj frequency accumulation% in order from the top (substitution process). Thereafter, only the data of the last line among those lines is adopted for the lines having the same value in the input attribute xj (duplicate processing). As a result, Tables 26 to 29 are created. For example, when creating the table 26 from the table 22 (a), if the sorting process 2 is performed in the table 22 (a), each value of the 1-x1 frequency accumulation% column and the 2-x1 frequency accumulation% column Each value is arranged with a blank space between each value while maintaining the respective order (Table 22 (b)). In the next substitution process, these blanks are processed (colored portion in Table 22 (c)). For example, since there is no value corresponding to x1 = 2 in the column of 2-x1 frequency accumulation%, x1 = 1 immediately above the column corresponding to x1 = 2 of 2-x1 frequency accumulation% is obtained by substitution processing. The value 25% of the 2-x1 frequency accumulation% corresponding to is substituted. Next, in this Table 22 (c), there are 5 x1, 2 is 2, 3 is 2, 4 is 3, and each x1 value is duplicated by duplication processing. , Adopt the data of the last line. For example, the value 50% corresponding to x1 = 1 of the last row is adopted as the value of 1-x1 frequency cumulative% corresponding to x1 = 1, and the value of 1-x1 frequency cumulative% corresponding to x1 = 2 is adopted as the final row. A value of 75% corresponding to x1 = 2 is adopted. Further, the value 100% corresponding to x1 = 4 in the last row is adopted as the value of 2-x1 frequency cumulative% corresponding to x1 = 4.

  Thus, as shown in Tables 26 to 29, for each value of the input attribute xj, the 1-xj frequency representing the ratio of the number of data whose input attribute is less than or equal to the numerical value in the first data group which is a non-defective data group Cumulative% (A; first frequency) and 2-xj frequency cumulative% (B; second frequency) representing the ratio of the number of data whose input attributes are less than or equal to the numerical value in the second data group that is a defective product data group ) Are calculated (S6).

  In the above steps 4 to 6, the xj frequency accumulation table of Table 26 to Table 29 is created in order to create the data row separation process (Tables 4 and 5) → the data string extraction process (Tables 6 to 13) → the arrangement. Replacement process 1 → Calculation process of 1-xj frequency accumulation% and 2-xj frequency accumulation% (Table 14 to Table 21) → Combination process (Table 22 (a), Table 23, Table 24, Table 25) → Reordering process 2 (Table 22 (b)) → Substitution processing (Table 22 (c)) → Duplicate processing (Table 26 to Table 29), but without performing these individual processing, it is directly performed collectively. The xj frequency accumulation table shown in Tables 26 to 29 may be calculated. Further, only some of the individual processes may be processed at once.

[Step 7]
Next, the frequency cumulative difference calculation unit 10 subtracts the 1-xj frequency cumulative% (A) of the non-defective product from the 2-xj frequency cumulative% (B) of the defective product for each value of the input attribute xj. Xj frequency cumulative difference (= B−A), and second xj frequency cumulative difference (= B−A) obtained by subtracting 2-xj frequency cumulative% (B) of defective products from 1−xj frequency cumulative percent (A) of non-defective products A difference between A and B, that is, (A−B) is calculated (S7). The first xj frequency cumulative difference (= B−A) and the second xj frequency cumulative difference (= A−B) are collectively referred to as an xj frequency cumulative difference. This xj frequency cumulative difference corresponds to the input attribute condition evaluation index in the claims.

  Tables 30 to 33 show the calculation results of the xj frequency cumulative difference (first xj frequency cumulative difference (= BA), second xj frequency cumulative difference (= AB)). Further, the value of the input attribute xj, the non-defective 1-xj frequency cumulative% (A), the defective 2-ch frequency cumulative% (B), the first xj frequency cumulative difference (= B−A), the second FIG. 3 to FIG. 6 show the relationship with the xj frequency cumulative difference (= A−B).

  The xj frequency cumulative difference (first xj frequency cumulative difference (= B−A), second xj frequency cumulative difference (= A−B)) in each numerical value of the input attribute xj indicates that the input attribute xj is less than or equal to the numerical value. Whether or not the first data group DA1 of good products and the second data group DA2 of defective products can be successfully separated by dividing the analysis data group DA00 ′ into two ranges into the range in which the input attribute xj exceeds the numerical value. It is an index representing

  That is, the first xj frequency cumulative difference (= B−A) in each numerical value of the input attribute xj is “if the input attribute xj is equal to or smaller than the numerical value, it is data belonging to the second data group DA2 of defective products and input. It represents the probability of the correlation rule that if the attribute xj exceeds the numerical value, the data belongs to the first non-defective data group DA1.

  Further, the second xj frequency cumulative difference (= A−B) in each numerical value of the input attribute xj is “data belonging to the second data group DA2 of defective products if the input attribute xj exceeds the numerical value, If the input attribute xj is equal to or less than the numerical value, it indicates the likelihood of the correlation rule that it is data belonging to the first good data group DA1.

  Each of the above two association rules is analyzed for each input attribute condition (input attribute xj is less than or equal to the numerical value or input attribute xj exceeds the numerical value) if the input attribute satisfies the input attribute condition. The data belongs to the second data group DA2 in the data group DA00 ′, and if the input attribute does not satisfy the input attribute condition, the data belongs to the first data group DA1 in the analysis data group DA00 ′. Corresponds to the association rule.

  The processes in steps 4 to 7 correspond to the first evaluation step in the claims. In addition, the data row separation unit 7, the data string extraction unit 8, the frequency calculation unit (frequency calculation unit) 9, and the frequency cumulative difference calculation unit (difference calculation unit) 10 used for the processes in steps 4 to 7 are patented. It constitutes the first evaluation means in the claims.

[Step 8]
For each input attribute xj, the input attribute condition determination unit 11 determines an xj frequency cumulative difference corresponding to each value of xj (first xj frequency cumulative difference and second xj frequency cumulative difference: input attribute condition) Among the evaluation indices, the condition (value of the input attribute xj and the type of xj frequency cumulative difference (first xj frequency cumulative difference or second xj frequency cumulative difference)) that maximizes the value Extract (S8). In Tables 30 to 33, the extracted conditions are colored.

  The value of the extracted input attribute xj will be referred to as an input attribute threshold value xj-th. As can be seen with reference to FIGS. 3 to 6, the input attribute threshold value xj-th is divided into a range of xj ≦ xj-th and a range of xj> xj-th, so that the non-defective first data group DA1. And the value of the input attribute xj that makes it easy to separate the defective product from the second data group DA2.

  When the first xj frequency cumulative difference (= B−A) is extracted as the type of xj frequency cumulative difference, the input attribute condition “xj ≦ xj−th” is included in the second data group DA2 of defective products. Correspondingly, when the second xj frequency cumulative difference (= A−B) is extracted, the input attribute condition “xj> xj−th” corresponds to the defective second data group DA2. That is, the input attribute condition determination unit 11 sets the input attribute condition (“xj ≦ xj−th” or “xj> xj−th”) having the maximum input attribute condition evaluation index for each input attribute xj. The input attribute condition satisfying the first correlation rule is determined.

  Table 34 shows the input attribute conditions determined by the input attribute condition determination unit 11 as conditions corresponding to the second data group DA2 of defective products for each input attribute xj. As an example, the input attribute condition “x2> 2” is determined for the input attribute x2, but this condition is separated from the non-defective first data group DA1 and the second data group DA2 of defective products is high. The conditions that can be detected with accuracy are shown. Also, “x2 ≦ 2”, which is an exclusive condition for the determined input attribute condition “x2> 2”, separates the defective second data group DA2 from the defective first data group DA1 with high accuracy. The conditions that can be detected are shown. These can be better understood with reference to FIG.

  As described above, the input attribute condition determination unit 11 determines the difference (xj frequency accumulation) between the first frequency (1-xj frequency accumulation% of non-defective products) and the second frequency (2-xj frequency accumulation% of defective products). Difference) is used as an input attribute condition evaluation index when separating the first data group of non-defective products and the second data group of defective products, so that the input satisfying the first correlation rule can be achieved with very simple processing. Attribute conditions can be determined.

In the above description, the processing from step 5 to step 8 is collectively performed for a plurality of input attributes. However, the value of j may be sequentially increased from 1 to 4 and the processing from step 5 to step 8 may be repeated. .
Here, although the xj frequency cumulative difference is used as the input attribute condition evaluation index, an index for evaluating the degree of data bias, for example, the improvement degree of the Gini index method described in Non-Patent Document 1, the information gain ( The input attribute condition may be determined using (Gain) or the like. However, from the viewpoint of reducing the calculation load and improving the calculation speed, it is preferable to use the xj frequency cumulative difference as the input attribute condition evaluation index. That is, the input attribute condition determination unit 11 of the present embodiment calculates only the xj frequency cumulative difference%, which is the difference between the non-defective 1-xj frequency cumulative% and the defective 2-xj frequency cumulative%. The input attribute condition evaluation index (corresponding to the improvement degree of the Gini index method) is obtained by simple processing. Therefore, as in the Gini index method, an enormous amount of arithmetic processing such as calculating the Gini index (Equation (1)) and the improvement level (Equation (2)) is performed for every branch condition pattern that the input attribute can take. Without performing this, only the arithmetic processing for obtaining the frequency cumulative difference is performed on the data (Tables 30 to 33) having the number of rows corresponding to the number of values that the input attribute can take. Therefore, even if the number of values that can be taken by one input attribute is on the order of tens of thousands to hundreds of thousands as in the data of the manufacturing process of a product such as an actual device (especially a semiconductor device), Table 30 to Since only the number of data rows in Table 33 is increased, almost no calculation load is applied, and processing can be performed in a short time.

  Further, the 1-xj frequency cumulative% of non-defective products and the 2-xj frequency cumulative% of defective products for each value of the input attribute xj respectively indicate the number of data whose input attributes are equal to or less than the numerical value in the corresponding data group. Since the data is normalized by the total number of data in the data group, the xj frequency cumulative difference%, which is the difference between them, is used as the input attribute condition evaluation index, whereby the non-defective product in the analysis data group (first data group) Even if the percentage of defective products and the proportion of defective products (second data group) are extremely different, the input attribute condition (each input) that separates good products from defective products without degrading the accuracy of the input attribute condition evaluation index The optimal branch condition in the attribute) can be determined with high accuracy.

[Step 9]
For each of the input attribute conditions (Table 34) determined by the input attribute condition determination unit 11 by the defective product data calculation unit 12, the input is performed in the classified basic data group DA00 (not the analysis data group DA00 ′). The number of data satisfying the attribute condition ("DA1 + DA2" column in Table 35) and the number of data (defective product data) satisfying the input attribute condition and corresponding to the defective second data group DA2 (Table 35) "DA2" column). The number of the “DA2” column represents the number of defects (number of defective product data) caused by each input attribute condition in the table, and corresponds to the number corresponding to the second data group in the claims.

  Next, the defective product separation degree is calculated by dividing the value in the “DA2” column of Table 35 by the value in the “DA1 + DA2” column. The defective product separation degree of each input attribute condition is the ratio of the number of data in which the second data group DA2 that is a defective product is included in the data satisfying the input attribute condition, that is, the input attribute in the classified basic data group DA00. This represents the ratio of the number of data corresponding to the second data group DA2 of defective products in the data that satisfies the conditions. The degree of defective product separation for each input attribute condition is the accuracy of defective product extraction based on the input attribute condition (defective rate when data belonging to the input attribute condition in the basic data group DA00 after classification is used as a population) Data ratio)), which corresponds to the second data group separation degree in the claims.

  Table 35 shows the total number of data in the basic data group DA00 after classification (= 12: “DA1 + DA2” column) and the number of defective products in the “Total” row, together with the calculation result by the defective data calculation unit 12. 10 is a table showing the number of second data groups DA2 (= 4: “DA2” column) and defective product content rate (= 4/12 = 0.333: “defective product separation” column). The defective product content rate is a ratio of the number of data in the second data group DA2 included in the basic data group DA00 after classification, and is a defective rate when all data in the basic data group DA00 after classification is a population. Represents.

  The meaning of each column in Table 35 can be easily understood by referring to FIG. 7A to FIG. FIGS. 7A to 7D are Venn diagrams showing the relationship between a set of data satisfying each input attribute condition in Table 35 and a set of defective second data group DA2.

  As described above, the input attribute conditions “x1> 2”, “x2> 2”, “x3” determined by the input attribute condition determination unit 11 as the cause of the problem event (the second data group DA2 of defective products). For “> 2” and “x4 ≦ 10”, the number of defects (the number corresponding to the second data group) and the degree of defective product separation (second data group separation degree) were calculated.

  In step 1 to step 9 described above, an input attribute condition that causes a problem event (defective product second data group DA2) is determined for each input attribute. However, the input attribute condition evaluation index (xj frequency cumulative difference) for each input attribute calculated in the process (steps 6 to 7) includes the influence of other input attributes as disturbances. There is a possibility that the accuracy of the input attribute condition evaluation index (xj frequency cumulative difference) has been lowered. Further, in the case where there are two types of factors of the problem event “xj ≦ xj−th1” and “xj> xj−th2” in a certain input attribute xj, only one of the conditions is determined. . In order to eliminate these points, it is preferable to further perform processing according to the following steps.

[Step 10]
For each input attribute threshold value xj-th (see Tables 30 to 33) extracted by the input attribute condition determination unit 11 (step 8), the frequency accumulation ratio calculation unit 14 performs the non-defective 1-xj frequency accumulation% (A). , A ratio of 2-xj frequency cumulative% (B) of defective products (= B / A: hereinafter referred to as frequency cumulative lower ratio), or a value obtained by subtracting 1-xj frequency cumulative% (A) of non-defective products from 100 Ratio (= (100−B) / (100−A)) (= (100−B) / (100−A)): (Referred to as cumulative frequency ratio) is calculated as the division rule evaluation value.

  When the input attribute condition determined by the input attribute condition determining unit 11 is of the type “xj ≦ xj−th” (when the first xj frequency cumulative difference is extracted by the input attribute condition determining unit 11) ), The frequency cumulative lower ratio (= B / A) is calculated as the division rule evaluation value. Here, the frequency cumulative lower ratio (= B / A) can be detected separately from the non-defective first data group DA1 and the defective second data group DA2 according to the input attribute condition “xj ≦ xj−th”. Represents a percentage.

  When the input attribute condition determined by the input attribute condition determining unit 11 is of the type “xj> xj−th” (when the second xj frequency cumulative difference is extracted by the input attribute condition determining unit 11) The frequency cumulative upper ratio (= (100−B) / (100−A)) is calculated as the division rule evaluation value. Here, the cumulative frequency ratio (= (100−B) / (100−A)) is separated from the non-defective first data group DA1 according to the input attribute condition “xj> xj−th”. The ratio at which the second data group DA2 can be detected is shown.

  In other words, the division rule evaluation value (frequency cumulative lower ratio or frequency cumulative upper ratio) for each input attribute condition is “if the input attribute xj satisfies the input attribute condition, the data is included in the second data group DA2.” This represents the probability of the second association rule.

  Table 36 shows division rule evaluation values (frequency cumulative lower ratio or frequency cumulative upper ratio) for each input attribute condition.

[Step 11]
The data dividing unit 15 extracts the input attribute condition having the maximum value from the division rule evaluation values (frequency cumulative lower ratio or frequency cumulative upper ratio: Table 36) in step 10 above.

  Referring to Table 36, input attribute condition “x2> 2” has the largest division rule evaluation value among all input attribute conditions, and division rule evaluation value = frequency cumulative upper ratio = ∞. This indicates that, under the input attribute condition “x2> 2,” the defective second data group DA2 can be detected by being completely separated from the non-defective first data group DA1.

  From another point of view, the input attribute condition “x2> 2” corresponds to the defective second data group DA2 regardless of the values of the other input attributes (x1, x3, x4). From determining the input attribute conditions of other input attributes (x1, x3, x4) (step 8), or calculating the input attribute condition evaluation index (xj frequency cumulative difference) (step 7), It may be a disturbance factor. In such a case, it is desirable to obtain the input attribute conditions of the other input attributes (x1, x3, x4) by excluding data corresponding to “x2> 2” from the analysis data group DA00 ′. As for the input attribute x2, in addition to “x2> 2,” the input attribute condition “x2 ≦ k” (k is an unknown number) may also be a cause of failure. This “x2 ≦ k” "Is to be extracted (determined), it is necessary to exclude the data corresponding to" x2> 2 "from the analysis data group DA00 'and determine the input attribute condition.

  Therefore, the data dividing unit 15 satisfies the analysis data group DA00 ′ based on the extracted input attribute condition “x2> 2”, the factor data group satisfying “x2> 2”, and “x2> 2”. It is divided into other data groups that do not exist (satisfying “x2 ≦ 2”). Table 37 shows the factor data group, and Table 38 shows the other data group.

  Here, the frequency cumulative lower ratio or the frequency cumulative upper ratio is calculated as the division rule evaluation value, but other evaluation indexes, for example, the improvement degree of the Gini index method described in Non-Patent Document 1, the information gain (Gain) or the like may be used.

  However, since the above division process is intended to exclude disturbance in determining the input attribute condition, “if the input attribute satisfies the input attribute condition, the data belongs to the second data group DA2 and the input Instead of the first correlation rule that the attribute belongs to the first data group DA1 if the attribute does not satisfy the input attribute condition, it is included in the second data group DA2 if the input attribute satisfies the input attribute condition. It is required to be based on the second association rule “data is”. From this viewpoint, it is preferable to use a frequency cumulative lower ratio or a frequency cumulative upper ratio as the division rule evaluation value.

[Step 12]
The analysis data group extraction unit 6 extracts another data group from the data group divided in step 11 as the next analysis data group DA00 ′. Then, the series of processing from step 3 to step 11 is repeated until the end condition determination unit 16 determines that the end condition is satisfied. The end condition determination unit 16 according to the present embodiment determines that the end condition is a case where the number of data in the second data group DA2 of defective products becomes 0 in the above-described step 4 during the iterative process. As described above, detailed factor analysis results can be obtained by repeatedly performing the process until the number of data in the second data group DA2 of defective products becomes zero.

  Note that the end condition is another end condition based on the number of data in the second data group DA2, for example, (1) when the number of data in the second data group DA2 is equal to or less than a predetermined number in the above step 4 during the repetitive processing. (2) The ratio of the number of data in the second data group DA2 to the number of data in the first data group DA1 in step 4 during the repetitive processing may be a predetermined ratio or less. Moreover, (3) It is good also as a case where the division | segmentation rule evaluation value of the input attribute conditions extracted at the said step 11 in the repetition process falls below a predetermined value. When such termination conditions are used, a simpler and sufficient factor analysis result can be obtained. Furthermore, when priority is given to obtaining a simple factor analysis result, the end condition is simply set to the case where the iterative process is performed a predetermined number of times, or the end condition determining unit 16 is omitted to perform the iterative process as much as possible. Or you may.

Table 39 (corresponding to Table 35 for the first time) shows the calculation result of the defective product data calculation unit 12 in Step 9 for the second time. In the second processing, since the other data group excluding data corresponding to “x2> 2” is the analysis data group, an input attribute condition “x2 ≦ 1” is extracted for the input attribute x2. (Determined).
Further, the division rule evaluation values (frequency cumulative lower ratio or frequency cumulative upper ratio) calculated in the second step 10 are shown in Table 40 (corresponding to the first time Table 36). In this example, the division rule evaluation value is a maximum of 4 in “x1> 2” and “x4 ≦ 10”, but the data division unit 15 has an input attribute condition that maximizes the division rule evaluation value. One of these is selected. For this selection criterion, there should be a certain rule. For example, the input attribute xj with the smallest number j is given priority, and “x1> 2” is selected. Then, the analysis data group DA00 ′ is divided into a factor data group that satisfies “x1> 2” and another data group that does not satisfy “x1> 2” (that satisfies “x1 ≦ 2”) (second step) 11).

  Next, among the data groups divided by the data dividing unit 15, another data group (a data group satisfying “x1 ≦ 2” in the second analysis data group) is analyzed data as the third analysis data group. It is extracted by the group extraction unit 6 (Table 41). However, since the analysis data group for the third time in Table 41 did not include defective product data (second data group DA2; y = 2), the iterative process was completed in Step 4 for the third time.

  A more detailed factor analysis result can be obtained by the repeated processing as described above than when the repeated processing is not performed.

  In addition, when iterative processing is not performed, even if the accuracy of the input attribute condition evaluation index (xj frequency cumulative difference) is low due to the influence of disturbance (the influence of other input attributes), the other data groups are sequentially analyzed. This problem can be solved by repeating the process.

  Furthermore, in a certain input attribute, when the factor of the output attribute condition corresponding to the second data group of the defective product is two types of “input attribute is below threshold” and “input attribute exceeds threshold” In addition, it is possible to extract both factors by sequentially performing repeated processing using another data group as an analysis data group. For example, for the input attribute x2 in this example, “x2> 2” (Table 35) can be extracted in the first process, and “x2 ≦ 1” (Table 39) can be extracted in the second process.

  Therefore, according to the above configuration, since only the other data group among the data group divided by the data dividing unit 15 is repeatedly executed as a new analysis data group, a series of previous processes are performed. The influence of the factor (input attribute condition) extracted in the process of repeating the above can be excluded, and a new factor of the output attribute condition corresponding to the second data group can be extracted with high accuracy.

[Step 13]
An input attribute condition list table (Table 35 and Table 39) that summarizes the input attribute conditions determined for each input attribute, the number of defects and the degree of defective product separation for each input attribute for each repetition process of Step 8. Table 42 shows the results of summarizing the above. In the input attribute condition list table of Table 42, all of a plurality of input attribute conditions for the same input attribute by the repetition processing of the input attribute condition determining unit 11 (step 8) are shown.

  The factor determination unit 17 causes a problem event (a factor of an output attribute condition corresponding to the second data group DA2) among a plurality of input attribute conditions for the same input attribute in the input attribute condition list table (Table 42). Only a condition having a high degree of importance, that is, a condition with a high degree of priority (hereinafter referred to as “priority”) in taking measures for solving the problem phenomenon is selected.

  In the present embodiment, the factor determination unit 17 repeats a plurality of input attribute conditions for the same input attribute in the input attribute condition list table (Table 42), that is, n times (where n is an integer of 2 or more) in step 8. Among the n input attribute conditions determined by the above, the input attribute condition that maximizes the defective product separation degree (second data group separation degree) is selected as the input attribute condition with high priority. However, when the input attribute conditions of a plurality of types of patterns are included in the n input attribute conditions, the factor determination unit 17 inputs the defective product separation degree (second data group separation degree) maximum in each pattern. Each attribute condition is selected as an input attribute condition having a high priority.

  Specifically, for the same input attribute, the input attribute condition that maximizes the defective product separation degree (second data group separation degree) in the first pattern “input attribute is below threshold” And one input attribute condition that maximizes the defective product separation degree (second data group separation degree) in the second pattern of “input attribute exceeds threshold”.

  In the case of the example in Table 42, finally, the five conditions shown in Table 43 (a) were selected as factors of the problem event (defective product second data group DA2). Table 43 (a) is a list of input attribute conditions determined (selected) by the factor determination unit 17 as the cause of the problem event, and this table is referred to as a determination factor list table. The determination factor list table is stored in the analysis result data storage unit 20.

  As described above, the factor determination unit 17 uses a plurality of input attributes in the same input attribute determined for each repetition process of step 8 based on a clear index of defective product separation (second data group separation). Since only the high-priority conditions are selected from among the conditions, the above process is repeated in order to obtain a detailed factor analysis result. The cause of the output attribute condition corresponding to the second data group can be determined by the accuracy.

  In selecting the high priority condition, the defective product separation degree (second data group separation degree) is used as an index for the following reason. That is, as described above, the degree of defective product separation under a certain input attribute condition represents the accuracy of defective product extraction according to the input attribute condition. Therefore, an input attribute condition with a large value is a value other than other input attributes (disturbances). This is because it is considered to be a condition determined with high accuracy without being affected by the above.

  If the process of the above two iterations is expressed in the form of a decision tree (similar to the conventional decision tree (FIG. 14)), it is as shown in FIG. Referring to FIG. 8, in this embodiment, for each branch of a decision tree, not only a branch condition but also an input attribute condition that causes a problem event (defective product second data group DA2) is obtained for all input attributes. (Processing by the input attribute condition determination unit 11 in step 8). Then, among all the input attribute conditions for the number of times of branching (for the number of iterations), input attribute conditions with a higher degree of defective product separation are narrowed down and determined as the final defect factor (factor determination in step 13). Processing by the unit 17).

  As described above, the input attribute condition that causes the problem event (defective product second data group DA2) is extracted not only for the branch condition in the decision tree but for all the input attributes. Even if there is, can be caught without missing the factor. In the final factor determination (processing by the factor determination unit 17), the factor is determined based on a clear index of defective product separation, so no matter how complicated the decision tree is, it is clear. The cause of the problem event can be grasped.

  The factor determination unit 17 may further perform the following processing. That is, in each input attribute condition in the determination factor list table of Table 43 (a), the defective product content rate of the basic data group DA00 after classification (= 0.333: “Defective product separation” column in the “Total” row) ) May be determined as information indicating a factor of the second data group DA2 of defective products. In the example of Table 43 (a), the four conditions other than “x2 ≦ 1” (defective product separation = 0.2) have a defective product separation higher than the defective product content in the basic data group DA00 after classification. These four conditions are selected (Table 43 (b)). The input attribute condition selected in this way is a condition that includes the defective second data group DA2 at a higher rate than the sample randomly selected from the classified basic data group DA00. Hereinafter, the table of Table 43 (b) is also referred to as a determination factor list table.

  In this example (data analysis with Table 1 as the basic data group DA), the number of data and the number of input attributes are small, and the influence between the input attributes is small. 43 (a) and Table 43 (b)) were not significantly different from the processing result (Table 35) of the input attribute condition determining unit 11 in Step 8 as a result. However, when the number of data and the number of input attributes are large, the input attribute conditions for each repetition process are often complicated as shown in the input attribute condition list table of Table 44. Table 44 is an example (dummy sample of the input attribute condition list table) in which four different input attribute conditions are determined for the input attribute xk1 by the four iterations of step 8.

  In the case of this example, the condition “xk1 ≦ 2.33” that maximizes the defective product separation degree (second data group separation degree) with respect to the first pattern “input attribute is equal to or less than threshold”. However, in the second pattern that “the input attribute exceeds the threshold”, the condition “xk1> 2.44” that maximizes the defective product separation degree (second data group separation degree) is It is selected by the factor determination unit 17 (Table 45).

  Note that if a factor is to be determined simply, the first pattern and the second pattern are not distinguished from each other, and the input attribute condition that maximizes the defective product separation degree is simply selected from the same input attributes. It may be selected (in the example of Table 44, “xk1> 2.44” is selected).

[Step 14]
Each of the input attribute conditions (each input attribute condition of the determination factor list table (Table 43 (a) or Table 43 (b)) determined (selected) by the factor determining unit 17 as the factor of the problem event. On the other hand, ranking is performed according to importance (priority) as a cause of a problem event (a factor causing an output attribute condition corresponding to the second data group DA2 of defective products). That is, an order (priority order) that should be prioritized in taking measures for solving the problem phenomenon is given (prioritization is performed). Hereinafter, the operation of the order determination unit 13 will be described by taking as an example a case in which prioritization is performed for each input attribute condition in the determination factor list table of Table 43 (b).

  The rank determination unit 13 first sets the first scoring so that the score increases in descending order of the number of defects (number of hits in the second data group) for each input attribute condition in the determination factor list table of Table 43 (b). In addition, the second scoring is given so that the score becomes higher in descending order of the defective product separation degree (second data group separation degree). Then, the total score of the first score and the second score is calculated. Table 46 is a table in which the first allocation score, the second allocation score, and the total score are written together in the determination factor list table of Table 43 (b) (this table is also referred to as a determination factor list table).

  Here, the first stipulation is the number of the order when the input attribute conditions in Table 46 are arranged in ascending order by the number of defects (that is, the order when ranking is performed in ascending order of the number of defects). Is the number of the order when the input attribute conditions in Table 46 are arranged in ascending order by the defective product separation degree (that is, the ranking when the defective product separation degree is ranked in ascending order).

  Note that the first and second points are not limited to this example. For example, the first score may be the number of defects itself, and the second score may be a value obtained by multiplying the defect separation degree by a predetermined coefficient.

  Next, according to the purpose of the analysis, the user designates which of the above three types of scores (first score, second score, and total score) to adopt to the data analyzer, and this In response to this, the order determining unit 13 prioritizes the input attribute conditions in descending order of the score.

  The determination factor list table (Table 46) is rearranged in descending order of priority determined by the rank determination unit 13, and the results displayed as the factor breakdown Pareto charts are shown in FIGS. 9 to 11, the number of defects caused by each input attribute condition (number of second data group DA2: number of corresponding second data group) is a bar graph, and the degree of defective product separation (second data group separation degree). Is shown as a line graph.

  In FIG. 9, the horizontal axis (input attribute condition) is arranged in order from the condition in which the first score is high (condition in which the number of defects is large). In FIG. 10, the horizontal axis (input attribute condition) is arranged in order from the condition where the second score is high (the condition where the defective product separation degree is large). In FIG. 11, the horizontal axis (input attribute condition) is in order from the condition where the total score of the first and second score points is high (the condition where the score is high considering both the number of defects and the degree of defect separation). Lined up. On the horizontal axis of each figure, the input attribute conditions having the same score are arranged on the left side with priority given to the condition where the number j of the input attribute Xj is young.

  By referring to the results shown in FIGS. 9 to 11, the user has a range of values for each of the input attributes x1 to x4. It is possible to judge immediately by looking at the causes of product characteristic defects. In addition, the order (priority order) in which measures should be taken can be determined from the first score, the second score, or the total score of the first score and the second score.

  Hereinafter, with reference to FIG. 9 to FIG. 11, the meaning of prioritization by the rank determination unit 13 based on the above three points (first score, second score, and total score) will be described.

(1) Prioritization based on first scoring When the purpose of analysis is to reduce as many as possible defects as much as possible, the user designates the data analyzer to adopt the first scoring. Then, the order determination unit 13 assigns priorities in order from an input attribute condition (input attribute condition with a high first score) having a large number of defects (corresponding number of second data groups) (see FIG. 9). The user can eliminate many of the currently occurring defects by taking countermeasures in order from the input attribute condition with the highest priority.

  For example, by taking measures against “x1> 2” or “x3> 2” having the first priority in FIG. 9 (Table 46), it is possible to eliminate three of the four defects. Yes (75% of all defects are eliminated).

  However, in this case, in the input attribute condition in which the first scoring is high (the number of defects is large), the second scoring is not necessarily high (the degree of defective product separation is also large). At the same time as resolution, there is a risk of dropping good products. For example, since the measure for “x3> 2” is a defective product separation degree of 0.43, four non-defective products are dropped simultaneously with the elimination of the three defects. That is, when the priority order based on the first scoring is adopted, the user refers to the defective product separation degree together with the priority order, and thereby finally determines the input attribute condition to be dealt with.

(2) Prioritization based on second scoring When the purpose of analysis is to separate defects from non-defective products and reduce defects with high accuracy, the user designates the data analyzer to adopt the second scoring. Then, the order determination unit 13 prioritizes the input attributes in descending order of input attribute conditions (input attribute conditions having a high second score) having a high degree of defective product separation (second data group separation) (see FIG. 10). The user can resolve the defect with high accuracy without dropping a non-defective product as much as possible by taking measures in order from the input attribute condition having the highest priority.

  For example, by taking measures against “x2> 2” (defective product separation degree = 1), which has the first priority in FIG. 10 (Table 46), it is possible to eliminate the defect without dropping any non-defective product. Can do.

  However, in this case, in the input attribute condition where the second scoring is high (the degree of defective product separation is large), the first scoring is not necessarily high (the number of defects is also large). There is a risk that only a part of the whole defect is resolved. That is, when the priority order based on the second score is adopted, the user refers to the number of defects together with the priority order, and thereby finally determines the input attribute condition to be countered.

(3) Prioritization based on the total score of the first and second scoring points Utilizing the advantages of prioritization based on the first scoring point and prioritization based on the second scoring point In addition, when the aim is to compensate for the shortcomings and reduce many defects with high accuracy from a comprehensive viewpoint, the user designates the data analyzer to adopt the total of the first and second points. To do. Then, the order determination unit 13 assigns priorities in descending order of the total points (scoring points considering both the number of defects and the degree of defective product separation) (see FIG. 11). The user can resolve many defects with high accuracy by taking countermeasures in order from the input attribute condition having the highest priority.

  For example, by taking measures against “x1> 2”, which has the first priority in FIG. 11 (Table 46), many defects can be eliminated with almost no non-defective products dropped. In other words, only 2 out of 8 non-defective products are dropped, and 3 out of 4 defects can be eliminated.

[Step 15]
The complex factor defect count calculation unit 18 calculates the number of defects due to the complex factor of two conditions among the input attribute conditions in the decision factor list table (Table 46) (Table 47). In Table 47, each of the title attribute and the title column shows each input attribute condition, and the number of defects (number of second data group DA2) due to the composite factor of the two input attribute conditions is shown at the intersection. It is shown.

  For example, the “x1> 2” row and the “x2> 2” column satisfy “x1> 2” and “x2> 2”, and the number of data corresponding to the defective second data group DA2 (= 1) ). Hereinafter, the table in Table 47 is referred to as a composite factor table.

[Step 16]
The numerical value-character data conversion unit 19 converts the numerical value of the input attribute threshold value xj-th in the decision factor list table (Table 46) and the composite factor table (Table 47) into character data as necessary. The conversion rule for character data is a rule that is the reverse conversion of the conversion in step 0, and is as follows.
(X1) 1 → A, 2 → B, 3 → C, 4 → D
(X2) 1 → a, 2 → b, 3 → c, 4 → d
(X3) Not converted (x4) Not converted Table 48 shows a determination factor list table in which the input attribute threshold value xj-th in the determination factor list table of Table 46 is converted into character data.

[Step 17]
Thus, the data analysis is completed, and the input attribute condition list table (Table 42), the decision factor list tables (Table 46 and Table 48), the composite factor table (Table 47) and various information in the data analysis process are finally obtained. The analysis result data is stored in the analysis result data storage unit 20 such as a hard disk. The analysis result data is appropriately sent from the analysis result data storage unit 20 to the output unit 21 such as a display device or a printing device, and a table (for example, Table 48), a decision tree (for example, FIG. 8), or a graph (for example, FIG. 9 to 11), it can be displayed on a display device or printed on a printing device.

  By referring to the graphs (FIGS. 9 to 11), the user has a range of values for each of the input attributes x1 to x4. It is possible to judge immediately by looking at the causes of product characteristic defects.

  In addition, the order of the measures to be taken is determined based on the ranking result of the rank determination unit 13 based on the first score, the second score, or the total score of the first score and the second score (FIG. 9 to FIG. 9). This can be determined with reference to the order of input attribute conditions in FIG. For example, referring to the ranking result (FIG. 11 and Table 46) of the ranking determination unit 13 based on the total score of the first score and the second score, the total score is the highest five points, “x1> 2 ”(“ x1 = C or D ”) or“ x2> 2 ”(“ x2 = c or d ”) can be determined as the first countermeasure condition. Note that it is only necessary to determine which of the two conditions is selected based on the feasibility of the countermeasure. For example, “x1> 2” (“x1 = C or D”) is selected as the first factor. To do. From FIG. 11 (or Table 46), this countermeasure can eliminate three defects out of the four defects without dropping only two non-defective products (defective product separation = 0.6). Can be expected).

  As for the content to be secondly countermeasured, the first factor (“x1> 2”, ie, “x1 = C or D”) and other factors are used using the composite factor table (Table 47). This can be determined by examining the degree of compounding. FIG. 12 shows the number of defects due to a composite factor with the first factor (“x1> 2”, ie, “x1 = C or D”) in the bar graph (number of defects) of each factor (input attribute condition) in FIG. It is shown with hatching. From FIG. 12, the condition “x2> 2”, that is, “x2 = c or d” has a defect that does not overlap with the first factor (“x1> 2”, ie, “x1 = C or D”), In addition, since the defective product separation degree (second data group separation degree) is high, there is a high possibility of an independent factor with respect to the first factor, and it can be read that this is an item to be secondly addressed.

  In FIG. 12, it is also possible to extract a composite factor (or a dependent factor) with the first factor (“x1> 2”, ie, “x1 = C or D”). In this example, “X3> 2” having a large hatching ratio (all bars are hatched) is extracted.

  As described above, in the present embodiment, the input attribute condition determination unit 11 determines the input attribute condition that causes a problem event (defective product second data group DA2) for all input attributes (steps). 8). Then, the factor determination unit 17 sets the plurality of input attribute conditions for the same input attribute determined for each repetition process of step 8 based on a clear index of defective product separation (second data group separation). On the other hand, among these, a condition with high priority is selected. For this reason, even if the above-described repetitive processing is performed to obtain a detailed factor analysis result, it does not become complicated, and it is in a very simple form, but with high accuracy and a cause of failure (corresponding to the second data group). The factor of the output attribute condition) can be determined.

  Further, the rank determination unit 13 performs the input attribute conditions (the input attribute conditions in the determination factor list table (Table 43 (a) or Table 43 (b)) determined by the factor determination unit 17 as the cause of the problem event. Since prioritization is performed based on clear indicators such as the first scoring point, the second scoring point, and the total score of the first and second scoring points, the priority of the factors that should be addressed The order can be clearly determined.

For this reason, a plurality of failure factors are extracted in the form of a hierarchical structure (product characteristic failures with y = Y appear everywhere in the decision tree, and the same input attribute has many branches under different conditions. The problem of the conventional decision tree (FIG. 14) that “it is difficult to determine which product attribute is bad because each input attribute is in a range of values” is solved. For example, the condition of product characteristic failure y = Y extracted in the conventional decision tree is
・ "X2 = c or d"
Or
“X2 = a or b” and “x1 = C or D” and “x3 ≧ 2.5”
On the other hand, the product characteristic defect condition (input attribute condition) determined in this embodiment is as shown in FIGS. 9 to 11 and Table 46 (Table 48).
・ “X1> 2” (C or D)
"X2>2" (c or d)
・ “X3> 2”
・ “X4 ≦ 10”
It is. That is, since the input attributes x1 to x4 each have a range of values, are the product characteristics bad? It is possible to make a quick determination by looking at the causes of product characteristic defects, and the first object of the present invention can be achieved.

  Furthermore, in the conventional decision tree (FIG. 14), it was difficult to understand what priority should be taken for each failure factor appearing in various layers, but in this embodiment, Since the order determination unit 13 performs prioritization based on clear indicators such as the first score, the second score, and the total score of the first score and the second score, measures are taken. The priority order of factors should be clearly determined. For example, “x1> 2” (“x1 = C or D”) or “x2” depending on the ranking result (FIG. 11) of the ranking determining unit 13 based on the total score of the first and second score points. > 2 ”(“ x2 = c or d ”) can be determined as a condition to be first addressed. That is, the second object of the present invention can be achieved.

  In the above-described embodiment, the decision tree (FIG. 8) is generated by a plurality of branches (repetition). However, if only one branch is sufficient, steps 10 to 12 are performed after step 9 is completed. Is skipped and the process of step 13 may be performed. In this case, the output (determination factor list table) of the factor determination unit 17 in step 13 is the same as the output (table 35) of the defective product data calculation unit 12 in step 9.

  In the above description, in step 12, the analysis data group extraction unit 6 extracts only the other data group from the divided data groups as the next analysis data group, but also extracts the factor data group as the analysis data group. The series of processing from step 3 to step 11 may be repeated. What is necessary is just to extract at least one of a factor data group and another data group as a new analysis data group.

  In the above description, the second data group DA2 is used as a defective product data group, and a data analysis example for extracting the cause of the failure has been described. However, the second data group DA2 is used as a good data group, and conditions for obtaining a good product are shown. It is good also as data analysis which extracts.

  The data analysis method described above can be realized by the computer executing a data analysis program including processes corresponding to S0 to S17 (steps 0 to 17) in FIG. Therefore, in the data analysis apparatus of FIG. 1, the data analysis program includes a computer, a character-numeric data conversion unit 2, a classification condition setting unit 3, a data classification unit 4, an analysis data group extraction unit 6, a data row separation unit 7, Data string extraction unit 8, frequency calculation unit 9, frequency cumulative difference calculation unit 10, input attribute condition determination unit 11, defective product data calculation unit 12, frequency cumulative ratio calculation unit 14, data division unit 15, end condition determination unit 16, This can be realized by functioning as the factor determination unit 17, the rank determination unit 13, the complex factor defect number calculation unit 18, and the numerical value-character data conversion unit 19.

  The data analysis program can be stored in a computer-readable recording medium and provided to the user. Thereby, the data analysis program can be easily provided to the computer. The recording medium may be a built-in medium built in the computer main body, or a removable medium configured to be separable from the computer main body. Examples of the built-in medium include ROM; rewritable nonvolatile memory such as flash memory; and hard disk. The removable media includes optical recording media such as CD-ROM and DVD; magneto-optical recording media such as MO; magnetic recording media such as floppy (registered trademark) disks, cassette tapes and removable hard disks; memory cards and the like. Examples of the medium include a rewritable non-volatile memory such as a medium; a medium including a ROM such as a ROM cassette.

  The program may be configured to be executed by CPU access, or after reading the program stored in the recording medium and transferring the read program to the program storage area of the built-in medium, the program on the built-in medium May be executed by CPU access. Further, the program is not limited to a program that is sold in a state of being stored in a computer-readable recording medium, but is sold in a format that is transferred to a user's computer via a communication network such as the Internet. It may be a thing.

[Modification 1]
In step 13 of the first embodiment, the factor determination unit 17 uses the defect separation degree as an index, and among the plurality of input attribute conditions (Table 42) for the same input attribute, the cause of the problem event (defective product A condition having a high degree of importance (priority) is selected as a factor that causes an output attribute condition corresponding to the second data group DA2. As described above, a defective product separation degree (second data group separation degree) is suitable as this index, but even if the following index is used, an effect similar to that of the first embodiment can be obtained. it can.

  That is, the factor determination unit 17 includes a first scoring that is given so that the score increases in descending order of the number of defects (corresponding to the second data group) among a plurality of input attribute conditions in the same input attribute, Using the total score with the second score given so that the score increases in descending order of the defective product separation degree (second data group separation degree), the input attribute condition that maximizes the total point is assigned the priority. You may make it select as conditions with high (importance as a factor of a problem event). However, there is a possibility that the input attribute condition of a certain input attribute selected according to this modification is strongly influenced by disturbance (the influence of other input attributes) as compared with the first embodiment. Therefore, the cause of the problem event may not be accurately extracted due to the influence of the disturbance.

  The factor determining unit 17 selects the input attribute condition that maximizes the number of defects (corresponding to the second data group) among the plurality of input attribute conditions for the same input attribute as the priority (the factor of the problem event). You may make it select as a condition with high importance. However, there is a possibility that the input attribute condition selected in this way is more strongly affected by disturbances (effects of other input attributes) than the first modification. Therefore, there is a higher possibility that the cause of the problem event cannot be accurately extracted due to the influence of the disturbance.

[Modification 2]
In step 7 of the first embodiment, for each input attribute condition related to the input attribute of the analysis data group, “if the input attribute satisfies the input attribute condition, the data belongs to the second data group, and the input attribute The xj frequency cumulative difference was calculated as an input attribute condition evaluation index representing the certainty of the first association rule that the data belongs to the first data group if the input attribute condition is not satisfied.

  As described above, it is preferable to use the xj frequency cumulative difference from the viewpoint of reducing the calculation load and improving the calculation speed. However, if the calculation load and the calculation speed are not considered as problems, the Gini index method described in Non-Patent Document 1 is used. May be used as an input attribute condition evaluation index.

  In this case, as the first evaluation means for calculating the input attribute condition evaluation index, the data row separation unit 7, the data string extraction unit 8, the frequency calculation unit 9, the frequency cumulative difference in the data analysis apparatus of the first embodiment. Instead of the calculation unit 10, as shown in FIG. 13, a Gini improvement degree calculation unit 101 and a polarity determination unit 102 may be provided. Further, the character-numeric data conversion unit 2 and the numeric-character data conversion unit 19 are unnecessary.

  The Gini improvement degree calculation unit 101 calculates the improvement degree similar to the equation (2) (FIGS. 15 to 18) for each of all branch conditions that can be taken by the input attribute of the analysis data group.

  In addition, when the node is divided into two child nodes, the polarity determination unit 102 determines which node corresponds to the second data group, and determines only the condition corresponding to the second data group at this time. Together with the degree of improvement (input attribute condition evaluation index), it is sent to the input attribute condition determination unit 11. Taking “branch between x1 = A, B and x1 = C, D” in FIG. 15 as an example, “x1 = A, B” corresponding to the first data group is not sent to the input attribute condition determination unit 11. Only “x1 = C, D” corresponding to the second data group is sent to the input attribute condition determination unit 11 together with the improvement degree (input attribute condition evaluation index) at this time. Note that the condition corresponding to the second data group determined by the polarity determination unit 102 may be a condition of a node having a higher probability of y = Y.

In the data analysis method according to the present modification, the steps 0 and 16 in the first embodiment are omitted, and the steps 4 to 7 are performed using the Gini improvement degree calculation unit 101 and the polarity determination unit. The second embodiment is the same as the first embodiment except that the processing is changed to 102. In step 8 of the present modification, the input attribute condition determination unit 11 determines the input attribute condition having the maximum improvement (input attribute condition evaluation index) for each input attribute xj as the first correlation rule. It is determined as an input attribute condition to be satisfied. In Step 10 of this modification, the frequency cumulative ratio calculating unit 14 sets 1-xj frequency cumulative% to the input attribute threshold value xj-th in each input attribute condition determined by the input attribute condition determining unit 11. 2-xj frequency cumulative% is calculated, and the frequency cumulative lower ratio or frequency cumulative upper ratio (division rule evaluation value) is calculated using the calculation result.
Although this modification is inferior to the first embodiment in calculation load and calculation speed, there is an advantage that it is not necessary to perform processing (step 0) for converting character data in the basic data group into numerical data.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  The present invention relates to an output attribute (object attribute) to be analyzed, for example, characteristics of a product manufactured in a manufacturing process, and an input attribute (description attribute) that is an attribute affecting the output attribute, for example, a manufacturing process condition, etc. It can be applied to the analysis of the causal relationship of, for example, it can be used for improving the manufacturing process.

It is a block diagram which shows the structure of the data analyzer which concerns on one Embodiment of this invention. It is a flowchart which shows the data analysis method which concerns on one Embodiment of this invention. FIG. 6 is a graph showing an example of the output of the frequency cumulative difference calculation unit 10 (step 7) in the data analysis apparatus according to the embodiment of the present invention, and the value of the input attribute x1 and the non-defective 1-x1 frequency cumulative% ( A) shows the relationship between 2-x1 frequency cumulative% (B) of defective products, first x1 frequency cumulative difference (BA), and second x1 frequency cumulative difference (AB). The graph of an example of the output of the frequency accumulation difference calculating part 10 (step 7) in the data analyzer which concerns on one Embodiment of this invention is represented with the value of the input attribute x2, and 1-x2 frequency accumulation% (non-defective product). A) shows the relationship between 2-x2 frequency cumulative percentage (B) of defective products, first x2 frequency cumulative difference (BA), and second x2 frequency cumulative difference (AB). FIG. 6 is a graph showing an example of the output of the frequency cumulative difference calculation unit 10 (step 7) in the data analysis apparatus according to the embodiment of the present invention, and the value of the input attribute x3 and the non-defective 1-x3 frequency cumulative% ( A) shows the relationship between 2-x3 frequency cumulative% (B) of defective products, first x3 frequency cumulative difference (BA), and second x3 frequency cumulative difference (AB). FIG. 6 is a graph showing an example of the output of the frequency cumulative difference calculation unit 10 (step 7) in the data analysis apparatus according to the embodiment of the present invention, and the value of the input attribute x4 and the non-defective 1-x4 frequency cumulative% ( A) shows the relationship between 2-x4 frequency cumulative% (B) of defective products, first x4 frequency cumulative difference (BA), and second x4 frequency cumulative difference (AB). (A)-(d) is the figure which expressed an example (Table 35) of the data output by the inferior goods data calculating part 12 (step 9) in the data analyzer which concerns on one Embodiment of this invention with the Venn diagram. is there. It is the figure which expressed the process of the input attribute condition determination (step 8) and the factor determination (step 13) in the data analysis method which concerns on one Embodiment of this invention in the form of a decision tree. The ranking determination unit 13 (step 14) in the data analysis apparatus according to the embodiment of the present invention rearranges the determination factor list table (Table 46) in descending order of the first score and displays the result as a factor breakdown Pareto chart. FIG. 5 is a diagram in which the number of defects (second data group hit count) for each input attribute condition is represented by a bar graph, and the defective product separation degree (second data group separation degree) is represented by a line graph. The ranking determination unit 13 (step 14) in the data analysis apparatus according to the embodiment of the present invention rearranges the determination factor list table (Table 46) in descending order of the second score and displays the result as a factor breakdown Pareto chart. FIG. 5 is a diagram in which the number of defects (second data group hit count) for each input attribute condition is represented by a bar graph, and the defective product separation degree (second data group separation degree) is represented by a line graph. The rank determination unit 13 (step 14) in the data analysis apparatus according to the embodiment of the present invention rearranges the determination factor list table (Table 46) in descending order of the total score of the first and second score points. The result is displayed as a factor breakdown Pareto chart, and the number of defects for each input attribute condition (number of corresponding second data groups) is represented by a bar graph and the degree of defective product separation (second data group separation degree) is represented by a line graph. is there. Using the example (Table 47) of data output from the composite factor defect number calculation unit 18 (step 15) in the data analysis apparatus according to the embodiment of the present invention, the number of defects for each input attribute condition (corresponding to the second data group) (Number) is a bar graph and defective product separation degree (second data group separation degree) is a line graph, which is a combination of the first factor (“x1> 2”, ie, “x1 = C or D”) The number of defects due to the factor is shown with hatching. It is a block diagram which shows the structure of the data analyzer which concerns on the modification 2 of this invention. It is a figure showing the conventional decision tree. It is a graph explaining the Gini index method which is the prior art described in Non-Patent Document 1, and shows the relationship between the branch condition of the input attribute x1 and the improvement degree of the Gini index method using the data group of Table 1 as the subject. is there. It is a graph explaining the Gini index method which is the prior art described in Non-Patent Document 1, and shows the relationship between the branch condition of the input attribute x2 and the improvement degree of the Gini index method using the data group of Table 1 as the subject. is there. It is a graph explaining the Gini index method which is the prior art described in Non-Patent Document 1, and shows the relationship between the branch condition of the input attribute x3 and the improvement degree of the Gini index method using the data group of Table 1 as a subject. is there. It is a graph explaining the Gini index method which is the prior art described in Non-Patent Document 1, and shows the relationship between the branch condition of the input attribute x4 and the improvement degree of the Gini index method using the data group of Table 1 as the subject. is there.

Explanation of symbols

1 basic data group storage unit 2 character-numeric data conversion unit 3 classification condition setting unit (classification condition setting means)
4 Data classification part (classification means)
5 Basic data group storage after classification 6 Analysis data group extraction unit (analysis data group extraction means)
7 Data row separator (first evaluation means)
8 Data string extraction unit (first evaluation means)
9 Frequency calculator (first evaluation means, frequency calculation means)
10 Frequency cumulative difference calculation unit (first evaluation means, difference calculation means)
11 Input attribute condition determining unit (input attribute condition determining means)
12 Defective product data calculation section 13 Rank determination section (rank determination means)
14 Frequency cumulative ratio calculation unit (second evaluation means)
15 Data division unit (division means)
16 End condition determination unit (end condition determination means)
17 Factor determination unit (factor determination means)
18 Compound factor failure number calculation unit 19 Numerical value-character data conversion unit 20 Analysis result data storage unit 21 Output unit 101 Gini improvement degree calculation unit 102 Polarity determination unit

Claims (17)

Data analysis for analyzing a causal relationship between the input attribute and the output attribute and extracting information indicating the causal relationship with respect to a basic data group that is a set of data composed of a plurality of input attributes and output attributes A device,
Classifying means for classifying the basic data group into a first data group and a second data group according to the value of the output attribute, and assigning a classification flag;
Analysis data group extraction means for extracting an analysis data group to be analyzed from the basic data group after classification by the classification means;
For each of the input attribute conditions related to the input attribute of each of the analysis data groups, “the first data group and the second data group by dividing the analysis data group according to the input attribute condition and its exclusive condition; If the input attribute satisfies the input attribute condition, the data belongs to the second data group in the analysis data group. If the input attribute does not satisfy the input attribute condition, the analysis data group A first evaluation means for calculating an input attribute condition evaluation index, which represents the probability of the first association rule that the data belongs to the first data group in
Input attribute condition determining means for determining an input attribute condition having the maximum input attribute condition evaluation index as an input attribute condition satisfying the first correlation rule for each of the input attributes of the analysis data group;
The number of data satisfying the input attribute condition in the basic data group and corresponding to the second data group with respect to at least a part of the plurality of input attribute conditions determined by the input attribute condition determining means And at least a second data group separation degree representing a ratio of the number of data corresponding to the second data group in data satisfying the input attribute condition in the basic data group A data analysis apparatus comprising: a rank determining unit that ranks according to importance as a factor of an output attribute condition corresponding to the second data group using one as an index. For each of the input attribute conditions determined by the input attribute condition determining means, “if the input attribute satisfies the input attribute condition, the data is included in the second data group in the analysis data group”. Second evaluation means for calculating a division rule evaluation value representing the probability of the second association rule;
Among the input attribute conditions determined by the input attribute condition determining means, based on the input attribute condition having the largest division rule evaluation value, the analysis data group is a factor data group that satisfies the input attribute condition Dividing means for dividing the data into other data groups that do not satisfy the input attribute condition;
Factor determining means for selecting an input attribute condition having high importance as a factor of the output attribute condition corresponding to the second data group from the input attribute conditions determined by the input attribute condition determining means; Including
The analysis data group extraction means extracts at least one of the data groups divided by the division means as a new analysis data group,
A series of processes consisting of processing by the analysis data group extraction means, processing by the first evaluation means, processing by the input attribute condition determination means, processing by the second evaluation means, and processing by the dividing means are repeated. To be executed,
The input attribute condition determining means is configured to determine a plurality of input attribute conditions for each of the input attributes by the above iterative process,
The factor determination means is a factor of an output attribute condition corresponding to the second data group from among a plurality of input attribute conditions determined by the input attribute condition determination means for each of the input attributes. Select input attribute conditions with high importance,
The rank determining means ranks the plurality of input attribute conditions selected by the factor determining means, which are part of the input attribute conditions determined by the input attribute condition determining means. The data analysis apparatus according to claim 1, wherein: The plurality of input attribute conditions determined by the input attribute condition determining unit includes a first pattern that “the input attribute is equal to or less than a threshold value” and a second pattern that “the input attribute exceeds the threshold value”.
The factor determining means includes
Among the plurality of input attribute conditions in the same input attribute, the input attribute condition having the high importance level in the first pattern and the input attribute condition having the high importance level in the second pattern are selected. The data analysis apparatus according to claim 2, wherein   The factor determining means selects an input attribute condition that maximizes the second data group separation degree among the plurality of input attribute conditions to be selected as the highly important input attribute condition. The data analysis apparatus according to claim 2, wherein the data analysis apparatus is provided.   The factor determination means includes a first scoring system that gives a higher score in descending order of the number of hits in the second data group among a plurality of input attribute conditions to be selected, and the second data The input attribute condition that maximizes the total score with the second score given so that the score becomes higher in descending order of the group separation degree is selected as the input attribute condition having the highest importance. The data analysis apparatus according to claim 2 or 3. The rank determining means includes
6. The ranking is performed on at least a part of the plurality of input attribute conditions determined by the input attribute condition determining means in descending order of the number of corresponding second data groups. The data analysis apparatus according to any one of the above. The rank determining means includes
The ranking is performed in order of descending second data group separation degree on at least a part of the plurality of input attribute conditions determined by the input attribute condition determining means. The data analysis device according to any one of 5. The rank determining means includes
A first scoring system that gives a higher score in descending order of the second data group hit count for at least some of the plurality of input attribute conditions determined by the input attribute condition determining means; 6. The ranking is performed according to any one of claims 1 to 5, wherein the ranking is performed in descending order of the total score with the second allocation point given so that the score increases in descending order of the second data group separation degree. Data analysis equipment.   The data analysis according to claim 2, wherein the analysis data group extraction unit extracts only the other data group from the data group divided by the division unit as a new analysis data group. apparatus. It further includes an end condition determining means for determining whether or not the end condition is satisfied,
When it is determined that the end condition is satisfied by the end condition determining means, the execution of the series of processes ends.
The end condition determination unit is configured to determine whether the number of data of the second data group in the analysis data group extracted by the analysis data group extraction unit is 0 as an end condition. Item 10. The data analysis device according to Item 9. The input attributes in the analysis data group are all numeric attributes,
The first evaluation means includes
For all the numerical values of each said input attribute,
In the first data group of the analysis data group, the ratio of the number of data whose input attribute is less than or equal to the numerical value is calculated as a first frequency,
In the second data group of the analysis data group, a frequency calculation means for calculating, as a second frequency, a ratio of the number of data whose input attribute is less than or equal to the numerical value;
For all the numerical values of each said input attribute,
The data analysis apparatus according to claim 1, further comprising a difference calculation unit that calculates a difference between the first frequency and the second frequency as the input attribute condition evaluation index. The second evaluation means includes
For each of the input attribute conditions determined by the input attribute condition determination means,
The ratio of the ratio of the number of data satisfying the input attribute condition in the second data group of the analysis data group to the ratio of the number of data satisfying the input attribute condition in the first data group of the analysis data group, The data analysis apparatus according to claim 2, wherein the data analysis apparatus calculates the division rule evaluation value. It further includes a classification condition setting means for setting a classification condition,
The data analysis apparatus according to claim 1 or 2, wherein the classification unit classifies the basic data group based on the value of the output attribute and the classification condition. The causal relationship between the input attribute and the output attribute is analyzed with respect to a basic data group that is a set of data composed of a plurality of input attributes and output attributes, using the data analysis device according to claim 1. And a data analysis method for extracting information indicating the causal relationship,
A classification step of classifying the basic data group into a first data group and a second data group according to the value of the output attribute, and assigning a classification flag by the classification means;
An analysis data group extraction step of extracting an analysis data group to be analyzed from the basic data group after classification by the classification means by the analysis data group extraction means;
For each of the input attribute conditions related to each of the input attributes of the analysis data group, the first evaluation means determines that the first data is separated by dividing the analysis data group by the input attribute condition and its exclusive condition. Group and the second data group, and if the input attribute satisfies the input attribute condition, the data belongs to the second data group in the analysis data group, and the input attribute satisfies the input attribute condition A first evaluation step for calculating an input attribute condition evaluation index that represents the probability of the first association rule that if it does not satisfy, the data belongs to the first data group in the analysis data group;
The input attribute condition determining means determines an input attribute condition having the maximum input attribute condition evaluation index as an input attribute condition satisfying the first correlation rule for each of the input attributes of the analysis data group. An input attribute condition determination step to be performed;
The order determination means satisfies the input attribute condition in the basic data group for at least a part of the plurality of input attribute conditions determined by the input attribute condition determination means, and the second data group A second data group corresponding number representing the number of data corresponding to the second data group, and a second data group representing a ratio of the number of data corresponding to the second data group among the data satisfying the input attribute condition in the basic data group And a ranking determination step of ranking by importance as an output attribute condition factor corresponding to the second data group using at least one of the data group separation degrees as an index. The input attribute is a manufacturing process condition and / or an in-line inspection result in a product manufacturing process,
The output attribute is a product quality determination result,
The data analysis method according to claim 14, wherein the second data group is a data group with a poor quality determination result. Data analysis for analyzing a causal relationship between the input attribute and the output attribute and extracting information indicating the causal relationship with respect to a basic data group that is a set of data composed of a plurality of input attributes and output attributes A program,
Computer
Classifying means for classifying the basic data group into a first data group and a second data group according to the value of the output attribute, and adding a classification flag;
Analysis data group extraction means for extracting an analysis data group to be analyzed from the basic data group after classification by the classification means;
For each of the input attribute conditions related to the input attribute of each of the analysis data groups, “the first data group and the second data group by dividing the analysis data group according to the input attribute condition and its exclusive condition; If the input attribute satisfies the input attribute condition, the data belongs to the second data group in the analysis data group. If the input attribute does not satisfy the input attribute condition, the analysis data group A first evaluation means for calculating an input attribute condition evaluation index that represents the probability of the first association rule that the data belongs to the first data group
Input attribute condition determining means for determining an input attribute condition having the maximum input attribute condition evaluation index as an input attribute condition satisfying the first correlation rule for each of the input attributes of the analysis data group;
The number of data satisfying the input attribute condition in the basic data group and corresponding to the second data group with respect to at least a part of the plurality of input attribute conditions determined by the input attribute condition determining means And at least a second data group separation degree representing a ratio of the number of data corresponding to the second data group in data satisfying the input attribute condition in the basic data group A data analysis program for functioning as a rank determining unit that ranks according to importance as a factor of an output attribute condition corresponding to the second data group using one as an index.   The computer-readable recording medium which recorded the data analysis program of Claim 16.

Images