JP5008623B2 - Method, apparatus, and program for classifying items according to how to proceed with processing of interest in business or work - Google Patents

Description

  The present invention relates to a technique for supporting business or work analysis by a computer using a work record in the business or work.

  In order for companies to efficiently perform various processes such as making quotations, checking inventory, and notifying delivery dates for matters such as orders from customers, internal approvals, orders to other companies, Create a business process model that expresses the rules for how to proceed, such as what processing is performed in what order, what processing and which processing is performed in parallel, and many business systems are created accordingly Is built and used.

When business is performed on an actually generated matter, the business can be performed efficiently and accurately by proceeding according to the business process model. Note that in this specification, what represents how the process has progressed for each case is referred to as a business process instance, and each execution of the process is referred to as an activity instance. In the description, in order to represent a specific process or activity instance, the process is represented as a _i , and an activity instance in a business process instance of a certain matter is represented as b _j . Here, i and j are integers of 1 or more assigned by an arbitrary method so as not to overlap each other and not to be missing. Even if the same process is executed multiple times, they are all handled as separate activity instances. In addition, even in a business performed according to the same business process model, a business process instance exists for each matter. FIG. 1 shows the relationship between a business process model and a business process instance.

  The importance of business process management (BPM) has been recognized in order to keep business operations in an appropriate state and to improve business efficiency through effective use of business systems. In BPM, the current business situation is grasped, problems are analyzed, a plurality of solutions are considered, and an effective solution is selected and reflected in an actual business or business system. These steps need to be performed continuously and repeatedly in order to be able to follow changes in the business and the environment surrounding the business.

  Of these steps, problems are often analyzed by interviewing business personnel or observing business operations. In addition, in order to analyze problems efficiently and objectively, based on the data obtained from the execution history of the business system, the index indicating the business efficiency and their variability are calculated, and the type of project, time zone There are also analysis methods and techniques for classifying and tabulating from various viewpoints such as regions, and investigating the relationship between them, and these are provided as a function of Business Intelligence (BI).

  The index indicating business efficiency is, for example, the time from when a matter occurs until the completion of all processing, the time that the person in charge actually processes, the time it takes to take over the matter between the persons in charge , Costs, etc. Furthermore, since there are cases where a business process instance of each case can be obtained directly from the execution history of the business system or by the method described in Patent Document 1, in Patent Document 2, a method of proceeding with processing is taken in from the viewpoint of classifying the case. A method and apparatus for quantitatively evaluating the relationship between several characteristics related to business efficiency and automatically extracting features with strong relations as a cause compensation for a decrease in business efficiency are shown.

  By classifying and totaling the cases in this way, it is possible to verify how the business efficiency differs depending on the category of the classification destination when the cases are classified. Further, by quantitatively evaluating the relationship between the viewpoint of classifying the case and the business efficiency, it is possible to find a cause for the decrease in the business efficiency.

  In addition to the “business” performed by business systems in a company, various functions of computers, such as Web browsing and service use, application software operation, and computer operation system operations, are arranged in order or in parallel. "Work" that is performed automatically or semi-automatically using, or "work" that can be regarded as conceptually composed of multiple steps even without using a computer, such as planting, fertilizing, watering, Manual work such as crop cultivation consisting of processes such as pruning, topdressing, and harvesting is also performed by regarding the objectives achieved in each work as a "project" and using the functions and steps as "processing". It is possible to express how to proceed in the same way as the above-described business process instance. Even in that case, the work is evaluated, such as work time and waiting time, purchase price on the Web service, the quality of documents created using application software, or the percentage of crops that became harvested or sick. By quantitatively verifying how the indicators to be processed differ depending on the difference in the processing method in the work, or by discovering the characteristics of the work progress that causes the difference in the indicators to evaluate the work, It is conceivable to deal with a lot of work in a more efficient way.

  In the present specification, a method of classifying each case according to a business or a process to be performed in the work including the “work” as described above will be described below. In other words, it is only necessary that the processing method is conceptually composed of a plurality of steps, and it is not limited whether it is in business or whether it is performed using a machine or a computer.

[Conventional technology]
Even for the same business, the details of the business process instance vary from case to case. Therefore, if all the differences in business process instances are differentiated when classifying matters by the way of processing for the purpose of aggregation or evaluation of the relationship with business efficiency, the categories of classification destinations increase. On the other hand, the cases classified into each category are very few, and a significant analysis result cannot be obtained. Accordingly, it is necessary to determine the viewpoint of classification of what is characteristic of the business process instance to express and distinguish the processing method, and to classify it while ignoring other differences at the time of classification.

In Patent Document 2, the following two types are used as features of a business process instance for expressing how to proceed.
(1) Which process was performed immediately before the execution of each process. Or what process was performed immediately after the execution of each process?
(2) What state was there before each process was executed? Or what kind of processing was executed in each state?

In (1), whether the analyst pays attention to a certain process in the business process model, and whether to match / mismatch how to proceed with the process by the process immediately before the execution of the focused process, or the process immediately after the execution Is specified, whether to proceed with the process is expressed accordingly, and a match / mismatch is determined. In FIG. 2, for some business process instances, the method of processing is expressed by using the process immediately before the execution of the process a ₂ , and the example of determining the match / mismatch and the process immediately after the execution of the process a ₃ are characterized. As an example, the method of processing is expressed, and a match / mismatch is determined.

In (2), the states to be distinguished are all defined as different states based on the execution status (executed / unexecuted) of various processes and their execution order. When the analyst specifies that attention is paid to a certain process in the business process model, the method of proceeding with respect to the case is expressed by the state before execution of the focused process, and matching / mismatching is determined. Alternatively, when the analyst specifies that attention is paid to a certain state among the predefined states, the process executed for the case is expressed by the processing executed in the focused state, and matching / mismatching is determined. FIG. 3 shows an example of how processing progress is expressed for a number of business process instances by using the state before execution of process a ₅ as a feature, and an example in which a match / mismatch is determined, and a process executed in state s ₃ are characterized. As an example, the method of processing is expressed, and a match / mismatch is determined.
JP 2006-338274 A JP 2008-117127 A

  In the method of representing the way of processing by the feature of (1) of Patent Document 2, attention is paid to which branch is used for transition in each conditional branch and its merge in the business process model. As a result, the business process instance for each item is expressed in a very simplified manner. Although knowledge about the business is not required and the processing can be performed mechanically, only the processing of the transition source and the transition destination of the focused processing is considered even though the processing is advanced. In some cases, sufficient analysis results can be obtained by this method. However, if the differences between business process instances are not distinguished, it is necessary to express how to proceed with features that reflect the differences a little more. There is also.

In FIG. 4, for example, the process immediately after the execution of a _9, processing a ₁₀ operational efficiency or transitions to either processing a ₁₁ transitions varies widely but, the processing before the execution of either itself transition process a ₉ to either a ₃ consider the case as highly dependent on whether it has been executed. Processing classified by the transition destination immediately after the execution of a _9, Doing relevance evaluation of the aggregation and operational efficiency, "the transition destination immediately after the execution of processing a _9, operational efficiency is greatly different" that (A) Can be verified and discovered. Separately, by classifying by the transition destination immediately after the execution of the process a ₂ and evaluating the relevance with the aggregation and the business efficiency, the “business efficiency is greatly increased by the transition destination immediately after the execution of the process a ₂ ”. “Different” (b) may be verified and discovered.

However, considering that the transition destination immediately after the execution of the process a ₉ is greatly influenced by the execution status of the process a ₃ , (a) is a simple consequence of (b). If we categorize them separately and evaluate their relevance to aggregation and business efficiency, we would like to know that “business efficiency varies greatly depending on whether or not process a ₃ has been executed before execution of process a ₉ ” (c) Until then, it cannot be verified or discovered. That is, in this state, it is impossible to classify cases by simultaneously distinguishing a plurality of differences in the processing method of interest.

Further, it is not sufficient to extend the method representing the method of proceeding with the feature (1) of Patent Document 2 as it is, and to express the method of proceeding depending on which process is performed immediately before or after a plurality of processes. For example, in the cases 1 and 2 of FIG. 5 in which each process is performed according to the business process model of FIG. 4, even if the execution order of the processes a ₂ and a ₉ is different, the process is considered to be the same, so the same It is classified into categories, and (c) cannot be verified or discovered. In other words, by combining the transition sources or transition destinations of a plurality of processes and representing the method of proceeding the process, the partial part of the process proceeding is still fragmented, and in what order the processes were executed, There is no distinction in the difference in order of processing such as which processing and which processing is executed in parallel.

The method of expressing the method of processing with the feature of (2) of Patent Document 2 is a method for solving these problems. For example, in the example of FIG. 5, the state is defined according to the execution status of the process a ₃ . Thus, how to proceed the overall processing up to processing a ₉ is performed can be represented verify or on minimum detail in the discovery of (c). On top of that, to classify projects depending were in what state before execution of the process a _9. In this way, in principle, business process instances can be differentiated with the details required for tabulation and evaluation of relevance with business efficiency, and cases can be classified.

However, it may be difficult to enumerate and define all the states you want to distinguish in advance. In particular, when there are many processes related to the determination of the state (process a _{3 in} the example of FIG. 5), the execution status of each process and the combination of the execution order that must be considered for defining the state become enormous. A lot of effort is required. In addition, it is not necessary to consider the combinations related to the execution status and execution order of the processes related to state determination that do not appear in the history of the analysis target, but the business process instance of the matter included in the actual analysis target If you don't examine it, you won't be able to judge it, and it will be difficult to reduce your labor.

  In order to solve the above problems, the present invention relates to a method, an apparatus, and a program for realizing a classification having the following features at the same time.

  Feature 1: Cases can be classified by ignoring differences in processing methods other than the processing of interest.

  Feature 2: Cases can be categorized by distinguishing differences in how to proceed with a plurality of processes of interest and differences in order relations of processes.

  Feature 3: The analyst can easily specify the viewpoint of the classification of what is expressed as a feature and how to proceed with the process.

  FIG. 6 shows the position of the present invention in the case classification method.

<Overview>
In the present invention, with respect to one or a plurality of processes of interest designated by an analyst, which processes are executed in what order, which processes and which processes are executed in parallel are compared. Thus, the matching / non-matching of the processing methods of the cases is determined, and the matching cases are classified into the same category, and different cases are classified into different categories. At this time, processing that is not focused on is ignored even if there is a difference in how the processing proceeds.

  Here, when a business process instance is regarded as a non-circular directed graph, the process of interest is the partial graph. Therefore, in order to classify the case, it is determined whether the subgraph of the acyclic directed graph is matched / mismatched (isomorphic). As an algorithm for determining isomorphism of a subgraph, for example, J.A. R. Ullman, “An algorithm for subgraph isomorphism”, Journal of the Association for Computing Machinery, 23 (1), pp. 199 An algorithm described in 31-42, 1976 (hereinafter referred to as Reference 1) is known.

  Determine whether the subgraph that represents the method of processing of interest that exists in the business process instance of one Matter also exists in the business process instance of the other Matter by determining the isomorphism of the subgraph Is possible. However, a partial graph representing how to proceed with the processing of interest may be divided into a plurality of parts in a business process instance of one item. When classifying cases with regard to how to proceed with the process of interest, it is necessary to determine match / mismatch by taking into account the order relationship of those subgraphs, and to determine match / mismatch as they are by determining the isomorphism of the subgraphs. I can't.

FIG. 7 shows an example in which matching / mismatching cannot be properly determined by determining the isomorphism of the partial graph. Although the method of proceeding with respect to a ₂ , a ₃ , a ₄ , a ₅ , a ₇ , and a ₈ of the cases 1, ₂ and ₃ is the same for each part, it is considered that they match in the isomorphism determination of the subgraph. , A ₂ , a ₃ , a ₄ , a ₅ are different from the order of a portion consisting of a ₇ , a ₈ .

  In order to distinguish the difference in the order relationship of the subgraphs that indicate how to proceed with the processing of interest in the business process instance, the processing that was not focused on was executed in what order. It is necessary to determine match / mismatch by ignoring which and which were executed in parallel, and considering what order relationship the processing focused on via them is There is. Therefore, in the present invention, the following processing is performed on the business process instance of the item to be classified.

(I) Anonymization process Anonymization of activity instances of non-focused processes eliminates the distinction between processes.

(Ii) Integration processing By integrating neighboring anonymized activity instances, a structure that expresses in what order non-focused processing was executed and which were executed in parallel. Exclude from business process instances.

  After that, the business process instances obtained as a result of these processes are compared with each other to determine whether or not the process progresses with respect to the case, and classify them. FIG. 8 shows an example in which anonymization processing and integration processing are performed on business process instances for classification.

  In addition, the process which compares the business process instances obtained as a result of the anonymization process and the integration process, and determines the match / mismatch of how to proceed the process with respect to the case is an isomorphism determination of the acyclic directed graph. As a method for realizing this, the algorithm described in Reference Document 1 described above, B.I. D. McKay, “Practical graph isomorphism”, Congressus Numeranitum, vol. 30, pp. 45-87, 1981 (hereinafter referred to as Reference 2), and the algorithm “naturity” is used. In the present invention, what algorithm is used for determining the isomorphism of an acyclic directed graph is not limited.

  Hereinafter, the anonymization process and the integration process will be described in detail.

<Detailed explanation>
《Anonymization process》
In the business process instance of each case, the process name of the activity instance of the process not designated by the analyst as the process of interest is changed to the same process name. In the following description, the process name is

Shall be changed to In addition, an activity instance whose process name is anonymized by this is called an anonymous activity instance, and an activity instance that is not anonymized is called a non-anonymous activity instance.

<Integration processing>
In order to be able to distinguish a plurality of differences in the processing method of interest, including a difference in the order relationship of processing, the integration processing needs to satisfy the following conditions.
Condition 1: The non-anonymous activity instance that is the transition source and the transition destination of the non-anonymous activity instance does not change before and after the integration process.
Condition 2: The order relationship of non-anonymous activity instances does not change before and after the integration process.
Condition 3: As long as the conditions 1 and 2 are satisfied, the number of anonymous activity instances included in the business process instances obtained as a result of the integration process is minimum.

Condition 1 is a condition for guaranteeing that the result does not change before and after the integration process when local match / mismatch is determined for the transition source process or the transition destination process of the activity instance of the process of interest. FIG. 9 shows an example of integration processing that does not satisfy condition 1. In this example, the transition from the process a ₁ to the process a ₃ was not directly performed before the integration process, but the transition is performed directly after the integration process.

Condition 2 is a condition for guaranteeing that the result does not change before and after the integration process when matching / mismatch is determined regarding the order relationship of the activity instances of the process of interest. Also, if the order relationship between any two non-anonymous activity instances in the business process instance of each matter is maintained before and after the integration process, the part representing the process of interest is a plurality of parts of the business process instance. Even in the case where they exist separately, the order relationship between these parts is also maintained. FIG. 10 shows an example of integration processing that does not satisfy condition 2. In this example, the processing a ₁ and the processing a ₂ are executed in parallel before the integration processing, but the processing a ₁ is performed before the processing a ₂ after the integration processing. Yes.

  Condition 3 is a condition for guaranteeing that the same result is always obtained when integration processing is performed for business process instances in which the processing methods of a plurality of processes that are in agreement include the order relationship of the processes. is there. FIG. 11 shows an example of integration processing that does not satisfy condition 3. In this example, the business process instances of the matters 1 and 2 are different from the business process instances as a result of the integration process, even though the process of interest, including the order relationship of the processes, is the same as the process of interest. It has become. In such a case, as a result of the integration process, the obtained business process instances are compared with each other, and in the process of determining the match / mismatch of how to proceed with the matter, these are determined to be mismatched, resulting in an incorrect classification result. End up.

  In order to satisfy these conditions, the present invention repeatedly applies the following rules shown in FIG. The order in which these are applied is not limited.

Integration rule 1
Applicable conditions The following holds true for anonymous activity instances b _u and b _v .

- from b _u to b _v are directly transition.

· B _v of the transition source b _u only.

Application operation All transition destinations of b _v are added as transition destinations of b _u and b _v is deleted.

Integration rule 2
Applicable conditions The following holds true for anonymous activity instances b _u and b _v .

- from b _u to b _v are directly transition.

・ The transition destination of b _u is only b _v .

Application operation All transition sources of b _u are added as transition sources of b _v and b _u is deleted.

Integration rule 3
Application Conditions The following holds true for activity instances _bu that may or may not be anonymized and anonymous activity instances b _v and b _w .

Direct transition from b _u to b _v and b _w

There is no direct or indirect transition from b _v to b _w .

• There is no direct or indirect transition from b _w to b _v .

Apply operation b _v all b _w of the transition source as the transition source and add, by adding all the transition destination b _w as the transition destination b _v, deletes the b _w.

Integration rule 4
Application Conditions The following holds true for the anonymous activity instances b _u and b _v and the activity instance b _w that may or may not be anonymized.

-Direct transition from b _u and b _v to b _w .

· B b _v directly or not also transition to indirectly to from _u.

• There is no direct or indirect transition from b _v to b _u .

Add all b _v of transition source as the transition source application operation b _u, add all transition destination b _v as the transition destination b _u, deletes the b _v.

Integration rule 5
Applicable conditions The following holds true for anonymous activity instances b _u and b _v .

- from b _u to b _v are directly transition.

- from b _u to b _v is also a transition to indirectly.

To delete a direct transition to the b _v from the application operation b _u.

Here, regarding the two activity instances b _u and b _v , “directly transitioning from b _u to b _v ” means that the transition is made without passing through other activity instances, that is, b. _This means that b _v is included in the transition destination of _u and b _u is included in the transition source of b _v . When a business process instance is expressed as a directed acyclic graph, this corresponds to the existence of a link from b _u to b _v . “Indirect transition from b _u to b _v ” means that the transition is via another activity instance, that is, b _v does not include b _{v in} the transition destination of b _u , and b _v B _u is not included in the transition source of, but b _v can eventually be reached if the transition destination activity instance of b _u , the activity instance of the transition destination, and so on are traced. When a business process instance is expressed as an acyclic directed graph, this corresponds to being able to reach b _v by following the link forward from b _u without using the link from b _u to b _v .

Procedures for handling business process instances as an adjacency matrix (hereinafter referred to as P) in order to clarify the specific method of determining applicability and applying operations when applying each of the integration rules 1 to 5 This will be described below. The adjacency matrix is a matrix whose (u, v) element p _{u, v} is 1 when there is a transition from b _u to b _v and 0 when there is no transition. Since the business process instance is a non-circular directed graph, it can be handled by expressions other than the adjacency matrix when realizing the process of determining whether to apply the integration rules 1 to 5 and performing the application operation. A procedure equivalent to the case of handling as an adjacency matrix can be considered. Therefore, the present invention is not limited to the method in the case where business process instances are handled as an adjacency matrix.

-Preparation-
Based on the adjacency matrix, the following is obtained in advance.

-Whole row sum Sum of each row in the adjacency matrix. This corresponds to the number of transition-destination activity instances for each activity instance.

Anonymous partial row sum The sum of each row of the portion related to the transition to the anonymous activity instance of the adjacency matrix. This corresponds to the number of transitional anonymous activity instances for each activity instance.

-Overall column sum Sum of each column in the adjacency matrix. This corresponds to the number of transition source activity instances for each activity instance.

Anonymous partial column sum The sum of each column of the portion related to the transition from the anonymous activity instance of the adjacency matrix. This corresponds to the number of transition source anonymous activity instances for each activity instance.

-Reachability matrix A matrix that represents whether each activity instance has transitioned directly or indirectly, or can be reached by following the links in order when the business process instance is regarded as an acyclic directed graph, A matrix whose (u, v) element q _{u, v} is 1 when b _u can reach b _v and 0 when it cannot be reached. Hereinafter, it is represented by Q.

The reachability matrix Q can be calculated by Q = (E _m + P) ^m−1 where m is the number of activity instances in the business process instance. However, in accordance with operation Boolean operation rule for each element of the matrix, sum the logical OR, AND is assumed ANDing, E _m is the unit matrix of m × m. In the present specification, hereinafter, an adjacency matrix added with an entire row sum, an anonymous partial row sum, an entire column sum, and an anonymous partial column sum is referred to as an expanded adjacency matrix. Since the adjacency matrix itself is not necessary in the following description, the expanded adjacency matrix is also represented by P. The expanded adjacency matrix for the business process instance after the anonymization process of FIG. 13 is shown in FIG. 14, and the reachability matrix is shown in FIG. However, for ease of viewing, the part that is 0 in the adjacency matrix and the reachability matrix is left blank.

-Integration rule 1-
〔Applicable condition〕
For the anonymous activity instance b _v , the total column sum and the anonymous partial column sum of the v-th column of the expanded adjacency matrix P are 1.

[Apply operation]
(1) Find the element p _{u, v} (where 1 ≦ u ≦ m) which is 1 in the v-th column of the expanded adjacency matrix P.
(2) For each column of the expanded adjacency matrix P (k-th column, where 1 ≦ k ≦ m), the anonymous partial column sum, the entire column sum, and the element p _{u, k} in the u-th row are updated as follows: .

(I) When both p _{u, k} and p _{v, k} are 1, 1 is subtracted from the anonymous partial sequence sum and the entire sequence sum.

(Ii) Update p _{u, k} with p _{u, k} ← p _{u, k} + p _{v, k} . However, the sum “+” means a logical sum.
(3) Delete the v-th column and the v-th row of the expanded adjacency matrix P.
(4) Recalculate the anonymous partial row sum and the total row sum for the (original) u-th row of the expanded adjacency matrix P. However, deleted columns are ignored.
(5) Delete the vth column and the vth row of the reachability matrix Q.

  The expanded adjacency matrix before applying the integration rule 1 to the business process instance of FIG. 13, the expanded adjacency matrix after application, and the reachability matrix are shown in FIGS. 16, 17, and 18, respectively. The instance is shown in FIG.

-Integration rule 2-
〔Applicable condition〕
For the anonymous activity instance _bu , the total row sum and the anonymous partial row sum of the u-th row of the expanded adjacency matrix P are 1.

[Apply operation]
(1) Find an element p _{u, v} (where 1 ≦ v ≦ m) which is 1 in the u-th row of the expanded adjacency matrix P.
(2) For each row of the expanded adjacency matrix P (k-th row, where 1 ≦ k ≦ m), the anonymous partial row sum, the entire row sum, and the element p _{k, v} of the v-th column are updated as follows.

(I) When p _{k, u} and p _{k, v} are both 1, 1 is subtracted from the anonymous partial row sum and the entire row sum.

(Ii) Update p _{k, v} with p _{k, v} ← p _{k, v} + p _{k, u} . However, the sum “+” means a logical sum.
(3) Delete the u-th column and u-th row of the expanded adjacency matrix P.
(4) Recalculate the anonymous partial column sum and the total column sum for the (original) column v of the expanded adjacency matrix P. However, deleted lines are ignored.
(5) Delete u-th column and u-th row of reachability matrix Q.

  The expanded adjacency matrix before applying the integration rule 2 to the business process instance in FIG. 19, the expanded adjacency matrix after application, and the reachability matrix are shown in FIGS. 20, 21, and 22, respectively. The instance is shown in FIG.

-Integration Rules 3-
〔Applicable condition〕
When the anonymous partial row sum of the u-th row of the expanded adjacency matrix P is 2 or more and two _{pu, v} and _{pu, w} are selected from the elements that are 1 in the row, the following condition is satisfied.

B _v and b _w are both anonymous activity instances.

In the reachability matrix Q, q _{v, w} and q _{w, v} are both 0.

[Apply operation]
(1) For each column of the expanded adjacency matrix P (the k-th column, where 1 ≦ k ≦ m), the anonymous partial column sum, the entire column sum, and the element p _{v, k} in the v-th row are updated as follows. .

(I) When p _{v, k} and p _{w, k} are both 1, 1 is subtracted from the anonymous partial sequence sum and the total sequence sum.

(Ii) Update p _{v, k} with p _{v, k} ← p _{v, k} + p _{w, k} . However, the sum “+” means a logical sum.
(2) For each row of the expanded adjacency matrix P (k-th row, where 1 ≦ k ≦ m), the anonymous partial row sum, the entire row sum, and the element p _{k, v} of the v-th column are updated as follows.

(I) When p _{k, v} and p _{k, w} are both 1, 1 is subtracted from the anonymous partial row sum and the entire row sum.

(Ii) Update p _{k, v} with p _{k, v} ← p _{k, v} + p _{k, w} . However, the sum “+” means a logical sum.
(3) Delete the w-th row and the w-th column of the expanded adjacency matrix P.
(4) Recalculate the anonymous partial row sum and the total row sum for the (original) row v of the expanded adjacency matrix P. However, deleted columns are ignored.
(5) Recalculate the anonymous partial column sum and the total column sum for the (original) column v of the expanded adjacency matrix P. However, deleted lines are ignored.
(6) For each column of the reachability matrix Q (the k-th column, where 1 ≦ k ≦ m and k ≠ v, w), replace each element q _{v, k} in the v-th row with q _{v, k} ← Update with q _{v, k} + q _{w, k} . However, the sum “+” means a logical sum.
(7) For each row of the reachability matrix Q (k-th row, where 1 ≦ k ≦ m and k ≠ v, w), replace each element q _{k, v} in the v-th column with q _{k, v} ← q Update with _{k, v} + q _{k, w} . However, the sum “+” means a logical sum.
(8) Delete the w-th row and the w-th column of the reachability matrix Q.

  The expanded adjacency matrix before applying the integration rule 3 to the business process instance of FIG. 23, the expanded adjacency matrix after application, the reachability matrix before application, and the reachability matrix after application are shown in FIGS. 25, 26, and 27 show business process instances obtained as a result, as shown in FIG.

-Integration Rule 4-
〔Applicable condition〕
When the anonymous partial column sum of the w-th column of the expanded adjacency matrix P is 2 or more and two _{pu, w} and p _{v, w} are selected from the elements that are 1 in the column, the following condition is satisfied.

B _u and b _v are anonymous activity instances.

In the reachability matrix Q, q _{u, v} and q _{v, u} are both 0.

[Apply operation]
(1) For each column of the expanded adjacency matrix P (k-th column, where 1 ≦ k ≦ m), the anonymous partial column sum, the entire column sum, and the element p _{u, k} in the u-th row are updated as follows. .

(I) When both p _{u, k} and p _{v, k} are 1, 1 is subtracted from the anonymous partial sequence sum and the entire sequence sum.

(Ii) Update p _{u, k} with p _{u, k} ← p _{u, k} + p _{v, k} . However, the sum “+” means a logical sum.
(2) For each row of the expanded adjacency matrix p (k-th row, where 1 ≦ k ≦ m), the anonymous partial row sum, the entire row sum, and the element p _{k, u} of the u-th column are updated as follows.

(I) When p _{k, u} and p _{k, v} are both 1, 1 is subtracted from the anonymous partial row sum and the entire row sum.

(Ii) Update p _{k, u} with p _{k, u} ← p _{k, u} + p _{k, v} . However, the sum “+” means a logical sum.
(3) Delete the v-th row and v-th column of the expanded adjacency matrix P.
(4) Recalculate the anonymous partial row sum and the total row sum for the (original) u-th row of the expanded adjacency matrix P. However, deleted columns are ignored.
(5) Recalculate the anonymous partial column sum and the total column sum for the (original) u-th column of the expanded adjacency matrix P. However, deleted lines are ignored.
(6) For each column of the reachability matrix Q (the k-th column, where 1 ≦ k ≦ m and k ≠ u, v), replace each element q _{u, k} in the u-th row with q _{u, k} ← Update with q _{u, k} + q _{v, k} . However, the sum “+” means a logical sum.
(7) For each row of the reachability matrix Q (k-th row, where 1 ≦ k ≦ m and k ≠ u, v), replace each element q _{k, u} in the u-th column with q _{k, u} ← q Update with _{k, u} + q _{k, v} . However, the sum “+” means a logical sum.
(8) Delete the vth row and vth column of the reachability matrix Q.

  The expanded adjacency matrix before applying the integration rule 4 to the business process instance of FIG. 28, the expanded adjacency matrix after application, the reachability matrix before application, and the reachability matrix after application are respectively shown in FIGS. 30, FIG. 31 and FIG. 32 show the business process instances obtained as a result, as shown in FIG. The application operation of the integration rule 4 is obtained by replacing v with u, w with v, and u with w in the application operation of the integration rule 3.

-Integration rule 5-
〔Applicable condition〕
For anonymous activity instance b _u , select one element p _{u, v} out of elements whose total row sum of the u-th row of the expanded adjacency matrix P is 2 or more and whose anonymous partial row sum is 1 or more and is 1 in that row At that time, the following conditions are satisfied.

B _u is an anonymous activity instance.

The inner product of the u-th row vector of the matrix temporarily set to p _{u, v} = 0 in the extended adjacency matrix P _{and the v} -th column vector of the matrix temporarily set to q _{u, v} = 0 in the reachability matrix Q Is 1. However, sum means logical sum and product means logical product.

[Apply operation]
(1) In the expanded adjacency matrix P, p _{u, v is set} to 0.
(2) The reachability matrix Q is not updated.

For the business process instance shown in FIG. 33, b ₁₂ is integrated into b ₁₁ according to the integration rule 1 and b ₁₅ is integrated into b ₇ according to the integration rule 2. The expanded adjacency matrix before application and the expanded adjacency matrix after application are shown in FIGS. 35 and 36, respectively, and the business process instances obtained as a result are shown in FIG.

  Note that the integration processing may be directly performed without performing the above-described anonymization processing. That is, the activity instances of processes other than the specified process in the business process instance are integrated. Specifically, any of the same integration rules 1 to 5 can be applied to the activity instance in the business process instance. You may make it integrate by applying repeatedly until it disappears. However, “anonymous activity instance” in the integration rule is read as “non-focused activity instance” which is an activity instance of processing other than the specified processing of interest, and “non-anonymous activity instance” is designated processing of interest. It is assumed that the activity instance is a “target activity instance”.

  When classifying matters according to the process progress in the business, assign the same process name to all activity instances of the process other than the target process specified by the analyst for the business process instance that represents the process progress of each matter Anonymize, and then integrate nearby anonymized activity instances to exclude their structures that indicate how they were executed or executed in parallel, and then their work By comparing the process instances and determining whether or not they match, and classifying the matching instances into the same category, and classifying the nonmatching categories into different categories, a classification method having the features 1 to 3 described above is realized.

  Further, in the present invention that realizes the above, if more attention processing is performed, it is possible to distinguish business process instances in more detail and obtain a classification result with fine granularity, while if less attention processing is performed, business processing is simplified more easily. It is possible to distinguish process instances and obtain a coarse-grained classification result. Therefore, it becomes possible to classify cases according to how to proceed with the processing in the business with various granularities regarding various processing in the business.

  With these effects, when the process is advanced by classifying the cases according to the present invention, and using the classification result to aggregate the index indicating the business efficiency and the variation for each category of the classification destination It will be possible to quickly verify whether the business efficiency is reduced according to the purpose of analysis and the intention of the analyst. In addition, by classifying cases according to the present invention, quantitatively evaluating the relevance between the classification result and business efficiency, and automatically extracting features with strong relevance as a cause compensation for business efficiency degradation, processing It will be possible to find a more direct cause of operational efficiency decline in relation to how to proceed.

  As a result, it becomes possible to analyze current business problems easily and accurately, and to enable rapid and continuous BPM, making it easy to keep business processes in an optimal state and operating costs. Can be reduced.

<Configuration of Invention Device and Processing Procedure of Invention Method>
FIG. 38 shows an example of an embodiment of an apparatus for classifying items according to how to proceed with a process of interest in business or work according to the present invention, here an example realized on a known computer (computer). In the figure, 1 is a history database, 2 is a classification target case designation unit, 3 is a classification target case storage unit, 4 is a target processing specification unit, 5 is a target processing storage unit, 6 is a business process instance acquisition unit, and 7 is a business process An instance storage means, 8 is an anonymization processing section, 9 is an integration processing section, 10 is a case classification section, 11 is a case classification result storage section, and 12 is a case classification result display section.

  The history database 1 is a database management system that controls reading / writing from a file stored on a hard disk in a computer that permanently holds data for storing / holding item information and business process instances thereof. is there.

  The classification target case designating unit 2 is a program that is executed by an arithmetic unit in a computer for allowing an analyst to specify a case to be classified or its conditions. The classification target item storage means 3 is a memory in the computer for holding the item specified by the classification target item specifying unit 2 or its conditions.

  The target process designation unit 4 is a program that is executed by an arithmetic device in the computer for causing the analyst to designate a process of interest. The attention process storage means 5 is a memory in the computer for holding the attention process designated by the attention process designation unit 4.

  The business process instance acquisition unit 6 is executed by an arithmetic unit in the computer for acquiring from the history database 1 the business process instance of the case specified by the classification target case specifying unit 2 and stored in the classification target case storage unit 3. It is a program. The business process instance storage means 7 is a memory in the computer for holding business process instances acquired by the business process instance acquisition unit 6 and processed by the anonymization processing unit 8 and the integration processing unit 9.

  The anonymization processing unit 8 is a process other than the target process in the business process instance held in the business process instance storage unit 7, that is, the process specified by the target process specification unit 4 and stored in the target process storage unit 5. It is a program executed by a computing device in a computer for anonymizing a process activity instance.

  The integrated processing unit 9 is executed by a computing device in the computer for integrating the neighboring anonymous activity instances in the business process instance after the anonymization processing held in the business process instance storage unit 7 in accordance with the integration rules 1 to 5. Program.

  The case classification unit 10 is executed by a computing device in a computer for classifying cases into categories based on the match / mismatch of business process instances after anonymization processing and integration processing held in the business process instance storage unit 7. It is a program. The case classification result storage means 11 is a memory in the computer for holding the result classified by the case classification unit 10.

  The case classification result display unit 12 is a program that is executed by an arithmetic unit in a computer for displaying cases classified into categories to an analyst.

  Whether these are in the same computer or in a plurality of computers via a network is not limited in the present invention.

  In the present invention, the completed case information and its business process instance are registered in the history database 1 before the analyst tries to classify the cases. After that, when classifying matters, designation of matters to be classified, designation of target processing, acquisition of business process instances, anonymization of activity instances of non-focused processing, integration of nearby anonymous activity instances, The classification and the display of the classification result are executed in this order.

  The overall processing flow of the present invention is shown in FIG. However, either the classification target item designation or the attention process designation may be performed first. In addition, anonymization of activity instances of non-focused processes, integration of nearby anonymous activity instances, and classification of cases into categories are performed for each case, and this series of processes is repeated for the number of cases to be classified. It is not limited whether the process is repeated for each process for the number of cases to be classified and the process proceeds to the next process.

  In the following, the role of each component, the details of processing, and details of the procedure will be described.

《History database》
The history database 1 holds information (case information) such as the type of case, the time zone where it occurred, and the region (case information) and a business process instance for each case in a database or file. FIG. 40 shows an example of contents and format held in the history database.

  It is assumed that each item has a unique identifier (item ID). Further, in order to hold the business process instance, the activity instance included in the business process instance of each matter is assigned a matter ID. Each activity instance is provided with an identifier (activity instance ID) for uniquely identifying the activity ID, in addition to the case ID. In addition, a process name for indicating which process each activity instance is executing is given. Transitions between activity instances are held as combinations of identifiers of transition source and transition destination activity instances. Note that items other than this may be added as item information and information about activity instances.

  The business process instances held in the history database 1 are assumed to be those directly stored by the business system at the time of business execution or those estimated from the execution history simply output in the execution order as in Patent Document 1. There is no limitation on how it is prepared. Also, the data format (diagram, text, etc.) in which the matter information and the business process instance are held is not limited.

《Category target item specification part》
The classification target item specifying unit 2 allows an analyst to specify a condition related to a case ID or item information via the input / output device of the computer that implements the apparatus of the present invention, and stores the result in the classification target item storage means 3 (see FIG. 39-S1). As a method to specify the conditions related to the case ID or case information, display a list of cases, select from among them, display a list of cases, time zones, and regions, etc. A method for specifying conditions related to project information by selecting a specific type, time zone, region, etc., or using a result obtained by using a tabulation tool or business process analysis tool different from the present invention. Although the method etc. are assumed, it does not specifically limit.

《Classification item storage means》
The classification target item storage unit 3 holds a condition related to item information for extracting a case from the set of item IDs of the items specified by the analyst in the classification target item specifying unit 2 or the item information table.

<< Focus process specification part >>
The attention process designation unit 4 designates one or a plurality of processes of interest to the analyst through the input / output device of the computer realizing the apparatus of the present invention, and saves the results in the attention process storage means 5 (FIG. 39). -S2). As a method of specifying the process of interest, a list of processes included in the analysis target business is displayed, a method of selecting from among them, and a graph structure representing the business process model are graphically displayed, and the range is selected. Although the method etc. are assumed, it does not specifically limit.

<< Target processing storage means >>
The attention process storage means 5 holds a set of processes designated by the analyst in the attention process designation unit 4.

<< Business process instance acquisition part >>
The business process instance acquisition unit 6 uses, as an extraction condition, the process ID of the case to be classified, which is stored in the classification target case storage unit 3, as the extraction condition, and the business process of the corresponding case from the history database 1. The instance is acquired and stored in the business process instance storage unit 7 in association with the case ID (FIG. 39-S3).

<< Business process instance storage means >>
The business process instance storage unit 7 holds the business process instance acquired by the business process instance acquisition unit 6 or the business process instance changed by the anonymization processing unit 8 or the integration processing unit 9. The business process instance may be held in the same table format as the activity instance table and the transition table shown in FIG. 40, or may be held as an (enlarged) adjacency matrix as shown in FIG. It may be held in the form of However, a case ID is associated with each business process instance.

《Anonymization processing unit》
In the anonymization processing unit 8, the original business process instance held in the business process instance storage unit 7 is read, and activity instances of processes that are not included in the set of processes held in the target processing storage unit 5 are read. The process name can be any identical process name that is different from any of the processes of interest (in this specification,

And the result is stored in the business process instance storage unit 7 (FIG. 39-S4).

《Integrated processing unit》
The integrated processing unit 9 reads the business process instance obtained as a result of the anonymization process held in the business process instance storage means 7 and repeatedly applies the integrated rules 1 to 5 until none of them can be applied. Is stored in the business process instance storage means 7 (FIG. 39-S5). FIG. 41 shows a detailed processing flow of the integration processing performed for one business process instance. However, the integration rules 1 to 5 may be applied in any order, and are not particularly limited.

-Step S11-
It is determined whether activity instances b _u and b _v satisfying the application condition of the integration rule 1 exist. If they exist, the operation of applying the integration rule 1 is performed on b _u and b _v .

-Step S12-
It is determined whether activity instances b _u and b _v satisfying the application rule of the integration rule 2 exist. If they exist, the application operation of the integration rule 2 is performed on b _u and b _v .

-Step S13-
It is determined whether activity instances b _u , b _v , and b _w satisfying the application condition of the integration rule 3 exist. If they exist, the operation of applying the integration rule 3 is performed on b _u , b _v , and b _w .

-Step S14-
It is determined whether activity instances b _u , b _v , and b _w satisfying the application condition of the integration rule 4 exist. If they exist, the application operation of the integration rule 4 is performed on b _u , b _v , and b _w .

-Step S15-
It is determined whether activity instances b _u and b _v satisfying the application condition of the integration rule 5 exist, and if they exist, the operation of applying the integration rule 5 is performed on b _u and b _v .

  Further, FIG. 42 shows a processing flow for determining whether or not the integration rule 1 is applicable and performing an application operation if applicable.

-Step S21-
If all anonymous activity instances have been investigated, the process ends. Otherwise, the process proceeds to step S22.

-Step S22-
Take one uninvestigated anonymous activity instance and _let it be b _v .

-Step S23-
If the transition source of b _v is only one anonymous activity instance, this is _regarded as b _u and the process proceeds to step S24. Otherwise, the process returns to step S21.

-Step S24-
All transition destinations of b _v are added as transition destinations of b _u .

-Step S25-
Delete anonymous activity instance b _v .

  Further, FIG. 43 shows a processing flow for determining whether or not the integration rule 2 is applicable and performing an application operation if it is applicable.

-Step S31-
If all anonymous activity instances have been investigated, the process ends. Otherwise, the process proceeds to step S32.

-Step S32-
Take one uninvestigated anonymous activity instance and _let it be b _u .

-Step S33-
If the transition destination of b _u is only one anonymous activity instance, this is set as b _v , and the process proceeds to step S34. Otherwise, the process returns to step S31.

-Step S34-
All b _u transition sources are added as b _v transition sources.

-Step S35-
Delete the anonymous activity instance _bu .

  Further, FIG. 44 shows a processing flow for determining whether or not the integration rule 3 is applicable and performing an application operation if applicable.

-Step S41-
If all activity instances have been investigated, the process ends. Otherwise, the process proceeds to step S42.

-Step S42-
Take one uninvestigated activity instance and _let it be b _u .

-Step S43-
If b _u transition destination anonymous activity instances there are two or more, the process proceeds to step S44, otherwise the process returns to the step S41.

-Step S44-
If two of the combination of the transition destination of anonymous activity instance of b _u all been examined returns to step S41, otherwise the process proceeds to step S45.

-Step S45-
Two combination of unexamined b _u of the transition destination of anonymous activity instance a set is taken out, and b _v, b _w.

-Step S46-
from b _v to b _w also, the process proceeds to step S47 when not in transition to also indirectly directly from b _w to b _v, otherwise it returns to step S44.

-Step S47-
All transition destinations of b _w are added as transition destinations of b _v .

-Step S48-
All transition sources of b _w are added as transition sources of b _v .

-Step S49-
Delete the anonymous activity instance b _w .

  Further, FIG. 45 shows a processing flow for determining whether or not the integration rule 4 can be applied and performing an application operation if applicable.

-Step S51-
If all activity instances have been investigated, the process ends. Otherwise, the process proceeds to step S52.

-Step S52-
Take one uninvestigated activity instance and _let it be b _w .

-Step S53-
If there are two or more anonymous activity instances that are transition sources of b _w , the process proceeds to step S54, and otherwise, the process returns to step S51.

-Step S54-
If all the two combinations of the anonymous activity instances of the transition source of b _w have been investigated, the process returns to step S51, and otherwise, the process proceeds to step S55.

-Step S55-
One set of two combinations of the anonymous activity instances of the transition source of unexamined b _w is taken out and set as b _u and b _v .

-Step S56-
from b _u to b _v also proceeds to step S57 when not in transition to be indirectly directly from b _v to b _u, otherwise returns to step S54.

-Step S57-
All transition destinations of b _v are added as transition destinations of b _u .

-Step S58-
All transition sources of b _v are added as transition sources of b _u .

-Step S59-
Delete anonymous activity instance b _v .

  Further, FIG. 46 shows a processing flow for determining whether or not the integration rule 5 is applicable and performing an application operation if applicable.

-Step S61-
If all anonymous activity instances have been investigated, the process ends. Otherwise, the process proceeds to step S62.

-Step S62-
Take one uninvestigated anonymous activity instance and _let it be b _u .

-Step S63-
If b _u transition destination anonymous activity instance of one or more process proceeds to step S64, otherwise the process returns to the step S61.

-Step S64-
If an anonymous activity instance of the transition destination of b _u all been examined returns to step S61, otherwise the process returns to step S65.

-Step S65-
An anonymous activity instance of the transition destination of unexamined b _u 1 piece is taken out, and b _v.

-Step S66-
When also transition indirectly to b _u from b _v proceeds to step S67, the otherwise returns to step S64.

-Step S67-
Delete the direct transition from b _u to b _v .

《Item classification department》
The case classification unit 10 compares the business process instances obtained as a result of the integration process held in the business process instance storage unit 7 to determine whether or not they match, and the matching cases do not match the same category. Items are classified into different categories, and the results are stored in the item classification result storage means 11 (FIG. 39-S6). In the present invention, a known algorithm described in Reference 2 or the like capable of determining the homomorphism of an acyclic directed graph is used as a method for determining match / mismatch of business process instances and determining a category of a classification destination.

<Item classification result storage means>
The case classification result storage means 11 holds the result of the case classification into categories by the case classification unit 10. FIG. 47 shows an example of contents and format held by the case classification result storage unit 11. Each item is given an identifier (category ID) indicating a classified category. When the name of each category is given by the case classification unit 10, the name may be added.

  In addition, as long as the classification destination category and the case are associated with each other, the data format is not limited.

<Item classification result display section>
The case classification result display unit 12 combines the case classification result held in the case classification result storage means 11 and the information in the case information table of the history database 1 using the case ID as a key, and implements the computer of the present invention. Are displayed to the analyst via the input / output device (FIG. 39-S7). As a format to be displayed, a table format, a format in which a tree structure having each category as a node and each item as an element included in the node is displayed in a diagram, but not particularly limited. In addition to the display of the case classification result, the case classification result display unit 12 displays the original business process instance for each case held in the history database 1 as an acyclic directed graph in response to the request of the analyst. The business process instances obtained as a result of the integration processing held in the business process instance storage means 7 in order to display the common characteristics of business process instances for the items classified into each category It is displayed graphically as a directed graph.

  Note that the present invention provides a program for realizing the functions shown in the configuration diagram of FIG. 38 or a program having the procedures shown in the flowcharts of FIGS. 39 and 41 to 46 via a medium or communication line in a known computer. It can also be realized by installing.

Explanatory diagram showing the relationship between the business process model and business process instances Explanatory drawing which shows an example of the expression of how to proceed by the feature of the business process instance Explanatory drawing which shows the other example of the expression of how to advance the processing by the feature of the business process instance Explanatory diagram showing an example of a business process model Explanatory drawing which shows the example of the expression of how to advance the process by the characteristic of the business process instance corresponding to the business process model of FIG. Explanatory drawing showing the positioning of the present invention in the case classification method Explanatory drawing which shows the example of the matter which cannot judge matching / disagreement appropriately with the isomorphism judgment of the subgraph An explanatory diagram showing an example of classifying cases by performing anonymization processing and integration processing for business process instances Explanatory drawing showing an example of integration processing in which the transition relationship between non-anonymous activity instances is not maintained Explanatory drawing showing an example of integration processing in which the order relationship between non-anonymous activity instances is not maintained Explanatory drawing showing an example of integration processing that does not satisfy the condition that minimizes the number of anonymous activity instances Explanatory drawing which shows rules 1-5 in integrated processing Explanatory drawing showing an example of a business process instance after anonymization processing Explanatory drawing which shows the expansion adjacency matrix with respect to the business process instance of FIG. Explanatory drawing which shows the reachability matrix with respect to the business process instance of FIG. Explanatory drawing which shows the expansion adjacency matrix before applying the integration rule 1 to the business process instance of FIG. Explanatory drawing which shows the expansion adjacency matrix after applying the integration rule 1 to the business process instance of FIG. Explanatory drawing which shows the reachability matrix after applying the integration rule 1 to the business process instance of FIG. Explanatory drawing which shows the business process instance of FIG. 13 after applying the integration rule 1 Explanatory drawing which shows the expansion adjacency matrix before applying the integration rule 2 to the business process instance of FIG. Explanatory drawing which shows the expansion adjacency matrix after applying the integration rule 2 to the business process instance of FIG. Explanatory drawing which shows the reachability matrix after applying the integration rule 2 to the business process instance of FIG. Explanatory drawing which shows the business process instance of FIG. 19 after applying the integration rule 2 Explanatory drawing which shows the expansion adjacency matrix before applying the integration rule 3 to the business process instance of FIG. Explanatory drawing which shows the expansion adjacency matrix after applying the integration rule 3 to the business process instance of FIG. Explanatory drawing which shows the reachability matrix before applying the integration rule 3 to the business process instance of FIG. Explanatory drawing which shows the reachability matrix after applying the integration rule 3 to the business process instance of FIG. Explanatory drawing which shows the business process instance of FIG. 23 after applying the integration rule 3 Explanatory drawing which shows the expansion adjacency matrix before applying the integration rule 4 to the business process instance of FIG. Explanatory drawing which shows the expansion adjacency matrix after applying the integration rule 4 to the business process instance of FIG. Explanatory drawing which shows the reachability matrix before applying the integration rule 4 to the business process instance of FIG. Explanatory drawing which shows the reachability matrix after applying the integration rule 4 to the business process instance of FIG. Explanatory drawing which shows the business process instance of FIG. 28 after applying the integration rule 4 Explanatory drawing which shows the business process instance before applying the integration rule 5 Explanatory drawing which shows the expansion adjacency matrix before applying the integration rule 5 to the business process instance of FIG. Explanatory drawing which shows the expansion adjacency matrix after applying the integration rule 5 to the business process instance of FIG. Explanatory drawing which shows the business process instance of FIG. 34 after applying the integration rule 5 The block diagram which shows an example of embodiment of the item classification device of this invention The flowchart which shows an example of embodiment of the item classification method of this invention Explanatory diagram showing an example of the contents held in the history database Flow chart showing details of integration processing Flowchart showing processing for determining whether or not to apply integration rule 1 and applying operation Flowchart showing processing for determining whether to apply integration rule 2 and applying operation Flowchart showing processing for determining whether to apply integration rule 3 and applying operation Flowchart showing processing for determining whether or not to apply integration rule 4 and applying operation Flowchart showing processing for determining whether or not to apply integration rule 5 and applying operation Explanatory diagram showing an example of case classification results

Explanation of symbols

  1: history database, 2: classification target case designation unit, 3: classification target case storage unit, 4: focused process designation unit, 5: focused process storage unit, 6: business process instance acquisition unit, 7: business process instance storage unit 8: anonymization processing unit, 9: integration processing unit, 10: case classification unit, 11: case classification result storage means, and 12: case classification result display unit.

Claims (13)

Classify cases according to how to proceed with the process of interest in the work or work using the classification target case specification unit, target process specification unit, business process instance acquisition unit, integration processing unit, and item classification unit realized on the computer A method,
A step for the classification target case designating section to let the analyst specify a case to be classified or its conditions;
A step in which a focused process designating unit designates a process focused on by an analyst;
A business process instance acquisition unit acquires the business process instance of the specified matter from the history database storing and holding the matter information and the business process instance;
The integration processing unit determines whether there is a combination of activity instances that satisfy the application conditions in the business process instance for each of a plurality of predetermined integration rules that consist of application conditions and application operations. Integrating an activity instance of a process other than the specified process of interest in the business process instance by performing an operation ;
A matter classification method comprising: a matter classification unit including a step of classifying matters into categories based on matching / non-coincidence of integrated business process instances. The item classification method according to claim 1, wherein the integration processing is performed by repeatedly applying to a activity instance in a business process instance until no predetermined integration rule can be applied. The predetermined integration rule is:
One non-target activity instance that is an activity instance of a process other than the specified target process is directly transitioning to another non-target activity instance, and the transition source of the other non-target activity instance is one When only the target activity instance is included, an integration rule 1 for adding all transition destinations of other non-target activity instances and deleting other non-target activity instances as transition destinations of one non-target activity instance,
When a transition is made directly from one non-focused activity instance to another non-focused activity instance, and the transition destination of one non-focused activity instance is only another non-focused activity instance, An integration rule 2 that adds all transition sources of one non-target activity instance as a transition source and deletes one non-target activity instance;
One non-focused or the specified target activity activity instance of the specified focused process is directly transitioning to another non-focused activity instance and another non-focused activity instance, and the other non-focused activity instance Has not made a direct or indirect transition to another non-focused activity instance, and has not made a direct or indirect transition from another non-focused activity instance to another non-focused activity instance. When all other non-focused activity instance transition sources are added as transition sources of other non-focused activity instances, transition destinations of other non-focused activity instances are further transitioned as transition destinations of other non-focused activity instances. All added, and integration rule 3 further remove other non interest activity instance,
There is a direct transition from one non-focused activity instance and another non-focused activity instance to another non-focused or focused activity instance, and directly from one non-focused activity instance to another non-focused activity instance. If it has not transited indirectly and has not transitioned directly or indirectly from another non-focused activity instance to another non-focused activity instance, it is An integration rule 4 that adds all transition sources of non-focused activity instances, adds all transition destinations of other non-focused activity instances as transition destinations of one non-focused activity instance, and deletes other non-focused activity instances ,
When a transition is made directly from one non-focused activity instance to another non-focused activity instance and also indirectly from one non-focused activity instance to another non-focused activity instance, The item classification method according to claim 2, further comprising an integration rule 5 that deletes a direct transition from the target activity instance to another non-target activity instance. Depending on how to proceed with the focused process in the work or work using the classification target case designating unit, focused process designating unit, business process instance acquisition unit, anonymization processing unit, integrated processing unit and case classifying unit realized on the computer A method for classifying projects,
A step for the classification target case designating section to let the analyst specify a case to be classified or its conditions;
A step in which a focused process designating unit designates a process focused on by an analyst;
A business process instance acquisition unit acquires the business process instance of the specified matter from the history database storing and holding the matter information and the business process instance;
Anonymization processing unit anonymizes activity instances of processing other than the specified processing of interest in the business process instance;
The integration processing unit determines whether there is a combination of activity instances that satisfy the application conditions in the business process instance for each of a plurality of predetermined integration rules that consist of application conditions and application operations. A step of integrating neighboring anonymous activity instances in an anonymized business process instance by performing an operation;
A matter classification method comprising: a matter classification unit including a step of classifying matters into categories based on matching / mismatching of anonymized and integrated business process instances. The case classification according to claim 4, wherein the integration process is performed by repeatedly applying to the activity instances in the anonymized business process instances until none of the predetermined integration rules can be applied. Method. The predetermined integration rule is:
When a transition is made directly from one anonymous activity instance to another anonymous activity instance and the transition source of the other anonymous activity instance is only one anonymous activity instance, the transition destination of one anonymous activity instance is another Integration rule 1 for adding all transition destinations of anonymous activity instances and deleting other anonymous activity instances,
When a transition is made directly from one anonymous activity instance to another anonymous activity instance, and the transition destination of one anonymous activity instance is only another anonymous activity instance, the transition source of the other anonymous activity instance is Integration rule 2 that adds all the transition sources of anonymous activity instances and deletes one anonymous activity instance,
A direct transition from one anonymous or non-anonymous activity instance to another anonymous activity instance and another anonymous activity instance, directly and indirectly from another anonymous activity instance to another anonymous activity instance If there is no transition, and there is no transition from another anonymous activity instance to another anonymous activity instance, either directly or indirectly, another anonymous activity instance as a transition source of another anonymous activity instance. Integration rule 3 to add all the transition sources of, add all transition destinations of other anonymous activity instances as transition destinations of other anonymous activity instances, and delete other anonymous activity instances, and
A direct transition from one anonymous activity instance and another anonymous activity instance to another anonymous or non-anonymous activity instance, either directly or indirectly from one anonymous activity instance to another anonymous activity instance If there is no transition and there is no direct or indirect transition from another anonymous activity instance to one anonymous activity instance, the transition source of another anonymous activity instance is the transition source of one anonymous activity instance. , Adding all transition destinations of other anonymous activity instances as transition destinations of one anonymous activity instance, and deleting other anonymous activity instances,
If you are transitioning directly from one anonymous activity instance to another anonymous activity instance and also indirectly transitioning from one anonymous activity instance to another anonymous activity instance, you can change from one anonymous activity instance to another The item classification method according to claim 5, further comprising: an integration rule 5 that deletes a direct transition to an anonymous activity instance. A device that classifies items according to how to proceed with a process of interest in business or work,
A classification target case designating section that allows an analyst to specify a matter to be classified or its conditions;
A focused process designating unit for designating a process focused on by the analyst;
A business process instance acquisition unit for acquiring the business process instance of the specified matter from the history database storing and storing the matter information and the business process instance;
For each of a plurality of predetermined integration rules that consist of application conditions and application operations, it is determined whether there is a combination of activity instances that satisfy the application conditions in the business process instance. the integration processing unit for integrating said specified activity instance of processing other than of interest was in the business process instance,
An item classification apparatus comprising at least an item classification unit for classifying items into categories based on matching / non-matching of integrated business process instances. The item classification apparatus according to claim 7, wherein the integration process is performed by repeatedly applying a predetermined integration rule to an activity instance in a business process instance until none of the predetermined integration rules can be applied. The predetermined integration rule is:
One non-target activity instance that is an activity instance of a process other than the specified target process is directly transitioning to another non-target activity instance, and the transition source of the other non-target activity instance is one When only the target activity instance is included, an integration rule 1 for adding all transition destinations of other non-target activity instances and deleting other non-target activity instances as transition destinations of one non-target activity instance,
When a transition is made directly from one non-focused activity instance to another non-focused activity instance, and the transition destination of one non-focused activity instance is only another non-focused activity instance, An integration rule 2 that adds all transition sources of one non-target activity instance as a transition source and deletes one non-target activity instance;
One non-focused or the specified target activity activity instance of the specified focused process is directly transitioning to another non-focused activity instance and another non-focused activity instance, and the other non-focused activity instance Has not made a direct or indirect transition to another non-focused activity instance, and has not made a direct or indirect transition from another non-focused activity instance to another non-focused activity instance. When all other non-focused activity instance transition sources are added as transition sources of other non-focused activity instances, transition destinations of other non-focused activity instances are further transitioned as transition destinations of other non-focused activity instances. All added, and integration rule 3 further remove other non interest activity instance,
There is a direct transition from one non-focused activity instance and another non-focused activity instance to another non-focused or focused activity instance, and directly from one non-focused activity instance to another non-focused activity instance. If it has not transited indirectly and has not transitioned directly or indirectly from another non-focused activity instance to another non-focused activity instance, it is An integration rule 4 that adds all transition sources of non-focused activity instances, adds all transition destinations of other non-focused activity instances as transition destinations of one non-focused activity instance, and deletes other non-focused activity instances ,
When a transition is made directly from one non-focused activity instance to another non-focused activity instance and also indirectly from one non-focused activity instance to another non-focused activity instance, The item classification method according to claim 8, further comprising: an integration rule 5 that deletes a direct transition from a target activity instance to another non-target activity instance. A device that classifies items according to how to proceed with a process of interest in business or work,
A classification target case designating section that allows an analyst to specify a matter to be classified or its conditions;
A focused process designating unit for designating a process focused on by the analyst;
A business process instance acquisition unit for acquiring the business process instance of the specified matter from the history database storing and storing the matter information and the business process instance;
An anonymization processing unit that anonymizes an activity instance of a process other than the specified process of interest in the business process instance;
For each of a plurality of predetermined integration rules that consist of application conditions and application operations, it is determined whether there is a combination of activity instances that satisfy the application conditions in the business process instance. , An integrated processing unit that integrates nearby anonymous activity instances in anonymized business process instances,
An item classification device comprising at least an item classification unit for classifying items into categories based on matching / non-coincidence of anonymized and integrated business process instances. The item classification according to claim 10, wherein the integration process is performed by repeatedly applying a predetermined integration rule to an activity instance in an anonymized business process instance until none of the predetermined integration rules can be applied. apparatus. The predetermined integration rule is:
When a transition is made directly from one anonymous activity instance to another anonymous activity instance and the transition source of the other anonymous activity instance is only one anonymous activity instance, the transition destination of one anonymous activity instance is another Integration rule 1 for adding all transition destinations of anonymous activity instances and deleting other anonymous activity instances,
When a transition is made directly from one anonymous activity instance to another anonymous activity instance, and the transition destination of one anonymous activity instance is only another anonymous activity instance, the transition source of the other anonymous activity instance is Integration rule 2 that adds all the transition sources of anonymous activity instances and deletes one anonymous activity instance,
A direct transition from one anonymous or non-anonymous activity instance to another anonymous activity instance and another anonymous activity instance, directly and indirectly from another anonymous activity instance to another anonymous activity instance If there is no transition, and there is no transition from another anonymous activity instance to another anonymous activity instance, either directly or indirectly, another anonymous activity instance as a transition source of another anonymous activity instance. Integration rule 3 to add all the transition sources of, add all transition destinations of other anonymous activity instances as transition destinations of other anonymous activity instances, and delete other anonymous activity instances, and
A direct transition from one anonymous activity instance and another anonymous activity instance to another anonymous or non-anonymous activity instance, either directly or indirectly from one anonymous activity instance to another anonymous activity instance If there is no transition and there is no direct or indirect transition from another anonymous activity instance to one anonymous activity instance, the transition source of another anonymous activity instance is the transition source of one anonymous activity instance. , Adding all transition destinations of other anonymous activity instances as transition destinations of one anonymous activity instance, and deleting other anonymous activity instances,
If you are transitioning directly from one anonymous activity instance to another anonymous activity instance and also indirectly transitioning from one anonymous activity instance to another anonymous activity instance, you can change from one anonymous activity instance to another The item classification device according to claim 11, further comprising an integration rule 5 that deletes a direct transition to an anonymous activity instance. The computer program to function as the respective parts of the project classification apparatus according to any one of claims 7 to 12.

Images