Information retrieval device

Abstract

To provide an information retrieval device that can deal with a case of regression.SOLUTION: An information retrieval device includes: a similar problem solution information retrieval part which retrieves similar problem solution information similar to new problem phenomenon information from past problem solution information; a constituent item change information extraction part which extracts past constituent item change information being the changed constituent item information as similar problem solution-time constituent item change information from the similar problem solution information when the past similar problem is solved; a similar problem solution neighborhood constituent item change information extraction part which extracts similar problem solution neighborhood constituent item change information in a neighboring relationship with each other from the similar problem solution-time constituent item change information; and a problem solution information extraction part which extracts constituent item information in a neighboring relationship with the similar problem solution information from the similar problem solution neighborhood constituent item change information, determines the possibility of regression of the corresponding past problem solution information, and outputs the past problem solution information that has the possibility of regression as the similar problem solution neighborhood problem solution information.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to an information retrieval device, and particularly to an information retrieval device that searches for the cause of a new problem phenomenon that occurs with changes in software and / or a system from past problem phenomena. from past problem phenomena.

Background Art

[0002] Due to the multifunctionalization of software, its scale and complexity are increasing. There is also an increasing trend to realize highly value-added functions by constructing a single system through the interlocking of multiple software and devices.

[0003] To reduce the software development cost of products, shortening the lead time of software development is required.

[0004] Especially in derivative development and forked development, while it is necessary to efficiently obtain information on past development, the past information is accumulating and becoming large-scale and complex, making efficient information acquisition increasingly difficult.

[0005] Therefore, for example, when problems such as software failures occur during development, a search mechanism using artificial intelligence (AI) technology or the like has been considered to efficiently obtain reference information from a vast amount of past problem-solving information.

[0006] For example, in Patent Document 1, in order to support the cause investigation when a new problem occurs, a method is disclosed that uses a small difference list obtained by subdividing the difference between a test-failed source code before the problem occurs and a test-passed source code after the problem occurs to identify combinations of small differences that result in test failure.

[0007] At this time, if the cause of the problem lies in the difference information between the test-failed source code and the test-passed source code, it leads to cause identification.

Prior Art Documents

Patent Documents

[0008] [Patent Document 1] Patent No. 6903533 [Overview of the project] [Problems that the invention aims to solve]

[0009] The technology disclosed in Patent Document 1 was unable to address regression cases in complex systems where changes in the external operating environment and other interconnected systems unintentionally caused regressions.

[0010] This disclosure aims to provide an information retrieval device that can handle regression cases. [Means for solving the problem]

[0011] An information retrieval device that searches for the cause of a new problem phenomenon that has newly occurred due to a change in software or system from past problem resolution information that has been resolved in the past and stored in a memory device, the past problem The system includes: a similar problem-solving information search unit that searches for similar problem-solving information similar to new problem phenomenon information from the solution information; a component change information extraction unit that extracts past component change information, which is component change information that was changed when a similar problem was solved in the past, from the similar problem-solving information as similar problem-solving component change information; a similar problem-solving neighbor component change information extraction unit that extracts similar problem-solving neighbor component change information that is in a neighboring relationship with each other from the similar problem-solving neighbor component change information; and a problem-solving information extraction unit that extracts component information that is in a neighboring relationship with the similar problem-solving information from the similar problem-solving neighbor component change information, determines the possibility of regression of the past problem-solving information corresponding to the extracted component information, and outputs the past problem-solving information that has the possibility of regression as similar problem-solving neighbor problem-solving information. [Effects of the Invention]

[0012] The search device described herein can handle regression cases and improve the probability of providing information that can be used as a reference when investigating the causes of new problem phenomena and formulating countermeasures. [Brief explanation of the drawing]

[0013] [Figure 1] This is a block diagram showing the overall configuration of the information retrieval device of Embodiment 1 according to this disclosure. [Figure 2] This figure shows an example of the input screen for the new problem / phenomenon information input section. [Figure 3] This figure shows the past problem-solving information stored in the past problem-solving information storage unit. [Figure 4] This figure shows similar problem-solving information stored in the past problem-solving information storage unit. [Figure 5] This figure shows the correspondence table used by the component change information extraction unit. [Figure 6] This diagram schematically shows the past component change information stored in the past component change information storage unit. [Figure 7] This figure shows the component change information stored in the past component change information storage unit when similar problems were resolved. [Figure 8] This figure shows the nearby component change information for solving similar problems, stored in the past component change information storage unit. [Figure 9] This diagram shows the correspondence table used by the problem-solving information extraction unit. [Figure 10] This figure shows the information on similar problems and neighboring problems stored in the past problem solving information storage unit. [Figure 11] This is a flowchart illustrating the processing in the component change information extraction unit. [Figure 12] This is a flowchart illustrating the processing in the neighboring component change information extraction unit for solving similar problems. [Figure 13] This is a flowchart illustrating the processing in the problem-solving information extraction unit. [Figure 14] This is a block diagram showing the overall configuration of the information retrieval device of Embodiment 2 according to this disclosure. [Figure 15] It is a diagram schematically showing a process sheet input from the process sheet input unit. [Figure 16] It is a diagram showing similar problem-solving neighboring problem-solving information and sorted similar problem-solving neighboring problem-solving information stored in the past problem-solving information storage unit. [Figure 17] It is a flowchart explaining the processing in the problem-solving information confirmation recommendation sorting unit. [Figure 18] It is a block diagram showing a partial configuration of a modified example of the information retrieval apparatus according to Embodiment 2 of the present disclosure. [Figure 19] It is a flowchart explaining the processing in the problem-solving information confirmation recommendation sorting unit. [Figure 20] It is a diagram showing a hardware configuration for realizing the information retrieval apparatuses according to Embodiments 1 and 2 of the present disclosure. [Figure 21] It is a diagram showing a hardware configuration for realizing the information retrieval apparatuses according to Embodiments 1 and 2 of the present disclosure.

Mode for Carrying Out the Invention

[0014] <Embodiment 1> FIG. 1 is a block diagram showing the overall configuration of an information retrieval apparatus 1000 according to Embodiment 1 of the present disclosure. As shown in FIG. 1, the information retrieval apparatus 1000 includes a component item change information extraction unit 10, a similar problem-solving neighboring component item change information extraction unit 20, a problem-solving information extraction unit 30, a past problem-solving information storage unit 50, a past component item change information storage unit 60, a new problem phenomenon information input unit 70, and a similar problem-solving information retrieval unit 100. <F

[0015] The past problem-solving information storage unit 50 stores similar problem-solving information 51 and similar problem-solving neighboring problem-solving information 52, and the past component item change information storage unit 60 stores similar problem-solving component item change information 61 and similar problem-solving neighboring component item change information 62.

[0016] <Similar Problem-Solving Information Retrieval Unit> The similar problem-solving information search unit 100 searches the past problem-solving information stored in the past problem-solving information storage unit 50 for new problem-phenomenon information input from the new problem-phenomenon information input unit 70, and extracts similar problem-solving information 51. New problem-phenomenon information is information about new problem-phenomenons that occur as a result of changes to the software and / or system. Figure 1 shows an example in which a parameter called the problem-phenomenon occurrence process 71 is associated with the new problem-phenomenon information.

[0017] More specifically, the similar problem-solving information search unit 100 performs morphological analysis on the elements of the new problem-phenomenon information obtained from the new problem-phenomenon information input unit 70 and the natural sentences contained in the elements of the past problem-solving information stored in the past problem-solving information storage unit 50, i.e., information on problem-phenomena that have been solved in the past, and converts them into word sequences. Then, it compares the word sequences obtained from each element of the new problem-phenomenon information and the past problem-solving information, and if the degree of similarity, i.e., the sentence similarity, is sufficiently large, it determines that the element of the past problem-solving information is similar to the element of the new problem-phenomenon information and extracts it as similar problem-solving information 51.

[0018] For example, if the degree of similarity is 90%, it will be extracted as similar problem-solving information 51, and if it is 50%, it will not be extracted. Alternatively, if the degree of similarity is 90% or higher, it will be extracted, and if it is less than 90%, it will not be extracted, but this is not the only option.

[0019] As explained above, by using morphological analysis, it is relatively easy to search for past problem-solving information and extract similar problem-solving information 51.

[0020] Figure 2 shows an example of the input screen for the new problem phenomenon information input unit 70. The new problem phenomenon information input from the new problem phenomenon information input unit 70 is a string representing a newly occurring problem phenomenon, and is entered as a search string for searching for useful information from past problem resolution information.

[0021] In the example in Figure 2, the search box for new problem phenomena against past problem-solving information has "The state of AA does not change when the XX process is executed" entered. Also, "Process Y1" is selected from the pull-down menu as the process in which the problem phenomenon occurs.

[0022] Figure 3 shows the past problem-solving information stored in the past problem-solving information storage unit 50. As shown in Figure 3, the past problem-solving information consists of a "number (No) column" for identifying problems that have occurred in the past, and a "content column" that shows the content and solution information of similar problems that have occurred in the past.

[0023] In the example in Figure 3, "1357" and "1358" are given as example identification numbers. Identification number 1357 describes an example of a problem phenomenon, "An error occurs when executing XXX," stating, "When XXX is executed under the conditions of BB, an internal error YY occurs, and processing stops." Identification number 1358 describes an example of a problem phenomenon, "The execution time of the XXX process is too long," stating, "The execution time of the XXX process is ZZ seconds, which does not meet the requirements."

[0024] Figure 4 shows the similar problem-solving information 51 stored in the past problem-solving information storage unit 50. As shown in Figure 4, the similar problem-solving information 51 consists of a "No. column" for identifying problems that have occurred in the past, a "search result column" showing the content and solution information of similar problems that have occurred in the past, a "text similarity column" showing the degree of similarity between new problem phenomenon information and the content described in the "search result column", and a "work process column" showing the process in which the problem occurred.

[0025] The "No. column" and "Search Results column" correspond to the "No. column" and "Content column" of the past problem-solving information shown in Figure 3, and the "Text Similarity column" shows the degree of agreement between word sequences obtained by morphological analysis of the natural text of each element of the new problem phenomenon information and the past problem-solving information performed by the similar problem-solving information search unit 100, i.e., text similarity.

[0026] In the example in Figure 4, for identification number 1357, the text similarity is "0.9" and the work process is "X1 process", while for identification number 1358, the text similarity is "0.85" and the work process is "Y1 process".

[0027] <Component Change Information Extraction Unit> The component change information extraction unit 10 reads the similar problem resolution information 51 stored in the past problem resolution information storage unit 50 and extracts the component change information 61 from when a similar problem was resolved using the correspondence table 80. The component change information 61 from when a similar problem was resolved is information about the component items that were changed when a similar problem that occurred in the past was resolved. The processing flow of the component change information extraction unit 10 will be explained later.

[0028] Figure 5 shows the correspondence table 80 used by the component change information extraction unit 10. The correspondence table 80 used by the component change information extraction unit 10 is a table that associates changed component information with each item of past similar problem solving information, and is pre-entered into the information retrieval device 1000 and stored in a storage device not shown.

[0029] In the example in Figure 5, the table is structured so that the "No. column" of the similar problem-solving information 51 shown in Figure 4 is associated with an identification number (CommitID) that uniquely identifies the "Component Change Information (Commit)" which holds the changed component information.

[0030] A Commit contains the data of the component's files before and after the change, as well as information representing the differences between the before and after changes. A Commit is read from a storage device (not shown) that stores data related to software or system change operations.

[0031] In the example in Figure 5, CommitID "1234" is associated with identification number 1357, and CommitID "1235" is associated with identification number 1358.

[0032] Figure 6 schematically shows the past component change information stored in the past component change information storage unit 60. The past component change information contains the component information changed in each commit, for each CommitID listed in the correspondence table 80 used by the component change information extraction unit 10.

[0033] In the example in Figure 6, the changes in Commit 1234 and Commit 7654 are schematically shown by file markers, with hatched file markers indicating the changes in both cases. Note that Figure 6 only shows Commit 1234 and Commit 7654, but this is merely an example.

[0034] Figure 7 shows the component change information 61 stored in the past component change information storage unit 60 when a similar problem was resolved. The component change information 61 when a similar problem was resolved holds detailed information of the commits listed in the correspondence table 80 used by the component change information extraction unit 10 shown in Figure 5. In the example in Figure 7, it holds information on component items registered as change information in Commit 1234, such as the source code group for implementing the software.

[0035] <Similar Problem Solving and Nearby Component Change Information Extraction Unit> The similar problem-solving nearby component change information extraction unit 20 reads the similar problem-solving component change information 61 stored in the past component change information storage unit 60 and extracts the similar problem-solving nearby component change information 62.

[0036] Figure 8 shows the similar problem-solving neighboring component change information 62 stored in the past component change information storage unit 60. The similar problem-solving neighboring component change information 62 is information that is in a neighboring relationship with each other among the past component change information. In the example in Figure 8, the similar problem-solving neighboring component change information 62 shows the change location 621 of Commit 1234 and the change location 622 of Commit 7654. This indicates that Commit 7654 is in the vicinity of Commit 1234 in the similar problem-solving information 51.

[0037] Figure 8 shows that in Commit 1234, change location 621 indicates that component group a was changed, and in Commit 7654, change location 622 indicates that component group b was changed. Component group a in Commit 1234, change location 621, and component group b in Commit 7654, change location 622, are neighbors of each other.

[0038] The Similar Problem Solving Nearby Component Change Information Extraction Unit 20 determines that components are neighbors if the proportion of components in component group a that were changed in Commit 1234 and component group b that were changed in Commit 7654 is equal to or greater than a threshold, and extracts them as Similar Problem Solving Nearby Component Change Information 62 from past component change information that are in a neighboring relationship with each other.

[0039] More specifically, as shown in Figure 6, the component items of multiple past component item change information from Commit 1234 to Commit 7654, stored in the past component item change information storage unit 60, are compared exhaustively to calculate the percentage of matching component items.

[0040] For example, if the component item is "source code for implementing software," even though it's the same source code, there are several possible units for the component item, such as "file," "function," or "line count." In the case of "function," for example, the "percentage of functions that were changed in Commit 1234 that were also changed in Commit 7645" is calculated and compared with a threshold to determine if they are neighbors. The processing in the similar problem solving neighboring component item change information extraction unit 20 will be explained further later.

[0041] As explained above, by comparing all the components of multiple past component change information, similar problem-solving neighboring component change information 62 that are in a neighboring relationship with each other can be extracted, thereby obtaining more accurate similar problem-solving neighboring component change information 62.

[0042] <Problem solving information extraction part> The problem-solving information extraction unit 30 reads the similar problem-solving neighboring component change information 62 stored in the past component change information storage unit 60 and extracts similar problem-solving neighboring problem-solving information 52 using the correspondence table 90.

[0043] Figure 9 shows the correspondence table 90 used by the problem-solving information extraction unit 30. The correspondence table 90 used by the problem-solving information extraction unit 30 is a table that associates a "No column" indicating past problem-solving information identified as a neighbor of similar problem-solving information 51 with a CommitID that uniquely identifies a Commit that holds information on components that were changed in order to solve a past problem identified as a neighbor of similar problem-solving information 51. The table is pre-entered into the information retrieval device 1000 and stored in a storage device not shown.

[0044] In the example in Figure 9, CommitID "7654" is associated with identification number 5123. Also, CommitID "9876" is associated with identification number 6987.

[0045] Figure 10 shows the similar problem-solving neighbor problem-solving information 52 stored in the past problem-solving information storage unit 50. As shown in Figure 10, the similar problem-solving neighbor problem-solving information 52 consists of a No. column 521 indicating past problem-solving information identified as a neighbor of the similar problem-solving information 51, a regression possibility search result column 522 indicating the possibility of regression, which is determined to be the cause of new problem phenomenon information input from the new problem phenomenon information input unit 70, an influence range similarity column 523 indicating the similarity with the new problem phenomenon information, and a work process column 524 indicating the work process in which the similar problem-solving neighbor problem occurred.

[0046] The similar problem-solving neighbor problem-solving information 52 is linked to elements in the similar problem-solving information 51 whose "No" matches the No column 521, and the "Content column" of the past problem-solving information becomes the regression possibility search result column 522.

[0047] In the example in Figure 10, it is shown that item 5123 is a possible regression that may cause identification number 1357. Furthermore, the problem phenomenon "garbled characters are displayed on screens A, B, and C" is described as "garbled Japanese characters when A, B, and C are displayed." The similarity of the scope of influence of this regression is given as "0.9," and the work process is identified as "process Y1."

[0048] Furthermore, in the example in Figure 10, it is shown that identification number 6987 is another item that may be a regression causing identification number 1357. Also, as an example of the problem phenomenon, "A communication error occurs in AAA," it is stated that "In the case of AAA, a communication error occurs." The similarity of the scope of influence of this regression is given as "0.85," and the work process is given as "Process Z1."

[0049] As explained above, the similar problem solving proximity problem solving information 52 is a search result for regression potential among past problem solving information that corresponds to component information in a neighboring relationship with the component information that was changed when a similar problem was solved in the past, and which has a potential regression potential.

[0050] The similar problem-solving and nearby problem-solving information 52 is presented to the user of the information retrieval device 1000 when the table shown in Figure 10 is displayed on the screen of an application or browser. The same applies to the past problem-solving information shown in Figure 3 and the similar problem-solving information 51 shown in Figure 4.

[0051] The determination that the cause of new problem phenomenon information is a regression is made by comparing the changed component groups in the similar problem-solving nearby component change information extraction unit 20, based on the percentage of matching changed component groups. For example, if the percentage of matching changed component groups is 90%, it is determined to be a regression; if it is 50%, it is determined not to be a regression. Alternatively, if the percentage of matching changed component groups is 90% or more, it is determined to be a regression; if it is less than 90%, it is determined not to be a regression, but this is not the only way to determine this.

[0052] As explained above, in the comparison of changed component groups in the similar problem solving nearby component change information extraction unit 20, it is possible to determine relatively easily whether the cause of the new problem phenomenon information is a regression by determining the percentage of matching changed component groups.

[0053] Furthermore, to determine whether the past problem information similar problem-solving information 51 is in the vicinity of the past problem-solving information, morphological analysis is used to compare the word sequences obtained from each element of the past problem-solving information and the similar problem-solving information 51. If the sentence similarity is sufficiently high, the past problem-solving information is judged to be in the vicinity of the similar problem-solving information 51. The criteria for judging sentence similarity can be the same as the criteria for judging sentence similarity used in the similar problem-solving information search unit 100 described earlier.

[0054] As explained above, by using morphological analysis, it is relatively easy to determine whether the past problem information similar to the problem-solving information 51 is in the vicinity of the past problem-solving information.

[0055] The influence scope similarity is the similarity between the new problem phenomenon information input from the new problem phenomenon information input unit 70 and the regression possibility search results, and can be calculated as the degree of agreement between word sequences obtained by morphological analysis. By using morphological analysis, the influence scope similarity can be obtained relatively easily.

[0056] Figure 11 is a flowchart illustrating the processing in the component item change information extraction unit 10. As shown in Figure 11, first, in step S101, the component item change information extraction unit 10 reads out the similar problem solving information 51 (Figure 4) stored in the past problem solving information storage unit 50.

[0057] Next, the component item change information extraction unit 10 reads the correspondence table 80 (Figure 5) (S102).

[0058] Next, the component item change information extraction unit 10 extracts the component item change information 61 corresponding to the similar problem resolution information by extracting items that correspond to the similar problem resolution information from the correspondence table 80 (S103), and then terminates the series of processes.

[0059] Figure 12 is a flowchart illustrating the processing in the Similar Problem Resolution Nearby Component Change Information Extraction Unit 20. As shown in Figure 12, first, in step S201, the Similar Problem Resolution Nearby Component Change Information Extraction Unit 20 reads the Similar Problem Resolution Time Component Change Information 61 stored in the Past Component Change Information Storage Unit 60.

[0060] Next, the Similar Problem Resolution Nearby Component Change Information Extraction Unit 20 extracts a change location 622 (Figure 8) that is near the change location 621 (Figure 8), i.e., the Similar Problem Resolution Nearby Component Change Information 622, from the past component change information stored in the past component change information storage unit 60 (step S202), and terminates the series of processes.

[0061] Figure 13 is a flowchart illustrating the processing in the problem-solving information extraction unit 30. As shown in Figure 13, first, in step S301, the problem-solving information extraction unit 30 reads out the similar problem-solving nearby component change information 622 stored in the past component change information storage unit 60.

[0062] Next, the problem-solving information extraction unit 30 reads the correspondence table 90 (Figure 9) (S302).

[0063] Next, the problem-solving information extraction unit 30 extracts the item corresponding to the similar problem-solving nearby component change information 622 of the similar problem-solving nearby component change information 622 stored in the past component change information storage unit 60 from the correspondence table 90, the item with identification number 5123 in the example of Figure 9, thereby extracting the similar problem-solving nearby component change information 52 that corresponds to the similar problem-solving nearby component change information 622 (S303), and the series of processes ends.

[0064] <Embodiment 2> Figure 14 is a block diagram showing the overall configuration of the information retrieval device 2000 of Embodiment 2 according to this disclosure. As shown in Figure 2, the information retrieval device 2000 has, in addition to the configuration of the information retrieval device 1000 shown in Figure 1, a problem-solving information confirmation recommendation sorting unit 40 and a process chart input unit 200, and is configured to store sorted similar problem-solving neighbor problem-solving information 53 in a past problem-solving information storage unit 50. Other components identical to those of the information retrieval device 1000 described using Figure 1 are denoted by the same reference numerals, and redundant explanations are omitted.

[0065] As shown in Figure 14, the problem-solving information confirmation recommendation sorting unit 40 generates sorted similar problem-solving neighbor problem-solving information 53, sorted in the order recommended for confirmation in order to solve the new problem, based on the new problem-phenomenon information input unit 70, the similar problem-solving neighbor problem-solving information 52 stored in the past problem-solving information storage unit 50, and the process schedule (work process information) input unit 200.

[0066] Figure 15 schematically shows the process schedule input from the process schedule input unit 200. As shown in Figure 15, the process schedule is such that the "Y1 process" is executed in parallel with the consecutive "X1 process" and "X2 process" for the same period, and then the "Z1 process", which has the same end date as the "Y1 process", starts executing partway through the "Y1 process".

[0067] Figure 16 shows the similar problem-solving neighbor problem-solving information 52 and sorted similar problem-solving neighbor problem-solving information 53 stored in the past problem-solving information storage unit 50. As shown in Figure 16, the sorted similar problem-solving neighbor problem-solving information 53 is generated by rearranging the similar problem-solving neighbor problem-solving information 52 based on the process table input unit 200.

[0068] Figure 16 shows an example of generating sorted similar problem-solving neighbor problem-solving information 53 by rearranging the similar problem-solving neighbor problem-solving information 52 shown in Figure 10 in the recommended order based on the process chart shown in Figure 15. Here, one example of a recommended order to check is to prioritize items with a high relationship between new problem phenomenon information and processes, i.e., similarity of impact scope, in order to prioritize regressions that are difficult to notice. However, here, priority is given to items of work processes that have a low relationship with the work processes of the new problem phenomenon information. Examples of evaluating the relationship between processes include considering the relationship to be low if the departments responsible for the processes are different, or considering the relationship to be low if the process relates to a derivative model rather than the same model.

[0069] In the example in Figure 16, when the work process where the problem phenomenon occurs is process Y1, and an algorithm is applied that prioritizes items with low process relationships, the priority of identification number 5123, which has a high process relationship (i.e., an influence scope similarity of 0.9) even though it is process Y1, decreases, while the priority of identification number 6987, which has a low process relationship (i.e., an influence scope similarity of 0.85) increases even though it is process Z1. Therefore, in the sorted similar problem-solving neighbor problem-solving information 53, identification number 6987 has a higher priority.

[0070] In this way, by rearranging the similar problem-solving and neighboring problem-solving information 52 in an order recommended for checking to resolve new problem phenomena based on the similarity of the scope of influence, regressions that are difficult to notice become easier to identify, for example.

[0071] Furthermore, by rearranging the similarity range information 52 so that smaller impact range similarities are ranked higher, regressions with small impact range similarities that are difficult to notice become easier to identify.

[0072] Figure 17 is a flowchart illustrating the processing in the problem-solving information confirmation recommendation sorting unit 40. As shown in Figure 17, first, in step S401, the problem-solving information confirmation recommendation sorting unit 40 reads out the similar problem-solving proximity problem-solving information 52 (Figure 10) stored in the past problem-solving information storage unit 50.

[0073] Next, the problem-solving information confirmation recommendation sorting unit 40 reads out the new problem phenomenon information input from the new problem phenomenon information input unit 70 (S402).

[0074] Next, the problem-solving information confirmation recommendation sorting unit 40 reads the process schedule (Figure 15) entered from the process schedule input unit 200 (S403).

[0075] Next, the problem-solving information confirmation recommendation sorting unit 40 generates sorted similar problem-solving neighbor problem-solving information 53 by rearranging the similar problem-solving neighbor problem-solving information 52 in the order recommended for confirmation based on the new problem phenomenon information and the process chart (step S404), and then terminates the series of processes.

[0076] <Variation> Figure 18 is a block diagram showing a partial configuration of a modified version of the information retrieval device 2000 of Embodiment 2 of the present disclosure. In this modified version, as shown in Figure 18, it further includes a cost database 210 of past problems that were required to solve problems, and the problem-solving information confirmation recommendation sorting unit 40 generates sorted similar problem-solving neighbor problem-solving information 54, which includes the costs required to obtain the similar problem-solving neighbor problem-solving information 52, instead of generating sorted similar problem-solving neighbor problem-solving information 53 as shown in Figure 14.

[0077] The costs incurred to obtain the similar problem solving and neighborhood problem solving information 52 include, for example, the modification costs 541 incurred for modifying the source code for solving the neighborhood problem and the evaluation costs 542 incurred for evaluating the operation after the source code modification.

[0078] Figure 18 shows an example in which the similar problem-solving neighbor problem-solving information 52 shown in Figure 10 is rearranged in the recommended order based on the process chart shown in Figure 15, and the correction cost 541 and evaluation cost 542 are read from the cost database 210 of past problem-solving and added to the items with identification numbers 5123 and 6987 to generate sorted similar problem-solving neighbor problem-solving information 54.

[0079] In the example in Figure 18, the item with identification number 6987 has a modified cost of 541, which is set to "100 person-hours," and an evaluation cost of 542, which is set to "200 person-hours." Similarly, the item with identification number 512 has a modified cost of 541, which is set to "20 person-hours," and an evaluation cost of 542, which is set to "50 person-hours."

[0080] Here, "person-hours" is a unit representing the number of people multiplied by the amount of work. 100 people for 1 hour, 50 people for 2 hours, and 1 person for 100 hours all equal "100 person-hours."

[0081] By adding correction cost and evaluation cost as parameters to the algorithm that determines the recommended order for review, items with higher costs can be given higher priority, thereby increasing the priority level of more important items.

[0082] Figure 19 is a flowchart illustrating the processing in the problem-solving information confirmation recommendation sorting unit 40. As shown in Figure 19, first, in step S401, the problem-solving information confirmation recommendation sorting unit 40 reads out the similar problem-solving proximity problem-solving information 52 (Figure 10) stored in the past problem-solving information storage unit 50.

[0083] Next, the problem-solving information confirmation recommendation sorting unit 40 reads out the new problem phenomenon information input from the new problem phenomenon information input unit 70 (S402).

[0084] Next, the problem-solving information confirmation recommendation sorting unit 40 reads the process schedule (Figure 15) entered from the process schedule input unit 200 (S403).

[0085] Next, the problem-solving information confirmation recommendation sorting unit 40 reads the correction cost 541 and the evaluation cost 542 from the cost database 210 of past problem-solving costs (S405).

[0086] Next, the problem-solving information confirmation recommendation sorting unit 40 rearranges the similar problem-solving neighbor problem-solving information 52 in an order recommended for confirmation based on the new problem phenomenon information and the process chart, and generates sorted similar problem-solving neighbor problem-solving information 53 by adding the correction cost 541 and the evaluation cost 542 (step S406), and ends the series of processes.

[0087] Furthermore, each component of the information retrieval devices 1000 and 2000 according to the embodiments 1 and 2 described above can be configured using a computer, and are realized by the computer executing a program. Specifically, the component item change information extraction unit 10, the similar problem solving nearby component item change information extraction unit 20, the problem solving information extraction unit 30, the problem solving information confirmation recommendation sort unit 40, and the similar problem solving information retrieval unit 100 of the information retrieval devices 1000 and 2000 are realized by a processing circuit 500, for example, shown in Figure 20. A processor such as a CPU or DSP (Digital Signal Processor) is applied to the processing circuit 500, and the functions of each unit are realized by executing a program stored in a memory device.

[0088] Furthermore, dedicated hardware can also be applied to the processing circuit 500. When the processing circuit 500 is dedicated hardware, it may include, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof.

[0089] The component item change information extraction unit 10, the similar problem solving nearby component item change information extraction unit 20, the problem solving information extraction unit 30, the problem solving information confirmation recommendation sorting unit 40, and the similar problem solving information search unit 100 of the information retrieval devices 1000 and 2000 can each be implemented in separate processing circuits, or these functions can be implemented together in a single processing circuit.

[0090] Figure 21 also shows the hardware configuration when the processing circuit 500 is configured using a processor. In this case, the functions of the component item change information extraction unit 10, the similar problem solving nearby component item change information extraction unit 20, the problem solving information extraction unit 30, the problem solving information confirmation recommendation sort unit 40, and the similar problem solving information retrieval unit 100 of the information retrieval devices 1000 and 2000 are realized by a combination of software, firmware, or software and firmware. The software is written as a program and stored in memory 520. The processor 510, which functions as the processing circuit 500, realizes the functions of each part by reading and executing the program stored in memory 520 (storage device). In other words, this program can be said to cause the computer to execute the procedures and methods of operation of the components of the information retrieval devices 1000 and 2000.

[0091] Here, memory 520 can be, for example, non-volatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), HDD (Hard Disk Drive), magnetic disk, flexible disk, optical disk, compact disk, minidisc, DVD (Digital Versatile Disc) and its drive device, or any storage medium that will be used in the future.

[0092] Furthermore, the memory 520 can function as a past problem resolution information storage unit 50 and a past component change information storage unit 60 for the information retrieval devices 1000 and 2000.

[0093] The above describes a configuration in which the functions of each component of the information retrieval devices 1000 and 2000 are realized by either hardware or software. However, this is not the only configuration; it is also possible to configure the information retrieval devices 1000 and 2000 in which some components are realized by dedicated hardware and other components are realized by software. For example, some components can be realized by a processing circuit 500 as dedicated hardware, while other components can be realized by a processing circuit 500 acting as a processor 510 reading and executing a program stored in memory 520.

[0094] As described above, the information retrieval devices 1000 and 2000 can realize the above-mentioned functions through hardware, software, or a combination thereof, and by connecting the new problem phenomenon information input unit 70 as a user interface, the information retrieval devices 1000 and 2000 become available for use by users.

[0095] Within the scope of this disclosure, it is possible to freely combine the embodiments, or modify or omit the embodiments as appropriate.

[0096] The above-described disclosure is summarized below as an appendix.

[0097] (Note 1) An information retrieval device that searches for the cause of a new problem phenomenon that arises as a result of software and / or system changes from past problem resolution information stored in a memory device, A similar problem-solving information search unit searches for similar problem-solving information that is similar to the new problem phenomenon information from the aforementioned past problem-solving information, From the aforementioned similar problem-solving information, a component change information extraction unit extracts past component change information, which is component change information that was changed when a similar problem was solved in the past, as component change information at the time of similar problem-solving. A similar problem-solving neighboring component change information extraction unit extracts similar problem-solving neighboring component change information that is in a close relationship with each other from the aforementioned similar problem-solving component change information, An information retrieval device comprising: a problem-solving information extraction unit that extracts component information that is in a proximity relationship with the similar problem-solving information from the similar problem-solving proximity component change information, determines the possibility of regression of the past problem-solving information corresponding to the extracted component information, and outputs the past problem-solving information that has the possibility of regression as similar problem-solving proximity problem-solving information.

[0098] (Note 2) The information retrieval device according to Appendix 1 further comprises a problem-solving information confirmation recommendation sorting unit that generates sorted similar problem-solving nearby problem-solving information by rearranging the similar problem-solving nearby problem-solving information in an order recommended for confirmation in order to resolve the new problem-solving phenomenon, based on the similar problem-solving nearby problem-solving information, the new problem-phenomenon information, and work process information.

[0099] (Note 3) The aforementioned problem-solving information confirmation recommendation sorting unit The information retrieval device described in Appendix 2 generates sorted similar problem-solving and neighboring problem-solving information by adding the correction costs incurred for solving past problems and the evaluation costs incurred for evaluating the operation after correction.

[0100] (Note 4) The aforementioned similar problem-solving information retrieval unit is: The aforementioned past problem The information retrieval device described in Appendix 1, which performs morphological analysis on natural sentences contained in the solution information and the new problem phenomenon information, converts them into word sequences, compares the word sequences with each other to determine sentence similarity, and determines similar problem solution information based on the sentence similarity.

[0101] (Note 5) The aforementioned similar problem solving nearby component item change information extraction unit The information retrieval device described in Appendix 1 compares all changed component items in multiple past component item change information, determines whether they are neighbors based on the proportion of matching changed component items, and extracts them as similar problem-solving neighbor component item change information if they are neighbors.

[0102] (Note 6) The aforementioned problem-solving information extraction unit, The information retrieval device described in Appendix 1, which performs morphological analysis on natural sentences contained in the aforementioned past problem-solving information and the aforementioned similar problem-solving information to convert them into word sequences, compares the respective word sequences to determine sentence similarity, and extracts component item information that is in a neighboring relationship with the aforementioned similar problem-solving information based on the sentence similarity.

[0103] (Note 7) The aforementioned problem-solving information extraction unit, The information retrieval device described in Appendix 1, which, in the comparison of the changed component items in the similar problem solving nearby component item change information extraction unit, determines the possibility of regression based on the matching rate of the changed component items.

[0104] (Note 8) The aforementioned information for solving similar problems and neighboring problems is, This includes the influence scope similarity, which is the similarity between the past problem resolution information and the new problem phenomenon information that may be causing the regression, The aforementioned problem-solving information confirmation recommendation sorting unit The information retrieval device described in Appendix 2 rearranges the information on similar problem solutions and neighboring problem solutions so that the order in which they are ranked is such that the smaller the similarity in the range of influence, the higher the ranking, as this is the recommended order for checking in order to resolve the aforementioned new problem phenomenon.

[0105] (Note 9) The aforementioned problem-solving information extraction unit, The information retrieval device described in Appendix 8, which performs morphological analysis on natural sentences contained in the past problem-solving information and the new problem-phenomenon information that may contain regressions, converts them into word sequences, compares the respective word sequences to determine sentence similarity, and uses the sentence similarity as the influence range similarity. [Explanation of Symbols]

[0106] 10 Component change information extraction unit, 20 Similar problem solving nearby component change information extraction unit, 30 Problem solving information extraction unit, 40 Problem solving information confirmation recommendation sort unit, 51 Similar problem solving information, 52 Similar problem solving nearby problem solving information, 53 Sorted similar problem solving nearby problem solving information, 61 Component change information at the time of similar problem solving, 62 Similar problem solving nearby component change information, 100 Similar problem solving information search unit, 541 Correction cost, 542 Evaluation cost.

Claims

1. An information retrieval device that searches for the cause of a new problem phenomenon that arises as a result of a change in software or system from past problem resolution information stored in a memory device, A similar problem-solving information search unit searches for similar problem-solving information that is similar to the new problem phenomenon information from the aforementioned past problem-solving information, From the aforementioned similar problem-solving information, a component change information extraction unit extracts past component change information, which is component change information that was changed when a similar problem was solved in the past, as component change information at the time of similar problem-solving. A similar problem-solving neighboring component change information extraction unit extracts similar problem-solving neighboring component change information that is in a close relationship with each other from the component change information when similar problems are solved, An information retrieval device comprising: a problem-solving information extraction unit that extracts component information that is in a proximity relationship with the similar problem-solving information from the similar problem-solving proximity component change information, determines the possibility of regression of the past problem-solving information corresponding to the extracted component information, and outputs the past problem-solving information that has the possibility of regression as similar problem-solving proximity problem-solving information.

2. The information retrieval device according to claim 1, further comprising a problem-solving information confirmation recommendation sorting unit that generates sorted similar problem-solving nearby problem-solving information by rearranging the similar problem-solving nearby problem-solving information in an order recommended for confirmation in order to resolve the new problem-solving phenomenon, based on the similar problem-solving nearby problem-solving information, the new problem-phenomenon information, and work process information.

3. The aforementioned problem-solving information confirmation recommendation sorting unit The information retrieval device according to claim 2, which generates sorted similar problem-solving neighbor problem-solving information by adding the correction cost required to solve past problems and the evaluation cost required to evaluate the operation after correction.

4. The aforementioned similar problem-solving information retrieval unit is: The information retrieval device according to claim 1, comprising: performing morphological analysis on natural sentences contained in the past problem-solving information and the new problem-phenomenon information, converting them into word sequences; comparing the respective word sequences to determine sentence similarity; and determining similar problem-solving information based on the sentence similarity.

5. The aforementioned similar problem solving nearby component item change information extraction unit The information retrieval device according to claim 1, which compares all changed component items in a brute-force manner with multiple past component item change information, determines whether they are neighbors based on the proportion of matching changed component items, and extracts them as similar problem-solving neighbor component item change information if they are neighbors.

6. The aforementioned problem-solving information extraction unit, The information retrieval device according to claim 1, comprising: performing morphological analysis on natural sentences contained in the past problem-solving information and the similar problem-solving information, converting them into word sequences; comparing the respective word sequences to determine sentence similarity; and extracting component information that is in a neighboring relationship with the similar problem-solving information based on the sentence similarity.

7. The aforementioned problem-solving information extraction unit, The information retrieval device according to claim 1, wherein in the comparison of the changed component items in the similar problem solving nearby component item change information extraction unit, the possibility of regression is determined by the matching rate of the changed component items.

8. The aforementioned information for solving similar problems and neighboring problems is, This includes the influence scope similarity, which is the similarity between the past problem resolution information and the new problem phenomenon information that may be causing the regression, The aforementioned problem-solving information confirmation recommendation sorting unit The information retrieval device according to claim 2, which rearranges the similar problem solving and nearby problem solving information so that the smaller the influence range similarity, the higher the ranking, as an order in which it is recommended to check for the resolution of the aforementioned new problem phenomenon.

9. The aforementioned problem-solving information extraction unit, The information retrieval device according to claim 8, comprising: performing morphological analysis on natural sentences contained in the past problem-solving information and the new problem phenomenon information that may contain regressions, converting them into word sequences; comparing the respective word sequences to determine sentence similarity; and using the sentence similarity as the influence range similarity.

Images