JP2019049870A - Similar program specifying device, similar program specifying method, and control program - Google Patents

Abstract

To provide a similar program specifying device, a similar program specifying method, and a control program capable of improving accuracy of specifying a similar program.SOLUTION: A similar program specifying device 10 comprises: a first calculator 11A for calculating the timing type unit similarity between two access patterns respectively corresponding to two constituent programs of the program pair for each access timing type with respect to each program pair among multiple programs, by using access correspondence relationship that associates the multiple programs with the access patterns for each access timing type of each program; and a second calculator 11B for calculating the program similarity for each program pair on the basis of the plurality of calculated timing type unit similarities and a weighting coefficient set corresponding to a combination of access target characteristic groups to which two configuration programs of each program pair respectively belong.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a similar program specifying device, a similar program specifying method, and a control program.

  As a company operates and maintains an IT system for a long time, there are times when the system is expanded or remodeled according to the purpose of the corporate activity. As a result, a variety of new and old software, hardware, and development techniques may be mixed and intricately intertwined. In such a situation, it may be difficult to grasp the entire software asset, and it may be very difficult to expand or modify the system.

  For example, in a development site, it is common to copy a base program and rewrite a part of the program to create a new program. The new program created in this way may have functions other than those of the program based on it, or it may be refactored (refactored) with the same function as the program based on it. Sometimes. Thus, in a system that has been maintained for many years, there may be a state in which new and old assets similar to each other are mixed. As a result, inventory of the test target asset or the transfer target asset can not be efficiently implemented at the time of repair or rebuilding, the estimation accuracy may be reduced, and the number of migration man-hours or test man-hours may be increased.

  On the other hand, conventionally, a software asset management apparatus has been proposed (for example, Patent Document 1). The software asset management apparatus of Patent Document 1 arranges information nodes, which are individual information units constituting a software asset, for blocks divided according to the degree of abstraction and the degree of detail, and indicates the relationship between each information node. It holds the asset management matrix set by the relationship link, which is a directed link.

JP, 2014-157598, A

  The present inventor has found that the above-mentioned prior art may not be able to accurately identify a similar program among software assets including a plurality of programs.

  That is, the program has an access pattern including an access destination of the program and an access purpose to the access destination in each of the plurality of access timing types. Then, the inventor found that the access timing type of the access pattern characterizing the program differs depending on the program. For this reason, it has been found that, in the above-described prior art in which the access timing type is not taken into consideration, a similar program can not be identified with high accuracy among software assets including a plurality of programs.

  An object of the present invention is to provide a similar program specifying device, a similar program specifying method, and a control program that can improve the specification accuracy of a similar program.

  The similar program identification device according to the first aspect of the present invention associates a plurality of programs with an access pattern including an access destination of each program and an access purpose to the access destination for each of a plurality of access timing types. Calculate the timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair for each program type of the plurality of programs using the above access correspondence relationship A first calculation unit, a plurality of timing type unit similarities calculated for each program pair by the first calculation unit, and a group combination of access destination characteristic groups to which two constituent programs of each program pair belong. With the corresponding set of weighting factors And a second calculation unit that calculates a program similarity for each program pair, and a program pair in which two configuration programs are similar based on the program similarity calculated by the second calculation unit. And a specific unit.

  A similar program identification method according to a second aspect of the present invention associates a plurality of programs with an access pattern including an access destination for each of a plurality of access timing types of each program and an access purpose to the access destination. Calculate the timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair for each program type of the plurality of programs using the above access correspondence relationship On the basis of the plurality of timing type unit similarities calculated for each program pair, and a weighting factor set corresponding to a combination of access target characteristic groups to which the two constituent programs of each program pair belong. About program pairs Calculating a program similarity, based on the program similarity, two configuration program identifies the program pair similar.

  A control program according to a third aspect of the present invention is an access in which a plurality of programs are associated with an access destination for each of a plurality of access timing types of each program and an access pattern including an access purpose to the access destination. Using the correspondence relationship, for each program pair of the plurality of programs, timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair is calculated for each access timing type, Each program pair is calculated based on the calculated plurality of timing type unit similarities for each program pair, and a weighting coefficient set corresponding to a combination of access target characteristic groups to which the two constituent programs of each program pair belong. About the program Calculating a beam similarity, based on the program similarity, two configuration program identifies the program pair similar, the process, to be executed by the similar program identifying device.

  According to the present invention, it is possible to provide a similar program specifying device, a similar program specifying method, and a control program that can improve the specifying accuracy of similar programs.

It is a block diagram showing an example of the similar program specific device of a 1st embodiment. It is a block diagram showing an example of the similar program specific device of a 2nd embodiment. It is a figure which shows an example of the system configuration information table of 2nd Embodiment. It is a figure which shows an example of the related link table of 2nd Embodiment. It is a figure which shows the structural example of the software asset represented by the system configuration | structure information table of FIG. 3, and the related link table of FIG. It is a figure which shows an example of the weighting coefficient table of 2nd Embodiment. It is a flow chart which shows an example of processing operation of a similar program specific device of a 2nd embodiment. It is a figure which shows the calculation result of program similarity. It is the graph which plotted the value of the program similarity with respect to the number of a program pair. It is the graph which plotted the value of the similarity degree without classification to the number of a program pair. It is the graph which plotted the value of compile time similarity to the number of a program pair. It is the graph which plotted the value of the runtime similarity to the number of a program pair. It is a figure showing the example of hardware constitutions of a similar program specific device.

  Hereinafter, embodiments will be described with reference to the drawings. In the embodiments, the same or equivalent elements will be denoted by the same reference signs, and overlapping descriptions will be omitted.

First Embodiment
<Configuration Example of Similar Program Identification Device>
FIG. 1 is a block diagram showing an example of the similar program identification device of the first embodiment. In FIG. 1, the similar program identification device 10 includes a control unit 11 and a storage unit 12. The control unit 11 includes a timing type unit similarity calculation unit (hereinafter may be referred to as a "first calculation unit") 11A and a program similarity calculation unit (hereinafter referred to as a "second calculation unit"). 11B and a similar program identification unit (hereinafter, may be referred to as a "first identification unit") 11C.

  The first calculating unit (timing type unit similarity calculating unit) 11A uses the “access correspondence relationship” to “access timing type” for each program pair among a plurality of programs included in the “similar program search range”. Each time, “timing type unit similarity” between two “access patterns” respectively corresponding to two configuration programs of the program pair is calculated. The “access correspondence relationship” is stored (held) in the storage unit 12. Here, the “access correspondence relationship” is a correspondence relationship in which a plurality of programs (that is, a plurality of program identifiers) are associated with access patterns for each of a plurality of access timing types of each program. Also, “a plurality of access timing types” include, for example, “at compile time” and “at program execution time”. Further, the definition element defining the “access pattern” includes, for example, “access destination” of the program and “access purpose” to the access destination. Further, the “access destination” includes, for example, a file, a table, a macro and the like. Also, the “access purpose” includes, for example, “quote”, “read”, “write”, and the like.

  That is, for example, when the program pair to be focused on is configured by the first configuration program and the second configuration program, the first calculation unit 11A determines the “compile time” access pattern of the first configuration program and the second configuration program. The similarity with the "compile time" access pattern is calculated as the "compile time" timing type unit similarity. Similarly, the first calculation unit 11A sets the similarity between the access pattern at the time of program execution of the first configuration program and the access pattern at the time of program execution of the second configuration program to the timing of "program execution". Calculated as type-unit similarity. Then, the first calculation unit 11A sequentially changes the target program pair among the plurality of programs included in the “similar program search range”, and calculates the timing type unit similarity for each target program pair.

  The second calculating unit (program similarity calculating unit) 11B is configured by the plurality of timing type unit similarities calculated for each program pair by the first calculating unit 11A, and the “access destination to which two constituent programs of each program pair belong. The program similarity for each program pair is calculated based on the "weighting coefficient set" corresponding to the group combination by characteristic group ". That is, for example, when the group combination of a certain program pair is the first group combination, the second calculating unit 11B compares the first group with respect to the “at compile time” timing type unit similarity of the program pair. Multiply the “compile time” weighting factors included in the weighting factor set corresponding to the combination. In addition, the second calculation unit 11B compares the weight coefficient set “at program execution time” included in the weight coefficient set corresponding to the first group combination with respect to the timing type unit similarity value “at program execution time” of the program pair. Cross over. Then, the second calculation unit 11B calculates the program similarity by adding the timing type unit similarity of “at compile time” and the timing type unit similarity of “at program execution” after being weighted respectively. Do. Here, the “access destination characteristic classified group” is a group corresponding to the access destination characteristic for each access timing type, and is, for example, a group for each subsystem or a group for each software package.

  The first specifying unit (similar program specifying unit) 11C may call a program pair (hereinafter, “similar program pair”) in which two configuration programs are similar based on the program similarity calculated by the second calculating unit 11B. Identify). For example, the first specifying unit 11C specifies the similar program pair based on the magnitude of the program similarity and the “similarity determination threshold”. Specifically, the first specifying unit 11C specifies a program pair having a program similarity equal to or higher than the “similarity determination threshold” as a similar program pair.

  As described above, according to the first embodiment, in the similar program identifying device 10, the first calculating unit 11A accesses the access destination and the access destination for each of the plurality of programs and the plurality of access timing types of each program. For each program pair among a plurality of programs, two accesses respectively corresponding to two configuration programs of the program pair for each access timing type, using an access correspondence relation associated with an access pattern including an access purpose of Calculate the timing type unit similarity between patterns. Then, the second calculating unit 11B determines a plurality of timing type unit similarities calculated for each program pair by the first calculating unit 11A, and group combinations of access destination characteristic groups to which two constituent programs of each program pair belong. Program similarity for each program pair is calculated based on the weight coefficient set corresponding to.

  With the configuration of the similar program identification device 10, the timing type unit similarity calculated for the program pair is weighted by considering the access destination characteristic for each access timing type for the two constituent programs of the program pair. Since the program similarity is calculated, the more accurate program similarity can be calculated.

  Further, in the similar program specifying device 10, the first specifying unit 11C specifies a program pair in which two configuration programs are similar based on the program similarity calculated by the second calculating unit 11B.

  The similar program specifying device 10 specifies the similar program pair based on the more accurate program similarity, so that the specifying accuracy of the similar program can be improved.

Second Embodiment
The second embodiment relates to a more specific embodiment.

<Configuration Example of Similar Program Identification Device>
FIG. 2 is a block diagram showing an example of the similar program identification device of the second embodiment. In FIG. 2, the similar program identification device 20 includes a control unit 21 and a storage unit 12. The control unit 21 includes an information collection unit 21A, a table management unit 21B, and a group combination specification unit 21C, in addition to the timing type unit similarity calculation unit 11A, the program similarity calculation unit 11B, and the similar program specification unit 11C. And a weighting factor identification unit 21D. Hereinafter, the group combination identification unit 21C may be referred to as a "second identification unit", and the weight coefficient identification unit 21D may be referred to as a "third identification unit".

  The information collection unit 21A analyzes the software asset and collects information on elements included in the software asset. The analysis of software assets by the information collection unit 21A can be performed by a general program analysis technique.

  The table management unit 21B updates the "system configuration information table" and the "relation link table" stored in the storage unit 12 using the information collected by the information collection unit 21A. FIG. 3 is a diagram showing an example of a system configuration information table of the second embodiment. FIG. 4 is a view showing an example of a relation link table of the second embodiment. FIG. 5 is a view showing a configuration example of software assets represented by the system configuration information table of FIG. 3 and the relation link table of FIG.

  As shown in FIG. 3, in the system configuration information table, the items "element management number", "element type", "element name", "group by access destination characteristics", "out link", and "in link" as items. "It is included. For example, the entry of the element number E1 indicates that the program 1 is included in the subsystem A and that the link of the link management numbers L1, L2, L3, and L4 is output from the program 1.

  That is, the "element management number" item indicates the number of the entry. Also, “element type” indicates the type of element, and is, for example, a program, a macro, a table, a file, and the like. Also, "element name" is the name of an element and can be used as an identifier of the element.

  The “access destination characteristic classified group” is a program group corresponding to the access destination characteristic for each access timing type, and is, for example, a group for each subsystem or software package. Here, the programs may be grouped into access target characteristic groups, for example, based on naming rules. Further, the correspondence between the program and the access target characteristic group may be collected by the information collection unit 21A from the setting file in which the correspondence is described. In the system configuration information table shown in FIG. 3, as an example, the program is divided into a subsystem A including a general task processing program and a subsystem B which is a common component commonly used in each task processing. , Are grouped.

  In addition, “output link” is a link starting from the element of the entry. Also, “incoming link” is a link whose end point is the element of the entry.

  In addition, as shown in FIG. 4, the relation link table includes, as items, “link management number”, “access purpose type”, “access timing type”, “start point information”, and “end point information”. There is. For example, the link management number L1 entry represents a link in which the element of the element management number E1 refers to the element of the element number E8 at compile time.

  That is, the "link management number" item indicates the number of the link. The “access purpose type” is an object for accessing the element represented by the “end point information” as the element represented by the “start point information”. "Access timing type" represents the type of timing when the element represented by "start point information" accesses the element represented by "end point information". For example, "compile time" and "program execution time" It contains. Also, "start point information" indicates an element that is the start point of the link. Also, "end point information" indicates an element that is the end point of the link.

  The “access correspondence relationship” and the “group correspondence relationship” are represented by the system configuration information table of FIG. 3 and the relationship link table of FIG. 4 described above.

  Here, as described above, the “access correspondence relationship” is a correspondence relationship in which a plurality of programs (that is, a plurality of program identifiers) are associated with access patterns for each of a plurality of access timing types of each program. . Referring to FIGS. 3 and 4, for example, program 1 of element management number E1 has an access pattern for accessing macro a of element management number E8 at the time of compilation for the purpose of citation. The program 1 of the element management number E1 accesses for reading the table A of the element management number E12 at the time of program execution, and accesses for reading the table C of the element management number E14, and the element management number E15 There is an access pattern to be accessed for the purpose of writing to the table D of FIG.

  The “group correspondence relationship” is a correspondence relationship in which a plurality of programs are associated with an access destination characteristic group corresponding to an access destination characteristic for each access timing type to which each program belongs. Referring to FIGS. 3 and 4, for example, programs 1, 2, 2 and 3 are associated with subsystem A, and programs 4, 5 and 5 'are associated with subsystem B. As described above, the subsystem A is a group of general task processing programs, and the subsystem B is a group of common parts commonly used in each task processing. The characteristic of business processing appears in an access pattern (that is, access destination characteristic) at the time of program execution because it tends to appear in a table or file to be accessed. Also, the feature of the common part appears in the access pattern at compile time (that is, the access destination characteristic) because there is less access to business tables and files. The programs 1, 2, 2 ', 3, 4, 5 and 5' are grouped into subsystem A and subsystem B according to the access destination characteristic for each access timing type.

  Returning to the explanation of FIG. 2, the second identification unit (group combination identification unit) 21C uses the above-mentioned “group correspondence relationship” to select 2 of each program pair of the plurality of programs included in the “similar program search range”. Identify a group combination of two configuration programs. For example, in the case of a program pair of program 1 and program 2, the specified group combination is combination AA. Further, in the case of the program pair of program 4 and program 5, the specified group combination is combination B-B. Further, in the case of the program pair of program 3 and program 4, the specified group combination is combination AB.

  The third specifying unit (weighting coefficient specifying unit) 21D specifies a weight coefficient set corresponding to a group combination of each program pair specified by the second specifying unit based on the “weight correspondence relationship”. The “weight correspondence relationship” is a correspondence relationship in which a plurality of group combinations are associated with a weight coefficient set between access timing types according to each group combination. The “weight correspondence relationship” is held in, for example, the storage unit 12 as a weight coefficient table illustrated in FIG.

  For example, in the case of group combination A-A, the weight coefficient to be multiplied with the timing type unit similarity of "at compile time" is "0.2", and for the timing type unit similarity of "at program execution". Thus, a weight coefficient set is identified, which has a weight coefficient to be multiplied by “0.8”. Here, as described above, since the characteristics of subsystem A tend to appear in the access pattern (that is, access destination characteristics) at the time of program execution, in the weighting coefficient set corresponding to group combination A-A, A large value is set to the weight coefficient corresponding to “time”. On the other hand, since the characteristics of subsystem B tend to appear in the access pattern at compile time (that is, access destination characteristics), the weight coefficient set corresponding to group combination B-B is the weight coefficient corresponding to "compile time" Is set to a large value. On the other hand, in the weight coefficient set corresponding to the group combination AB, the weight coefficient corresponding to "compile time" is equal to the weight coefficient corresponding to "program execution time". That is, the weight coefficient set corresponding to the group combination AB is equivalent to the average of the weight coefficient set corresponding to the group combination A-A and the weight coefficient set corresponding to the group combination B-B.

  Then, the second calculating unit (program similarity calculating unit) 11B determines the plurality of timing type unit similarities calculated for each program pair by the first calculating unit 11A and the “weighting coefficient specified by the third specifying unit 21D. Based on the “set”, as described in the first embodiment, the program similarity for each program pair is calculated.

<Operation Example of Similar Program Identification Device>
An example of the processing operation of the similar program identification device having the above configuration will be described. Here, in particular, processing until calculation of the program similarity is described. FIG. 7 is a flowchart showing an example of the processing operation of the similar program identification device of the second embodiment.

  In the similar program specification device 20, the first calculation unit (timing type unit similarity calculation unit) 11A selects one program pair among a plurality of programs included in the “similar program search range” as a program pair of interest. (Step S101). Here, it is assumed that the program pair of program 1 and program 2 is selected as the program pair of interest.

<Calculation processing of timing type unit similarity>
Next, the first calculator 11A calculates the timing type unit similarity of the target program pair (step S102).

  For example, the first calculation unit 11A first counts, for each access timing type, the type of link related to the target program pair. In the case of the program pair of program 1 and program 2, the total number of links whose access timing type is "compile time" is two, the outgoing link of link management number L1 and the outgoing link of link management number L4. The outgoing link is counted as one since the elements of the end point and the access purpose type are the same. The total number of links whose access timing type is "program execution" is six of the outgoing links of link management numbers L2, L3, L4, L6, L7, L8, but the output of link management numbers L3, L7 The links are counted as one because the end elements and the access purpose types are the same. Therefore, in the case of the program pair of program 1 and program 2, there are five types of links whose access timing type is "program execution time". A plurality of incoming links are counted as one when the start point element and the access purpose type are the same.

  Next, the first calculator 11A counts the type of link common to the two configuration programs of the program pair of interest. In the case of the program pair of program 1 and program 2, the link having the access timing type “at compile time” and common to both configuration programs is the outgoing link of link management number L1 and the outgoing link of link management number L4. Also in this case, as in the above case, the outgoing link of the link management number L1 and the outgoing link of the link management number L4 are counted as one since the elements of the end point and the access purpose type are the same. Further, in the case of the program pair of program 1 and program 2, the link having the access timing type “at the time of program execution” and common to both configuration programs is the outgoing link of link management number L3 and the outgoing link of link management number L7. . Also in this case, as in the above, the outgoing link of the link management number L3 and the outgoing link of the link management number L7 are counted as one since the elements of the end point and the access purpose type are the same.

Then, the first calculation unit 11A performs the following calculation to calculate the “at compile time” timing type unit similarity and “at the time of program execution” timing type unit similarity for the program pair of interest. calculate.
"Compile time" timing type unit similarity = 1/1 = 1
At program execution timing type unit similarity = 1/5 = 0.2

<Specific processing of group combination>
Next, the second specification unit (group combination specification unit) 21C specifies a group combination of the two configuration programs of the program pair of interest using the “group correspondence relationship” (step S103). In the case of the program pair of program 1 and program 2, since any of the configuration programs belongs to subsystem A, group combination A-A is specified.

<Specific process of weighting coefficient set>
Next, the third specifying unit (weighting coefficient specifying unit) 21D specifies a weight coefficient set corresponding to the group combination of the program pair of interest specified in step S103 based on the “weight correspondence relationship” (step S104). In the case of the program pair of program 1 and program 2, the weighting coefficient set (0.2, 0.8) corresponding to the group combination A-A is specified in the weighting coefficient table shown in FIG.

<Calculation processing of program similarity>
Next, the second calculating unit (program similarity calculating unit) 11B determines the plurality of timing type unit similarities calculated for the target program pair in step S102 and the “weighting coefficient set” specified in step S104. Program similarity for the target program pair is calculated (step S105). Specifically, in the case of the program pair of program 1 and program 2, the program similarity is calculated as follows.
Program similarity = timing type unit similarity of “at compile time” × weight coefficient of “at compile time” + timing type unit similarity of “at program execution” × weight coefficient of “at program execution” = 1 × 0.2 + 0. 2 × 0.8 = 0.36

  Next, the first calculator 11A determines whether all program pairs in the “similar program search range” have been selected as the target program pair (step S106). If all the program pairs have been selected as the target program pair (YES in step S106), the flow of FIG. 7 ends.

  If all the program pairs have not been selected as the target program pair (NO in step S106), the first calculation unit 11A selects another program pair not selected as the target program pair as the target program pair (step S107). Then, the flow returns to step S102. Then, for the newly selected focused program pair, the processing from step S102 to step S105 described above is performed.

<Calculation result of program similarity>
The calculation result of the program similarity degree calculated according to the above processing procedure will be described. FIG. 8 is a diagram showing calculation results of program similarity. In FIG. 8, “combination” corresponds to the above program pair, and “weighted similarity” corresponds to the above program similarity. In addition, in FIG. 8, “classification similarities”, “compile time similarity”, and “run time similarity” are described for comparison. The "compilation time similarity" corresponds to the above-mentioned "compilation time" timing type unit similarity, and the "execution time similarity" corresponds to the above "program execution time" timing type unit similarity. Also, “classification similarity” is a value obtained by dividing the number of common link types for the program pair by the total number of link types without considering the access timing type. That is, in the case of the program pair of program 1 and program 2, "classification similarities" are calculated by the following formula.
No classification Similarity = (1 + 1) / (1 + 5) = 0.33

Here, the configuration of the software asset shown in FIG. 5 has the following features.
(1) The program included in the subsystem A is a general task processing program, and accesses a table or file associated with each task.
(2) The program included in the subsystem B is a common component commonly used in each business process, and does not access tables and files related to business.
(3) Table C is a table storing messages, in which all the messages in the system are stored. Therefore, it is accessed from all programs.
(4) The program 2 'is obtained by refactoring the program 2, that is, the code which can be used by other functions is cut out as a macro (macros b and c). There is no change in the accessed table because the process is the same in business. These two programs are to be detected as similar programs.
(5) The program 5 'is also a refactoring of the program 5. These two programs are also to be detected as similar programs.

  FIG. 9 is a graph in which program similarity values are plotted against program pair numbers. As shown in FIG. 9, the program similarity of the program pair of combination No. 3 (program 2, 2 ') and the program pair of combination No. 8 (programs 5, 5') to be identified as similar programs This value is larger than the program similarity of the program pair. Therefore, by setting the value of the above-mentioned "similarity determination threshold" to, for example, 0.8, it is possible to accurately specify a program pair whose configuration programs are similar.

  On the other hand, as shown in FIG. 10 to FIG. 12, the configuration programs are similar regardless of which of “classification similarities”, “compile time similarity” and “run time similarity” are used. Unable to identify program pair. FIG. 10 is a graph in which the value of similarity without classification is plotted against the number of program pairs. FIG. 11 is a graph in which compile-time similarity values are plotted against program pair numbers. FIG. 12 is a graph in which the values of runtime similarity are plotted against the numbers of program pairs.

  As described above, according to the second embodiment, in the similar program specifying device 20, the first calculation unit 11A uses the above-mentioned "access correspondence relationship" to set a plurality of programs included in the "similar program search range". For each program pair among them, timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair is calculated for each access timing type. Then, the second specifying unit 21C specifies a group combination of two configuration programs of each program pair of the plurality of programs using the “group correspondence relationship”. Then, the third specifying unit 21D specifies the weighting coefficient set corresponding to the group combination of each program pair specified by the second specifying unit based on the above-mentioned “weight correspondence relationship”. Then, based on the plurality of timing type unit similarities calculated for each program pair by the first calculation unit 11A and the “weighting coefficient set” specified by the third specifying unit 21D, the second calculation unit 11B Program similarity for each program pair is calculated.

  With the configuration of the similar program specification device 20, weighting is performed in consideration of the access destination characteristic for each access timing type for two configuration programs of the program pair, on the timing type unit similarity calculated for the program pair. Since the program similarity is calculated, the more accurate program similarity can be calculated.

<Modification>
The following modification can be applied to the similar program identification device 20 described above.

  <1> The storage unit 12 may additionally hold an excluded link table that defines information on links not used for similarity determination. Then, in the <timing type unit similarity calculation process> described above, the link held in the excluded link table may be excluded from both the link type count target and the common link type count target. In this way, links can be excluded that do not contribute to representing program features because related links are provided from most programs as in the message table. As a result, it is possible to further improve the calculation accuracy of the program similarity.

  <2> In the above description, the weighting coefficient set of group combination AB is previously set in the weighting coefficient table, but it is not limited to this, and the weight of group combination A-A in the weighting coefficient table A third identification unit (weighting coefficient identification unit) 21D may derive from the coefficient set and the weighting coefficient set of the group combination B-B.

Other Embodiments
<1> In the first and second embodiments, the control unit (11, 21) and the storage unit (12) are described as being included in one similar program identification device (10, 20). However, it is not limited to this. For example, the control unit (11, 21) and the storage unit (12) may be included in separate devices.

  <2> FIG. 13 is a diagram showing an example of the hardware configuration of the similar program identification device. In FIG. 13, the similar program identification device 100 includes a processor 101 and a memory 102. The control units 11 and 21 of the similar program identification devices 10 and 20 described in the first embodiment and the second embodiment are realized by the processor 101 reading and executing a program stored in the memory 102. In addition, the storage unit 12 may be realized by the memory 102.

  In addition, this invention is not limited to said embodiment, It is possible to change suitably in the range which does not deviate from the meaning.

10, 20 Similar program identification device 11, 21 Control unit 11A Timing type unit similarity calculation unit (first calculation unit)
11B Program similarity calculation unit (second calculation unit)
11C Similar Program Identification Unit (First Identification Unit)
12 storage unit 21A information collection unit 21B table management unit 21C group combination identification unit (second identification unit)
21D weight coefficient identification unit (third identification unit)

Claims (4)

Of the plurality of programs, using an access correspondence relationship in which a plurality of programs are associated with an access destination for each of a plurality of access timing types of each program and an access pattern including an access purpose to the access destination. A first calculation unit configured to calculate, for each program pair, a timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair for each access timing type;
Based on the plurality of timing type unit similarities calculated for each program pair by the first calculation unit, and the weighting coefficient set corresponding to the group combination of the access destination characteristic group to which the two constituent programs of each program pair belong. A second calculation unit that calculates the program similarity for each program pair;
A first specification unit that specifies a program pair in which two configuration programs are similar based on the program similarity calculated by the second calculation unit;
Similar program identification device equipped with. A group of two constituent programs of each program pair based on a group correspondence relationship in which the plurality of programs are associated with the access destination characteristic group corresponding to the access destination characteristic for each access timing type to which each program belongs. A second identification unit that identifies the combination;
Corresponds to a group combination of each program pair identified by the second identification unit based on a weight correspondence relationship in which a plurality of group combinations are associated with a weight coefficient set between access timing types according to each group combination A third identification unit that identifies a set of weighting coefficients to be
Further equipped,
The second calculation unit sets a plurality of timing type unit similarity degrees calculated for each program pair by the first calculation unit, and a weighting coefficient set corresponding to a group combination of each program pair specified by the third specification unit. Calculate the program similarity for each program pair based on
The similar program identification device according to claim 1. Of the plurality of programs, using an access correspondence relationship in which a plurality of programs are associated with an access destination for each of a plurality of access timing types of each program and an access pattern including an access purpose to the access destination. For each program pair, for each access timing type, calculate timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair,
Each program pair is calculated based on the calculated plurality of timing type unit similarities for each program pair, and a weighting coefficient set corresponding to a combination of access target characteristic groups to which the two constituent programs of each program pair belong. Calculate the program similarity for
The two component programs identify similar program pairs based on the program similarity.
Similar program identification method. Of the plurality of programs, using an access correspondence relationship in which a plurality of programs are associated with an access pattern of each program for each of a plurality of access timing types and an access pattern including an access purpose to the access destination. For each program pair, for each access timing type, calculate timing type unit similarity between two access patterns respectively corresponding to two configuration programs of the program pair,
Each program pair is calculated based on the calculated plurality of timing type unit similarities for each program pair, and a weighting coefficient set corresponding to a combination of access target characteristic groups to which the two constituent programs of each program pair belong. Calculate the program similarity for
The two component programs identify similar program pairs based on the program similarity.
A control program that causes a similar program specification device to execute processing.

Images