JP5772437B2 - Data structure extraction program and data structure extraction device - Google Patents

Description

  The present invention relates to a data structure extraction program and a data structure extraction device.

  As a conventional method of extracting and outputting a data structure, for example, a tree structure that frequently appears using a pattern extraction technique is extracted from a plurality of data having a tree structure as a sample data structure, and the extracted data structure is similar to the sample data structure. There is one that calculates the degree (for example, see Non-Patent Documents 1 and 2). Using these methods, it is possible to output a data structure having a similarity equal to or higher than a predetermined threshold.

  Non-Patent Document 1 compares the arrangement of leaf nodes of each tree structure when comparing between tree structures composed of an internal node that branches from a root node as a vertex and a leaf node at the bottom. Disclosed is a method for calculating a minimum number of procedures (hereinafter referred to as “edit distance”) such as insertion, deletion, and replacement of data required for transforming from one leaf node sequence to another leaf node sequence. . Using this edit distance, tree structures having an edit distance smaller than a predetermined value are set as similar tree structures.

  Non-Patent Document 2 includes not only the leaf nodes of a tree structure but also the insertion and deletion of data required for transformation from one tree structure to the other, including root nodes and internal nodes between tree structures. , A method of calculating the minimum number of procedures such as replacement (hereinafter referred to as “Tree Edit distance”) is disclosed. This tree edit distance calculation method increases the amount of calculation compared to the method of Non-Patent Document 1 described above, but calculates a more reliable value as the similarity between tree structures in order to consider other than leaf nodes.

Gusfield, Dan (1997). Algorithms on strings, trees, and sequences: computer science and computational biology. Cambridge, UK: Cambridge University Press. ISBN0-521-58519-8. Philip, Bille.A survey on tree edit distance and related problems.Journal Theoretical Computer Science, Volume 337 Issue 1-3, 9 June 2005.Elsevier Science Publishers Ltd. Essex, UK

  An object of the present invention is to use a frequency of common data and data other than the common data among a plurality of appearing data structures, determine a similar one from the data structure, and output the data structure extraction program It is to provide a data structure extraction device.

  In order to achieve the above object, an aspect of the present invention provides the following data structure extraction program and data structure extraction device.

[1]
Extracting means for extracting a plurality of types of patterns appearing at a frequency equal to or higher than a predetermined threshold from a plurality of supports ;
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is support for a pattern having mutually common to all sets of support similar, support for the pattern in which the with each other is similar If the set of no support other than support for the common support in the set does not exist, and setting means for setting the number of said common support as a second threshold value,
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is a support which is common to all sets of support patterns corresponding with each other are similar, the pattern in which the with each other is similar If a supported set of no support other than support for the common support in the set is present, it is determined that the pattern pattern to be output of a set of support without the support of non-support for the common output Judgment means,
Functions as an output means for outputting a pattern in which the output determination unit determines that a pattern to be output, a pattern in which the setting means is extracted by the extraction means with a frequency of more than the second threshold value from the support that targets Data structure extraction program

[2] Extraction means for extracting a plurality of types of patterns appearing at a frequency equal to or higher than a predetermined threshold from a plurality of supports ;
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is support for a pattern having mutually common to all sets of support similar, support for the pattern in which the with each other is similar If the set of no support other than support for the common support in the set does not exist, and setting means for setting the number of said common support as a second threshold value,
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is a support which is common to all sets of support patterns corresponding with each other are similar, the pattern in which the with each other is similar If a supported set of no support other than support for the common support in the set is present, it is determined that the pattern pattern to be output of a set of support without the support of non-support for the common output Judgment means,
A pattern in which the output determination unit determines that a pattern to be output, and output means said setting means outputs a pattern that is extracted by the extraction means with a frequency of more than the second threshold value from the support that targets Data structure extraction device having.

  According to the first or second aspect of the present invention, it is possible to accurately extract a data structure having similar characteristics from a plurality of appearing data structures as compared with the case where this configuration is not adopted.

FIG. 1 is a diagram showing an example of the configuration of a data structure extraction apparatus according to an embodiment of the present invention. FIG. 2 is a schematic diagram illustrating an example of a configuration of a DPC data pattern. FIG. 3 is a schematic diagram illustrating an example of a configuration of a pattern extracted from DPC data. FIG. 4 is a flowchart showing the pattern extraction operation of the data structure extraction apparatus. FIG. 5 is a schematic diagram illustrating an example of the configuration of a support vector. FIG. 6 is a diagram for explaining the support vector classification operation. FIGS. 7A to 7C are schematic diagrams illustrating an example of the table mapped by the support vector classification unit and the output contents output by the output unit 107. FIG. 8A to 8C are schematic diagrams illustrating an example of the table mapped by the support vector classification unit and the output contents output by the output unit 107. FIG. FIG. 9 is a flowchart illustrating an operation example of the data structure extraction apparatus. FIG. 10 is a flowchart showing an example of the pattern output operation of the data structure extraction apparatus. FIGS. 11A and 11B are schematic diagrams illustrating an example of patterns extracted by the data structure extraction device.

(Configuration of data structure extraction device)
FIG. 1 is a diagram showing an example of the configuration of a data structure extraction apparatus according to an embodiment of the present invention.

  The data structure extraction device 1 is configured by a CPU and the like, and controls each unit and executes various programs, and a storage unit configured by a storage medium such as an HDD (Hard Disk Drive) or a flash memory and stores information. 11, for example, used for analyzing patient clinical information and electronic data of medical practice.

  The control unit 10 executes a data structure extraction program 110 to be described later, whereby a pattern extraction unit 100, a support vector generation unit 101, a support vector classification unit 102, a common support calculation unit 103, a threshold setting unit 104, a pattern re-extraction unit 105, functions as the output judging means 106, the output means 107, and the like.

  The pattern extraction unit 100 is an example of an extraction unit, and has a data structure from a plurality of comparison target data (hereinafter referred to as “support”) included in the DPC data 111 described later from the storage unit 11 as electronic data to be analyzed. As an example, a plurality of sample tree structures (hereinafter referred to as “patterns”) having a patient name as a root node as a target are acquired. Here, a pattern refers to a pattern extracted as a common tree structure that appears at a predetermined frequency (first threshold) in a plurality of data having a tree structure using a pattern extraction technique.

  The pattern extraction unit 100 extracts a common E file and F file, which will be described later, as a pattern among a plurality of supports by using a known pattern extraction technique. As a well-known pattern extraction technique, for example, Prefix Span, BIDE, CloSpan or the like of sequential pattern mining or sub-tree mining can be used.

  The support vector generation unit 101 vectorizes a support set for each pattern extracted by the pattern extraction unit 100, and generates a support vector to be described later.

  The support vector classification unit 102 calculates a correlation coefficient between the support vectors generated by the support vector generation unit 101, and classifies the support vectors into a plurality of groups based on the correlation function.

  The common support calculation unit 103 calculates the number of supports common among the patterns belonging to the group classified by the support vector classification unit 102.

  The threshold setting means 104 is included in the common support set when there is a common support set among patterns belonging to the same support vector group and there is another support other than the common support for each pattern. The number of supports is set as the second threshold.

  The pattern re-extraction unit 105 is another example of the extraction unit, and extracts a pattern that the pattern extraction unit 100 did not extract based on the second threshold set by the threshold setting unit 104. Instead of using the pattern re-extraction unit 105, the pattern extraction unit 100 may extract a pattern based on the second threshold value.

  If there is a set of supports common to patterns belonging to the same support vector group and there is another support other than the support common to one pattern, the output determination means 106 does not output that one pattern. to decide.

  The output unit 107 deletes the pattern determined not to be output by the output determination unit 106 from the pattern extracted by the pattern extraction unit 100, adds the pattern extracted by the pattern re-extraction unit, and outputs the result as a pattern extraction result.

  The storage unit 11 stores a data structure extraction program 110, DPC data 111, threshold information 112, and the like that cause the control unit 10 to operate as each unit described above.

  The DPC data 111 is an electronic data set of patient clinical information and medical practice in a nationally unified format that can be analyzed. Patient clinical information includes, for example, patient basic information, disease name, technique, various score / stage classifications, etc., and medical practice information includes medical practice, pharmaceuticals, medical materials, implementation date, frequency / quantity, clinical department, ward And insurance type.

  The DPC data 111 includes data called format 1, E file, and F file as basic data. Form 1 includes patient clinical information, names of wounds, surgical procedures, adjuvant treatment, and the like. The E file is information such as the implementation date, number of times, clinical department, ward, order doctor, and the like. The F file is a detailed content of the E file, and is information such as an action, a medicine, a material, and a quantity.

  In the present embodiment, a tree structure including a patient as a root node, date / time data belonging to the patient and an E file as an internal node, and an F file belonging to the E file as a leaf node is supported. A support data structure that appears at a frequency higher than the specified frequency is acquired as a pattern, a similarity is calculated between the acquired patterns, and a set of a plurality of patterns is extracted based on the calculated similarity.

  The threshold information 112 stores a predetermined first threshold used by the pattern extraction unit 100 and a second threshold set by the threshold setting unit 104.

  FIG. 2 is a schematic diagram illustrating an example of a configuration for supporting the DPC data 111 from which the pattern extraction unit 100 obtains a pattern.

  The plurality of supports 200a, 200b,... Acquired from the DPC data 111 includes date / time data 22 belonging to a patient, an E file 21 belonging to the date / time data 22, and an F file belonging to the E file 21, and constitutes a tree structure. To do.

  FIG. 3 is a schematic diagram showing an example of a configuration of a pattern extracted from the DPC data 111. As shown in FIG.

  The patterns 2a and 2b extracted from the DPC data 111 include date / time data 22 belonging to a patient, an E file 21 belonging to the date / time data 22, and an F file belonging to the E file 21, and constitute a tree structure.

(Operation of data structure extraction device)
Hereinafter, the operation example of the data structure extraction apparatus will be described by dividing into (1) pattern extraction operation, (2) support classification operation, and (3) pattern output operation, with reference to each drawing.

  FIG. 9 is a flowchart illustrating an operation example of the data structure extraction apparatus.

(1) Pattern Extraction Operation First, the pattern extraction unit 100 acquires a plurality of supports that are targets for pattern extraction from the DPC data 111 in the storage unit 11 (S1).

  FIG. 4 is a flowchart showing the pattern extraction operation of the data structure extraction apparatus.

  Hereinafter, in order to simplify the description, a case where six supports 200A to 200F that are simply displayed as shown in FIG. 4 are acquired will be described.

Pattern extraction means 100 extracts from the support 200A~200D patterns _{2 1} containing E files 21a and 21b, support 200B, extracts a pattern _{2 2} containing E files 21c and F files 20c from 200E and 200F.

(2) support classification Operation Next, support vector generating unit 101 generates a set of support by vectorizing the support vector for each of the patterns 2 ₁ and 2 ₂ a pattern extracting means 100 is extracted (S4).

  FIG. 5 is a schematic diagram illustrating an example of the configuration of a support vector.

Support vector SV1 is a support vector for pattern _{2 1,} support 200A, the component of 200B ... sequentially with vectors of are described. For example, it supports 200A is vector component for including a pattern _{2 1} is "1", supports 200F vector component contains no pattern _{2 1} is "0". The support vector SV2 is described in the same manner.

  Next, the support vector classification unit 102 calculates an inner product as a correlation coefficient between the support vectors (S4). The inner product of the support vectors SV1 and SV2 is 1 / √6.

  Next, the support vector classification unit 102 clusters (classifies) the support vectors as described below (S5). In the example described below, it is assumed that seven supports are extracted in step S1, support vectors SV1 to SV7 are generated for each of the seven supports, and inner products between the support vectors SV1 to SV7 are calculated.

  FIG. 6 is a diagram for explaining the support vector classification operation.

  As shown in FIG. 6, the support vector classification means 102 generates a matrix with each inner product. A large inner product value indicates that the patterns corresponding to the support vectors are similar. Therefore, the support vector classifying unit 102 clusters the support vectors according to the value of the inner product, and classifies the support vectors into cluster 1 of support vectors SV1 to SV5 and cluster 2 of support vectors SV6 and SV7.

  Next, the support vector classification unit 102 maps the classified support groups on the table with patterns. The mapping results are shown below.

(3) Pattern Output Operation FIG. 10 is a flowchart showing an example of the pattern output operation of the data structure extraction device. FIGS. 7A to 7C are schematic diagrams showing examples of the table mapped by the support vector classification unit and the output contents output by the output unit 107.

  First, the common support calculation unit 103 calculates the number of common supports between clustered support groups (S11). In the example mapped as shown in FIG. 7 (a), the support of “support group 1” is schematically shown as shown in FIG. 7 (b). The number, here “8”, is calculated as the common support number.

  Next, when each pattern includes support other than the common support as in “support group 1” (S12; Yes), the threshold setting unit 104 sets the threshold information 112 using the number of common support as the second threshold ( S13).

  Next, the pattern re-extraction means 105 re-extracts “Pattern 8” corresponding to the common support of “Support Group 1” under a condition that is smaller than the first threshold and greater than the second threshold (S14).

  Next, the above operation is repeated for “support group 2” (S16). “Support group 2” is a case where each pattern includes no support other than the common support (S12; No).

  8A to 8C are schematic diagrams illustrating an example of the table mapped by the support vector classification unit and the output contents output by the output unit 107. FIG.

  In the example mapped as shown in FIG. 8A, the support of “support group 2” is schematically shown as shown in FIG. 8B. “Pattern 6” including support can be determined to be less meaningful (less important) than “Pattern 7” in which common support is dominant, so it is determined not to output “Pattern 6” (S15). .

  Next, the output unit 107 converts “pattern 1” to “pattern 5” extracted by the pattern extraction unit 100 for “support group 1” determined to include support other than common support in each pattern in step S12. In addition, “pattern 8” re-extracted by the pattern re-extraction means 105 is output as the output content 107a of “support group 1” as shown in FIG. 7C (S17).

  In addition, the output unit 107 excludes “pattern 6” that is determined not to be output by the output determination unit 106 for “support group 2” that is determined not to include support other than common support in each pattern in step S12. “Pattern 7” is output as the output content 107b of “support group 2” as shown in FIG. 8C (S17).

[Example]
FIGS. 11A and 11B are schematic diagrams illustrating an example of patterns extracted by the data structure extraction apparatus 1.

As shown in FIG. 11 (a), E file 21g, support of patients with only the pattern _{2 3} including 21h and F file 20g is "four", pattern 2 containing E file 21g, 21h and F file 20g, a 20h support "eight" of a patient having _3, E files 21g and F file 20g, support for patients with only the pattern _{2 3} including 20h consider the case where there "four".

At this time, when the first threshold value is "12", the pattern extracting means 100, pattern _{2 4} includes patterns _{2 3} and _{2 5,} the number of support patterns _{2 3} pattern _{2 3} and _{2 4} number of support the sum is "12", because the number of support patterns 2 ₃ is the sum of the number of the support pattern 2 ₅ 2 ₄ is "12", to extract the pattern 2 ₃ beauty 2 _5, the number of support " 8 "pattern 2 ₄ a is not extracted.

However, the pattern ₂₄ is a pattern to be extracted because it is a highly important pattern with more patients than the patterns 2 ₃ and 2 ₅ . Here, the threshold setting unit 104 determines that there is common support in the patterns ₂₃ and ₂₅ extracted by the pattern extraction unit 100, and sets the second threshold to “8” which is the number of common supports.

As a result, the pattern re-extraction unit 105 extracts a high pattern 2 ₄ importance to pattern extraction unit 100 has not extracted.

Next, as shown in FIG. 11 (b), E file 21g, support of patients with only the pattern _{2 3} including 21h and F file 20g is "two people", E file 21g, 21h and F file 20g, a 20h consider the case of support of the patient is present "10 people" with pattern 2 ₃ including.

At this time, when the first threshold value is "9", the pattern extracting means 100, pattern 2 ₄ includes patterns 2 _3, the number of support patterns 2 ₃ is the sum of the number of support patterns 2 ₃ and 2 ₄ Since it is “12” and the number of supports of the pattern 2 ₃ is “10”, the patterns 2 _{3 and} 2 ₄ are extracted.

However, the pattern 2 ₃ are the lower patterns with less importance the patient compared to the pattern 2 ₄ is a pattern that should not be extracted. Therefore, the output determination unit 107 determines that the pattern that should not be output pattern 2 _3.

[Other embodiments]
The present invention is not limited to the above embodiment, and various modifications can be made without departing from the spirit of the present invention. For example, the present invention is not applied only to the DPC data 111, but can be similarly applied to a set of data having an arbitrary data structure represented by a tree structure.

  In addition, the data structure extraction program 110 used in the above embodiment may be read from a storage medium such as a CD-ROM into a storage unit in the apparatus, or from a server apparatus connected to a network such as the Internet. You may download to the memory | storage part. Moreover, you may implement | achieve part or all of the means 100-107 used by the said embodiment with hardware, such as ASIC.

DESCRIPTION OF SYMBOLS 1 Data structure extraction apparatus 10 Control part 11 Storage part 20 F file 21 E file 22 Date and time data 100 Pattern extraction means 101 Support vector generation means 102 Support vector classification means 103 Common support calculation means 104 Threshold setting means 105 Pattern re-extraction means 106 Output Judgment means 107 Output means 110 Data structure extraction program 111 DPC data 112 Threshold information

Claims (2)

Computer
Extracting means for extracting a plurality of types of patterns appearing at a frequency equal to or higher than a predetermined threshold from a plurality of supports ;
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is support for a pattern having mutually common to all sets of support similar, support for the pattern in which the with each other is similar If the set of no support other than support for the common support in the set does not exist, and setting means for setting the number of said common support as a second threshold value,
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is a support which is common to all sets of support patterns corresponding with each other are similar, the pattern in which the with each other is similar If a supported set of no support other than support for the common support in the set is present, it is determined that the pattern pattern to be output of a set of support without the support of non-support for the common output Judgment means,
Functions as an output means for outputting a pattern in which the output determination unit determines that a pattern to be output, a pattern in which the setting means is extracted by the extraction means with a frequency of more than the second threshold value from the support that targets Data structure extraction program Extracting means for extracting a plurality of types of patterns appearing at a frequency equal to or higher than a predetermined threshold from a plurality of supports ;
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is support for a pattern having mutually common to all sets of support similar, support for the pattern in which the with each other is similar If the set of no support other than support for the common support in the set does not exist, and setting means for setting the number of said common support as a second threshold value,
In each set of support having a pattern extracted by the extracting unit by using the first threshold value, there is a support which is common to all sets of support patterns corresponding with each other are similar, the pattern in which the with each other is similar If a supported set of no support other than support for the common support in the set is present, it is determined that the pattern pattern to be output of a set of support without the support of non-support for the common output Judgment means,
A pattern in which the output determination unit determines that a pattern to be output, and output means said setting means outputs a pattern that is extracted by the extraction means with a frequency of more than the second threshold value from the support that targets Data structure extraction device having.

Images