Method for determining condition of category division of key performance indicator, and computer and computer program therefor

Abstract

Embodiments include determining a condition of category division of a key performance indicator. Aspects include dividing multiple pieces of workflow execution information into a division number for the type of the key performance indicator and generating a graph for each of the multiple groups after the division, from workflow execution information belonging to the group. Aspects also include calculating graph edit distances for respective two arbitrary graphs and generating a new group by merging two graphs having the shortest graph edit distance. Aspects further include calculating a graph edit distance for the new group and repeating the generation of the new group and the calculation of the graph edit distance for the new group, until the graph edit distance is equal to or more than a certain fixed edit distance.

What is claimed is: 
     
         1 . A method for determining a condition of category division of a key performance indicator (KPI) from multiple pieces of workflow execution information, each of the multiple pieces of workflow execution information having at least one set of a type of the KPI and a value associated with the type, the method, executed by a computer, comprising:
 (a) dividing the multiple pieces of workflow execution information into a division number for the type of the KPI;   (b) generating a graph for each of the multiple groups after the division into the division number, from workflow execution information belonging to the group;   (c) calculating graph edit distances for respective two arbitrary graphs;   (d) generating a new group by merging two graphs having the shortest graph edit distance;   (e) calculating a graph edit distance for the new group; and   (f) repeating the steps (d) to (e) until the graph edit distance is equal to or more than a certain fixed edit distance, and determining a value to define the new group, as the condition of the category division.   
     
     
         2 . The method according to  claim 1 , wherein the condition of the category division is determined for each type of the KPI. 
     
     
         3 . The method according to  claim 1 , wherein when the type of the KPI is numeric data, the value to define the new group is a value in a range from a minimum value to a maximum value of the numeric data, and the value in the numeric data is determined as the condition of the category division. 
     
     
         4 . The method according to  claim 1 , wherein when the type of the KPI is category data, the value to define the new group is a name of a category to define the new group, and the name of the category is determined as the condition of the category division. 
     
     
         5 . The method according to  claim 1 , further including determining the division number for the type of the KPI. 
     
     
         6 . The method according to  claim 5 , wherein the step of determining the division number includes a step of determining the division number using a statistical method, in response that the type of the KPI is numeric data. 
     
     
         7 . The method according to  claim 5 , wherein the step of determining the division number includes a step of determining the division number from a category number for the type or the value associated with the type, in response that the type of the KPI is category data. 
     
     
         8 . The method according to  claim 1 , wherein the step (a) of dividing the multiple pieces of workflow execution information into the division number further includes a step of sorting the multiple pieces of workflow execution information based on the value associated with the type, in response that the type of the KPI is numeric data, and
 the step (a) of dividing the multiple pieces of workflow execution information into the division number comprises a step of dividing the multiple pieces of workflow execution information after the sort into the division number.   
     
     
         9 . The method according to  claim 1 , wherein the step (b) of generating the graph further includes a step of saving a condition of a region of the generated graph. 
     
     
         10 . The method according to  claim 1 , wherein the step (c) of calculating the graph edit distances for the respective two arbitrary graphs includes:
 (c-1) acquiring a combination of two arbitrary graphs;   (c-2) calculating the graph edit distance for the combination of the two arbitrary graphs; and   (c-3) repeating the steps (c-1) and (c-2) for the remaining combinations of the two arbitrary graphs.   
     
     
         11 . The method according to  claim 1 , wherein the step (d) of generating the new group includes a step of generating the new group from multiple pieces of workflow execution information respectively belonging to the two graphs having the shortest graph edit distance. 
     
     
         12 . The method according to  claim 1 , wherein the step (d) of generating the new group further includes a step of merging respective regions of two groups having the shortest graph edit distance, and saving a condition of a region of the new group. 
     
     
         13 . The method according to  claim 1 , wherein the step (c) of calculating the graph edit distances for the respective two arbitrary graphs further includes a step of recording the calculated graph edit distances in a table. 
     
     
         14 . The method according to  claim 13 , wherein the step (d) of generating the new group further includes a step of adding the generated new group in the table. 
     
     
         15 . The method according to  claim 14 , wherein the step (d) of generating the new group further includes a step of deleting two graphs belonging to two groups having the shortest graph edit distance, from the table. 
     
     
         16 . The method according to  claim 15 , wherein the step (e) of calculating the graph edit distance for the new group further includes a step of recording, in the table after the addition of the new group, the graph edit distance calculated for the new group. 
     
     
         17 . The method according to  claim 14 , wherein the steps of repeating the steps (d) to (e) until the graph edit distance is equal to or more than the certain fixed edit distance includes a step of repeating the steps (d) to (e) until table entries of the table have an edit distance equal to or more than the certain fixed edit distance. 
     
     
         18 . The method according to  claim 14 , wherein the table is a table of K×K.