Business support apparatus, business support method, and business support program

Abstract

To support the monitoring of fraudulent accounting by switching account items.SOLUTION: A cluster generation unit assigns a plurality of pieces of recorded data, which are recorded for each task and include at least an account item and a recorded amount, to a cluster that is a plurality of groups, according to the similarity between pieces of the recorded data. A detection unit detects, for each cluster, rare cost items, which are account items whose amount composition ratio that is the ratio of the recorded amount of each recorded data to the total amount of the recorded amounts of the recorded data for the entire cluster, is below a cost determination ratio threshold. An abnormality determination unit determines, as an abnormality, the recorded amount for the rare cost item that is outside a predetermined amount range. A display control unit then controls the display of the abnormality determination result data, including the total amount determined to be abnormal, the task, and the cluster to which the work belongs, on a display unit. Thus, the monitoring of fraudulent accounting by switching account items can be supported.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a business support device, a business support method, and a business support program. [Background technology]

[0002] Patent Document 1 (JP 2019-67086 A) discloses a financial analysis device that can accurately and efficiently detect anomalies contained in accounting data. In this financial analysis device, a first vector generation unit generates a first vector whose elements are the fluctuation values ​​of each account item within a first period of the accounting data. An estimation unit estimates the fluctuation values ​​of each account item within the first period based on multiple first vectors within a second period that includes multiple first periods.

[0003] The residual detection unit detects the residual between the fluctuation value and the actual fluctuation value, and the anomaly candidate identification unit extracts the fluctuation value of a specific account item in a specific first period where a value correlated with the residual exceeds a threshold. The journal entry limitation unit generates a second matrix in which second vectors, each of which has as elements the fluctuation values ​​of multiple account items for each journal entry in the specific first period, are arranged row-wise. The journal entry extraction unit extracts journal entries from the second matrix that include account items where a value correlated with the residual exceeds a threshold. The anomaly detection unit then detects anomalies contained in the extracted journal entries, and the anomalous journal entry extraction unit extracts journal entries in which anomalies are detected. This makes it possible to accurately and efficiently detect anomalies contained in accounting data. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2019-67086 Summary of the Invention [Problem to be solved by the invention]

[0005] In this case, fraudulent entries may be made by switching account items. This requires accounting staff to rely on their own experience to detect fraudulent entries, which makes monitoring difficult.

[0006] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a business support device, a business support method, and a business support program that can support the work of monitoring fraudulent accounting by switching account items. [Means for solving the problem]

[0007] In order to solve the above-mentioned problems and achieve the objectives, the business support device of the present invention has a cluster generation unit that assigns multiple accounting data, including at least account items and accounting amounts, recorded for each task, to multiple groups, or clusters, according to the similarity between the accounting data; a detection unit that detects, for each cluster, rare cost items, which are account items whose amount composition ratio, which is the ratio of the accounting amount of each accounting data to the total amount of the accounting amounts of the accounting data for the entire cluster, is below a cost determination ratio threshold; an abnormality determination unit that determines, as abnormal, an accounting amount for a rare cost item that is outside a specified amount range; and a display control unit that controls the display of abnormality determination result data, including the accounting amount, task, and cluster to which the task belongs, that are determined to be abnormal, on a display unit.

[0008] In addition, in order to solve the above-mentioned problems and achieve the objective, the business support method of the present invention includes a cluster generation step in which a cluster generation unit assigns multiple recorded data, including at least account items and recorded amounts, recorded for each task, to multiple groups, or clusters, according to the similarity between the recorded data; a detection step in which a detection unit detects, for each cluster, rare cost items, which are account items whose amount composition ratio, which is the ratio of the recorded amount of each recorded data to the total amount of the recorded amounts of the recorded data for the entire cluster, is below a cost determination ratio threshold; an abnormality determination step in which an abnormality determination unit determines, as an abnormality, an account amount that is outside a predetermined amount range for a rare cost item; and a display control step in which a display control unit controls the display of abnormality determination result data, including the recorded amount, task, and cluster to which the task belongs, determined to be abnormal.

[0009] In addition, in order to solve the above-mentioned problems and achieve the objective, the business support program of the present invention causes a computer to function as a cluster generation unit that assigns multiple recorded data, including at least account items and recorded amounts, recorded for each task, to multiple groups, or clusters, according to the similarity between the recorded data; a detection unit that detects, for each cluster, rare cost items, which are account items whose amount composition ratio, which is the ratio of the recorded amount of each recorded data to the total amount of the recorded amounts of the recorded data for the entire cluster, is below a cost determination ratio threshold; an abnormality determination unit that determines, as abnormal, an account amount for a rare cost item that is outside a specified amount range; and a display control unit that controls the display of abnormality determination result data, including the recorded amount, task, and cluster to which the task belongs, determined to be abnormal on a display unit. [Effects of the Invention]

[0010] The present invention can support the work of monitoring fraudulent accounting through the substitution of account items. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a block diagram illustrating a hardware configuration of a task assistance device according to an embodiment. [Figure 2]FIG. 2 is a diagram illustrating an example of an abnormality determination definition storage unit provided in the task support device according to the embodiment. [Figure 3] FIG. 3 is a diagram illustrating an example of a clustering result table provided in the task assistance device according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of an abnormality determination result table provided in the task assistance device according to the embodiment. [Figure 5] FIG. 5 is a diagram illustrating an example of an abnormality determination result message table provided in the task assistance device according to the embodiment. [Figure 6] FIG. 6 is a diagram illustrating an example of a construction data table provided in the task assistance device according to the embodiment. [Figure 7] FIG. 7 is a diagram showing the first half of a construction cost data table provided in the task support device according to the embodiment. [Figure 8] FIG. 8 is a diagram showing the second half of the construction cost data table provided in the task support device according to the embodiment. [Figure 9] FIG. 9 is a diagram illustrating an example of a construction budget data table provided in the business support device according to the embodiment. [Figure 10] FIG. 10 is a flowchart showing the flow of the monitoring operation for fraudulent accounting due to substitution of account items in the business support device according to the embodiment. [Figure 11] FIG. 11 is a diagram showing how the cluster generating unit acquires the abnormality determination definition data. [Figure 12] FIG. 12 is a diagram showing how the cluster generating unit acquires construction data of the construction progress status and the judgment range indicated by the abnormality judgment definition data. [Figure 13] FIG. 13 is a diagram showing how the cluster generation unit acquires construction cost data for each construction project whose range target is "construction progress status" and whose judgment range is "completed" from the construction cost data table 16. [Figure 14] FIG. 14 is a diagram for explaining the one-hot encoding process. [Figure 15]FIG. 15 is a diagram showing an example of data before the robust standardization process. [Figure 16] FIG. 16 is a diagram showing an example of data after the robust standardization process and the one-hot encoding process. [Figure 17] FIG. 17 is a diagram illustrating an example of a clustering result table provided in the task assistance device according to the embodiment. [Figure 18] FIG. 18 is a diagram showing a group of source data for clustering processing in which three centers of gravity are set randomly. [Figure 19] FIG. 19 is a diagram showing a state in which each piece of source data for clustering processing is sorted into three clusters based on three randomly set centers of gravity. [Figure 20] FIG. 20 is a diagram for explaining the operation of resetting the center of gravity data for each cluster. [Figure 21] FIG. 21 is a diagram for explaining the calculation of an index that expresses the quality of clustering. [Figure 22] FIG. 22 is a diagram for explaining the cluster reorganization process. [Figure 23] FIG. 23 is a diagram for explaining how rare cost items are recognized based on the cost amounts of each cluster. [Figure 24] FIG. 24 is a diagram for explaining how rare cost items are recognized based on the number of cost details recorded in each cluster. [Figure 25] FIG. 25 is a diagram showing how the cost amount of the rare cost item for each construction project in each cluster is detected from the construction project cost data table. [Figure 26] FIG. 26 is a diagram illustrating an example of a construction-specific rare cost item cost amount storage unit that stores the cost amount of the rare cost item for each construction project in each cluster. [Figure 27] FIG. 27 is a diagram for explaining an abnormality determination method using an interquartile range abnormality determination algorithm. [Figure 28]Figure 28 is a diagram showing an example of a rare cost item cost amount storage unit for each construction project that stores the upper threshold used in the interquartile range abnormality determination algorithm and the normal / abnormal classification that results in an abnormality determination of the cost amount of each construction project. [Figure 29] Figure 29 is a diagram showing an example of an abnormality determination result table in which abnormality determination result data for construction projects, clusters, cost account items, etc. corresponding to cost amounts determined to be abnormal are stored, and an abnormality determination result message table in which abnormality determination result message data is stored. [Figure 30] FIG. 30 is a diagram showing a state in which abnormality determination result data and abnormality determination result message data are displayed in a list. [Figure 31] FIG. 31 is a diagram showing how desired abnormality determination result data and abnormality determination result message data are designated from the list of abnormality determination result data and abnormality determination result message data. [Figure 32] FIG. 32 is a diagram showing an example of a display screen of a chart in a state where the desired abnormality determination result data and abnormality determination result message data that have been specified are displayed. [Figure 33] FIG. 33 is a diagram showing an extraction condition setting screen and a parameter storage section in which various data of the set extraction conditions is stored. [Figure 34] FIG. 34 is a diagram showing an example of various data that are the source of display of chart (1). [Figure 35] FIG. 35 is a diagram showing an example of various data that are the source of display of chart (2). [Figure 36] FIG. 36 is a diagram for explaining the process of aggregating the cost amounts of each cost account item in each cluster. [Figure 37] FIG. 37 shows data that aggregates the cost amounts of each cost account item for each cluster, and is the data that is the source of display for Chart (3). [Figure 38] FIG. 38 is another diagram showing data that is the aggregate of the cost amounts of each cost account item for each cluster, and is the source data for displaying Chart (3). [Figure 39] FIG. 39 is a diagram showing an example of various data that are the source of display of chart (4). [Figure 40] FIG. 40 is a diagram for explaining an example of displaying chart (4) by specifying another cluster and another cost account item. [Figure 41] FIG. 41 is a diagram showing the data that is the display source of the chart (4) generated based on other specified clusters and other cost account items. [Figure 42] FIG. 42 is a diagram showing an example of various data that are the source of display of chart (5). [Figure 43] FIG. 43 is a diagram showing an example of data that is the source of display of chart (5), including the cost amounts of the specified cluster and cost account items. [Figure 44] FIG. 44 is a diagram showing an example of a bar graph of Chart (1). [Figure 45] FIG. 45 is a diagram showing an example of the display of chart (2). [Figure 46] FIG. 46 is a diagram showing an example of a bar graph of Chart (3). [Figure 47] FIG. 47 is a diagram showing an example of the display of chart (4). [Figure 48] FIG. 48 is a diagram showing an example of the display of chart (5). [Figure 49] FIG. 49 shows examples of display of charts (1) to (5) displayed based on clusters, construction works, and rare cost items in which abnormalities have been detected. [Figure 50] FIG. 50 shows examples of display of charts (1) to (5) displayed based on the cost account title and cluster specified as extraction conditions. [Figure 51] FIG. 51 is a diagram for explaining the display operation of chart (2) corresponding to the graph operated on chart (1). [Figure 52] FIG. 52 is a diagram for explaining the display operation of charts (4) and (5) corresponding to the graph operated on chart (3). DETAILED DESCRIPTION OF THE INVENTION

[0012] A business support device according to an embodiment of the present invention will be described in detail below with reference to the drawings. This embodiment is an example in which the present invention is applied to a business support device that supports accounting work for construction work, which is an example of "work (project)." However, the present invention is not limited to the following embodiment.

[0013] (overview) In recent years, corporate internal control standards have been revised, resulting in increased demand for strengthened controls and monitoring. With fraud in corporate management still on the rise, and with the aging and turnover of personnel, there is a demand for an environment that allows even unskilled users to check data.

[0014] Meanwhile, many cases of accounting fraud have been reported in the construction industry. In particular, there are many cases of fraudulent manipulation and recording of construction costs with the aim of concealing loss-making construction projects or falsifying financial statements. For example, there is fraud in which incurred construction costs are recorded under an account other than the account they should be recorded under (account substitution). This account substitution is often carried out using an account that is rarely recorded for that construction project (account with few opportunities to record = rare cost account). This is because rare cost accounts are often not subject to verification for correct journal entry entries.

[0015] The cost structure (amount or ratio) of rare cost items tends to be similar between construction projects with similar characteristics (type, period, etc.). Therefore, by focusing on construction projects with large amounts of rare cost items when comparing similar construction projects, it is possible to detect fraudulent accounting where account items have been swapped.

[0016] Here, the judgment of similar construction projects is often highly dependent on the experience of accounting staff, etc. Furthermore, as the number of construction projects increases, the workload on accounting staff also increases.

[0017] The business support device of the embodiment automatically "determines whether construction projects are similar," "determines rare cost items," and "detects construction projects where account items may have been swapped." In other words, since "construction projects that tend to have large amounts in rare cost items may have had account items swapped," the business support device of the embodiment detects construction projects that have distinctive amounts in rare cost items compared to other construction projects and prompts the user to investigate, thereby supporting fraudulent accounting monitoring work.

[0018] Specifically, construction projects with similar characteristics (type, duration, etc.) tend to have similar cost structures (amounts or ratios). For this reason, the business support device of the embodiment groups similar construction projects, detects rare cost items in each group, and detects construction projects in which a large number of these rare cost items are recorded.

[0019] In other words, the business support device of the embodiment automatically performs the following steps: "grouping construction projects with similar characteristics" → "determining rare cost items in each group" → "detecting construction projects for which the amount of rare cost items recorded is abnormally high compared to other construction projects in the same group."

[0020] Furthermore, the business support device of the embodiment displays detailed data on the screen for construction work in which an abnormally large amount of the detected rare cost item is recorded, enabling the detected construction work to be analyzed. This makes it possible to detect construction work in which account items may have been swapped without relying on the experience of an accounting staff member, thereby speeding up the work and preventing it from becoming dependent on one person.

[0021] Furthermore, when detecting similar construction projects, the business support device of the embodiment automatically groups similar construction projects together by using a "clustering method," which is a machine learning algorithm that groups each construction project data based on the similarity between the construction project data. This reduces the burden on accounting staff of detecting similar construction projects and prevents the work from becoming dependent on one person.

[0022] Furthermore, the business support device of the embodiment uses a "clustering method" to group construction work, and then calculates the component ratio of the cost amount for each group. Then, it determines the rare cost items for each group (each cluster) based on the rule that "account items whose component ratio of the calculated cost amount is below a certain value are considered rare cost items." This automates the determination of account items that are rare cost items without relying on the experience of the accounting staff, preventing the inconvenience of this decision being made by a specific individual.

[0023] (Hardware configuration) The hardware configuration of the business support device 1 according to this embodiment is shown in FIG. 1. As shown in FIG. 1, the business support device 1 according to this embodiment includes a storage unit 2, a control unit 3, a communication interface unit 4, and an input / output interface unit 5. An input device 6 and an output device 7 are connected to the input / output interface unit 5. The output device 7 may be a display unit such as a monitor device (including a home television), a printer, or a speaker device. The input device 6 may be a keyboard device, a mouse device, a microphone device, or a monitor device that cooperates with a mouse device to provide a pointing device function. The communication interface unit 4 is connected to a network, such as a wide area network like the Internet or a private network like a local area network (LAN).

[0024] The storage unit 2 may be, for example, a read-only memory (ROM), a random access memory (RAM), a hard disk drive (HDD), or a solid state drive (SSD). The storage unit 2 stores a fraudulent accounting monitoring program, which is an example of a business support program that supports the monitoring of fraudulent accounting by switching account items. This fraudulent accounting monitoring program includes a "clustering method" algorithm that is used as an example of an algorithm for detecting the similarity of construction work (tasks) in the business support device 1 of the embodiment.

[0025] The storage unit 2 also has an abnormality determination definition storage unit 11, a clustering result table 12, an abnormality determination result table 13, an abnormality determination result message table 14, a construction data table 15, and a construction cost data table 16, which are each storage areas. The storage unit 2 also has a construction budget data table 17, a clustering processing original data storage unit 18, a clustering result table 19, a cluster-by-cluster cost composition ratio storage unit 20, a construction-by-construction rare cost item cost amount storage unit 21, and a parameter storage unit 22, which are each storage areas.

[0026] As shown in FIG. 2, the abnormality determination definition storage unit 11 stores definition identification information (definition ID), range target, determination range, clustering explanatory variables (quantitative variables), clustering explanatory variables (qualitative variables), number of clusters, proportion threshold for scarce cost determination, abnormality determination algorithm, quartile multiplier, determination item, group unit, etc.

[0027] The business support device 1 of the embodiment performs anomaly detection using three algorithms: a clustering method, a rule-based method, and an anomaly detection algorithm. The clustering method is a method for grouping similar construction projects. The rule-based method focuses on the cost composition ratio of each cost account item within each group and determines that cost account items below a certain composition ratio are rare cost items. The anomaly detection algorithm focuses on rare cost items in each group and detects construction projects with abnormally high cost amounts for rare cost items.

[0028] For this reason, the anomaly determination definition storage unit 11 is set with clustering explanatory variables (quantitative variables), clustering explanatory variables (qualitative variables), and the number of clusters used in the clustering method. The clustering explanatory variables (quantitative variables) are set with items that serve as the basis for clustering the construction work. In the example of FIG. 2, the contract amount and construction period are set. The clustering explanatory variables (qualitative variables) are set with items that serve as the basis for clustering the construction work. In the example of FIG. 2, the type of construction work is set. The number of clusters is the number of clusters to be generated, and in the example of FIG. 2, the number of clusters is set to "4 (generate four groups)".

[0029] Furthermore, the anomaly determination definition storage unit 11 is set with a ratio threshold for determining rare costs, which is used in the rule-based method. The ratio threshold for determining rare costs is a threshold for detecting cost accounts that are rare cost items when focusing on the cost composition ratio for each cost account within each group. In the example of Figure 2, "1%" is set as the ratio threshold for determining rare costs. In other words, cost accounts with a cost composition ratio of "1% or less" are determined to be rare cost items, as described below.

[0030] It is also possible to statistically determine rare cost items by using the "interquartile range" algorithm instead of the rare cost determination percentage threshold. In this case, items equivalent to the anomaly determination algorithm below are used.

[0031] Furthermore, the algorithm to be used as the abnormality determination algorithm, quartile magnification, determination item, and group unit are set in the abnormality determination definition storage unit 11. In the example of Fig. 2, the interquartile range algorithm is set as the algorithm to be used as the abnormality determination algorithm in order to detect construction work in which rare cost items are recorded at abnormally high levels.

[0032] The quartile multiplier is a parameter required when employing the interquartile range algorithm; the larger the value set, the wider the range of values ​​considered normal. In the example in Figure 2, it is set to "1.5." It is preferable to set a predetermined initial value for this quartile multiplier when the system is running, and then adjust (tune) the value according to the detection results obtained.

[0033] The judgment item is the item to which the interquartile range algorithm is applied. In the example of Figure 2, the "cost amount" is set as the judgment item. The group unit is the unit for making anomaly judgments. In the example of Figure 2, it is set that anomaly judgments are made based on cost accounts on a cluster basis. In other words, when "cluster and cost accounts" are set as the group unit, the values ​​of records with cost accounts within the same cluster are compared to make anomaly judgments.

[0034] Next, the clustering result table 12 is a table in which the grouped construction works are stored, and definition IDs, detection targets, and clusters are stored in association with each other as shown in Fig. 3. In the example of Fig. 3, construction works A to E are assigned to "cluster A," and construction works F to J are assigned to "cluster B."

[0035] Next, the abnormality determination result table 13 is a table in which the construction work and cost account item, cluster, cost amount, upper limit value, etc. that have been determined to be abnormal are stored together with the definition ID, as shown in Fig. 4. The example in Fig. 4 is an example in which the cost amount of the cost account item "miscellaneous expenses" for construction work A belonging to cluster A is 100,000 yen compared to the upper limit value of 12,000 yen, and therefore detected as "abnormal" and stored.

[0036] Next, as shown in Fig. 5, an abnormality determination result message corresponding to the registered contents of the abnormality determination result table 13 is automatically generated by the control unit 3 and stored in the abnormality determination result message table 14. The example in Fig. 5 is an example in which a determination result message such as "Construction: Construction A, cost account item: miscellaneous expenses detected" is automatically generated based on the registered contents of the abnormality determination result table 13 shown in Fig. 4, and a determination result detail message such as "The cost amount within cluster A is abnormal (upper limit value 12,000 yen)" is automatically generated, and each is stored in the abnormality determination result message table 14.

[0037] The construction data table 15 stores detailed information about each construction project, as shown in Fig. 6. In the example of Fig. 6, the construction project name, business establishment name, department name, person in charge name, payee name, supplier name, construction type, construction period (months), contract amount, and construction progress information are associated with each other and stored for each construction project.

[0038] As shown in Figures 7 and 8, the construction cost data table 16 stores the cost details, recording date, product name, detailed notes, transaction details, etc. recorded for each construction project. Figure 7 shows the first half of the construction cost data table 16, and Figure 8 shows the second half of the construction cost data table 16. The example shown in Figure 7 shows an example in the case of construction A, where material costs of 1 million yen, labor costs of 1.1 million yen, outsourcing costs of 1.2 million yen, and miscellaneous expenses of 90,000 yen and 10,000 yen, respectively, were recorded between March 1, 2023 and March 9, 2023.

[0039] The construction budget data table 17 is a table that stores the budget amount allocated to each construction project, as shown in Fig. 9. The example in Fig. 9 shows an example in which a budget amount of "10,000 yen" is allocated to miscellaneous expenses for construction A and construction C, a budget amount of "8,000 yen" is allocated to miscellaneous expenses for construction B, a budget amount of "6,900 yen" is allocated to miscellaneous expenses for construction D, and a budget amount of "6,200 yen" is allocated to miscellaneous expenses for construction E.

[0040] (Functional configuration of business support device) Next, the control unit 3 executes a fraudulent accounting monitoring program, which is an example of a business support program stored in the memory unit 2, and thereby functions as an abnormal value definition execution unit 30 and an alert list confirmation unit 31, as shown in Figure 1.

[0041] The abnormal value definition execution unit 30 includes a cluster generation unit 32, a detection unit 33, an abnormality determination unit 34, and a result data update unit 35. The alert list confirmation unit 31 includes a result data acquisition unit 36 ​​and a display control unit 37. The display control unit 37 includes a list screen display unit 38 and an alert details display unit 39.

[0042] The cluster generation unit 32 assigns multiple pieces of accounting data (construction data) including at least account items and accounting amounts recorded for each task (construction work) to clusters, which are multiple groups, according to the similarity between the accounting data (see Figures 18 to 22).

[0043] The detection unit 33 detects rare cost items for each cluster, which are account items whose amount composition ratio, which is the ratio of the recorded amount of each recorded data to the total amount of the recorded amount of the recorded data for the entire cluster, is below the cost determination ratio threshold (see Figures 23 and 24).

[0044] The abnormality determination unit 34 determines that the recorded amount of the scarce cost item that is outside a predetermined amount range is abnormal (see FIGS. 25 to 28).

[0045] The display control unit 37 controls the display of the abnormality determination result data including the total amount, work, and cluster to which the work belongs that are determined to be abnormal on the display unit (output device 7) (see FIGS. 30 and 32).

[0046] In addition, the cluster generation unit 32 selects, from among the multiple recorded data, recorded data that will serve as the basis for generating each cluster, the number of which is the same as the number of clusters to be generated, and sets them as selected data (data A1, B1, and C1 shown in Figure 19), and assigns each of the other recorded data to a cluster of selected data with a similar value, thereby assigning the recorded data to one of the clusters.

[0047] In addition, the display control unit 37 displays on the display unit, together with the abnormality determination result data, a predetermined number of corresponding work (construction) and recorded amount (cost amount) in descending order of the recorded amount of rare cost items within the cluster to which the recorded data corresponding to the recorded amount determined to be abnormal belongs (see Figure 34).

[0048] In addition, the display control unit 37 reads out detailed data (construction cost data shown in Figures 7 and 8) including the work date, work entity, account item, recorded amount, and use of the recorded data corresponding to the recorded amount determined to be abnormal from the memory unit (construction cost data table 16 in Figure 1) and displays it on the display unit.

[0049] Furthermore, the display control unit 37 displays the total amount of the amounts posted to each account item for each task in each cluster on the display unit (see FIG. 38).

[0050] In addition, the display control unit 37 controls the display unit to display the recorded amounts of rare cost items for each task in the cluster to which the task belongs, including the recorded amounts of rare cost items determined to be abnormal (see Figure 39).

[0051] Furthermore, the display control unit 37 displays at least one or more of the work, account item, or recorded amount corresponding to the abnormality determination result data in a display format different from the display format of other objects (see FIGS. 44 to 47).

[0052] In addition, when a desired account item and a desired cluster are specified, the display control unit 37 reads out from the memory unit the recorded amount of the specified account item for each task belonging to the specified cluster and displays it on the display unit (see Figures 40 and 52).

[0053] In addition, the display control unit 37 displays on the display unit the total amount of each recorded amount for each account item of each task in the cluster to which the task belongs, including the recorded amount of the rare cost item determined to be abnormal, and the amount composition ratio for each total amount, which is the ratio of each total amount to the added amount of each total amount (see Figure 42).

[0054] In addition, when a desired account item and a desired cluster are specified, the display control unit 37 displays on the display unit the total amount of the recorded amounts of the specified account items of each task belonging to the specified cluster, and the amount composition ratio, which is the ratio of the recorded amount of the specified account item of each task to the total amount (see Figure 43).

[0055] (Fraudulent accounting monitoring operation) Next, the monitoring operation for fraudulent accounting due to the substitution of account items in the business support device 1 of the embodiment will be described. Fig. 10 is a flowchart showing the flow of this monitoring operation. Based on the fraudulent accounting monitoring program stored in the storage unit 2, the control unit 3 functions as an abnormal value definition execution unit 30 to an alert list confirmation unit 31, thereby executing each process of steps S1 to S3 shown in Fig. 10.

[0056] In this example, the abnormal value definition execution unit 30 to the alert list confirmation unit 31 are described as being realized based on software, which is a fraudulent accounting monitoring program. However, some or all of the abnormal value definition execution unit 30 to the alert list confirmation unit 31 may be realized by hardware. In either case, the same effects as those described below can be obtained.

[0057] (Step S1: Abnormality detection execution process) First, in step S1, the cluster generation unit 32 refers to the abnormality determination definition storage unit 11 and acquires the abnormality determination definition data to be used this time, as shown in FIG. 11(b), for each definition ID to be executed when the definition is executed, as shown in FIG. 11(a).

[0058] Next, the cluster generation unit 32 refers to the abnormality determination definition data master table for each acquired definition ID, and recognizes the range target and determination range. In the example of FIG. 11(b), the range target is "construction progress status," and the determination range is "completion." Therefore, as shown in FIG. 12, the cluster generation unit 32 refers to the construction data table 15, and detects construction data for which the construction progress status is "completion."

[0059] Next, the cluster generation unit 32 refers to the construction cost data table 16 as shown in Figure 13 and obtains the construction cost data for each construction project (construction project detected from the construction data table 15) whose range target is "construction progress status" and whose judgment range is "completed."

[0060] Next, the cluster generation unit 32 refers to the anomaly determination definition storage unit 11 and recognizes the "clustering explanatory variables (quantitative variables)" and the "clustering explanatory variables (qualitative variables)." In the example of FIG. 11(b), the "clustering explanatory variables (quantitative variables)" are "contract amount, construction period," and the "clustering explanatory variables (qualitative variables)" are "type of construction."

[0061] The cluster generation unit 32 performs a robust standardization process as preprocessing on the "contract amount, construction period" recognized as the "clustering explanatory variables (qualitative variables)", and performs a one-hot encoding process using a dummy variable as preprocessing on the "construction type" recognized as the "clustering explanatory variables (quantitative variables)".

[0062] Robust standardization is used in machine learning to align the scales of multiple numeric items. When performing preprocessing for robust standardization, the cluster generation unit 32 calculates "Xmedian," which is the median of the entire data, and "IQR," which is the interquartile range, and converts each numeric value based on the formula "X1 = (X - Xmedian) / IQR." Note that "X" is the data before robust standardization, and "X1" is the data after robust standardization.

[0063] The "construction period (months)" and "contract amount" of each construction project shown in Figure 15 are the data before robust standardization processing. In contrast, the "construction period (standardized)" and "contract amount (standardized)" data of the original data for clustering processing of each construction project shown in Figure 16 are the data after robust standardization processing.

[0064] The one-hot encoding process is a process for converting the "clustering explanatory variables (qualitative variables)" into data that can be easily handled by a machine learning algorithm. When performing the one-hot encoding process, the cluster generation unit 32 converts the "clustering explanatory variables (qualitative variables)" into data expressed as "0" or "1", as shown in FIG. 14. The data for "whether it is type A or not" and "whether it is type B or not" in the original data for clustering process for each construction project shown in FIG. 16 is the data after the one-hot encoding process.

[0065] When referring to the construction data table 15 shown in FIG. 15, the "construction type" of construction A to construction E is "construction type A." Therefore, as shown in FIG. 16, the "construction type" field for construction A to construction E is "construction type A or not" is "1," and the "construction type" field for construction F to construction J is "construction type B." Similarly, when referring to the construction data table 15 shown in FIG. 15, the "construction type" of construction F to construction J is "construction type B or not" is "1," and the "construction type" field for construction F to construction J is "construction type B or not" is "0," as shown in FIG. 16. In this way, the cluster generation unit 32 converts the "clustering explanatory variables (qualitative variables)" into data expressed as "0" or "1."

[0066] Next, the cluster generation unit 32 refers to the anomaly determination definition storage unit 11 shown in FIG. 11(b) and recognizes the "number of clusters." In the example of FIG. 11(b), the "number of clusters" is "3." Therefore, the cluster generation unit 32 generates a total of three clusters (groups), "Cluster A," "Cluster B," and "Cluster C," as shown in FIG. 17, from the clustering process original data of each construction project that has been subjected to robust standardization processing or one-hot encoding processing, as shown in FIG. 16. Then, the cluster generation unit 32 stores clustering result data indicating the assigned cluster (Cluster A, Cluster B, etc.) for each construction project (detection target) with the definition ID shown in FIG. 11(b) in the clustering result table 19, as shown in FIG. 17.

[0067] In the business support device 1 according to the embodiment, clustering is performed using the "k-means method." Specifically, the "circles" shown in FIG. 18 represent the clustering process source data for each construction project. The cluster generation unit 32 randomly sets the number of centroids (three in this example) from the clustering process source data for each construction project, the number of which is equal to the "number of clusters" set in the anomaly determination definition storage unit 11. Each of the clustering process source data A1, B1, and C1 shown in FIG. 18 is the clustering process source data set as a centroid. Hereinafter, each of the clustering process source data A1, B1, and C1 will be referred to as centroid data A1, centroid data B1, and centroid data C1.

[0068] Next, the cluster generation unit 32 sorts the clustering processing source data that are close in distance into clusters of the centroid data A1, B1, and C1 that are close in distance from the centroid data A1, B1, or C1. Fig. 19 shows a total of three clusters (number of clusters = 3) generated by sorting the clustering processing source data in this way: cluster A of centroid data A1, cluster B of centroid data B1, and cluster C of centroid data C1. The curved lines in Fig. 19 are imaginary lines, and are illustrated so that the boundaries between each cluster are clear. These imaginary lines are not actually formed.

[0069] Next, the cluster generating unit 32 resets new centers of gravity for cluster A, cluster B, and cluster C based on the clustering processing source data assigned to each of them (resets the center of gravity for each cluster). Center of gravity data A2, B2, and C2 shown in Fig. 20 are the center of gravity data newly set based on the clustering processing source data constituting each cluster.

[0070] Next, as shown in FIG. 21, the cluster generation unit 32 calculates the sum of squares of the distances between each clustering processing source data and the centroid data of the cluster to which it belongs (WCSS: Within Cluster Sum Squared) as an index (defined in advance) expressing the quality of the clustering.

[0071] Next, as shown in Fig. 22, based on the newly set centroid data A2, B2, and C2, the cluster generating unit 32 reorganizes each piece of clustering process source data into the cluster of centroid data A2, the cluster of centroid data B2, and C2, as described with reference to Fig. 19. This reorganization process may result in some of the clustering process source data being moved and assigned to another cluster, as shown by the dotted circle in Fig. 22.

[0072] As a clustering method, for example, a hierarchical clustering method (a method for hierarchically grouping data) such as the "group averaging method" or "Ward's method" may be used. By using this hierarchical clustering method, it is possible to eliminate the need to set the number of clusters in advance.

[0073] Once the source data for the clustering process has been clustered in this manner, the detection unit 33 calculates the component ratio of the cost amount of the cost account item for each cluster. The detection unit 33 also refers to the abnormality determination definition storage unit 11 shown in FIG. 11(b) and recognizes the cost determination ratio threshold. In this example, the abnormality determination definition storage unit 11 is set to "1%." The detection unit 33 determines that a cost account item with a component ratio of "1% or less," which is the cost determination ratio threshold, is a rare cost item (rare cost item).

[0074] Specifically, as shown in Fig. 13, the construction cost data table 16 stores detailed cost information for construction A to construction E that belong to cluster A. The detection unit 33 calculates the total amount of the cost item for each account item for construction A to construction E. That is, the cost amounts for the material cost account items for construction A to construction E are "1 million yen," "1.3 million yen," "1.6 million yen," "1.7 million yen," "2.1 million yen," and "2.5 million yen," respectively.

[0075] Therefore, the detection unit 33 calculates the total cost amount of the material cost account items for constructions A to E as "1 million yen + 1.3 million yen + 1.6 million yen + 1.7 million yen + 2.1 million yen + 2.5 million yen = 10.2 million yen." The detection unit 33 then stores the calculated total material cost amount (cost amount) in the cluster-by-cluster cost composition ratio storage unit 20 shown in FIG.

[0076] Similarly, the cost amounts of the miscellaneous expense account items for Construction A to Construction E are "90,000 yen and 10,000 yen," "9,000 yen," "8,000 yen," "7,000 yen," and "4,000 yen and 2,000 yen," respectively. Therefore, the detection unit 33 calculates the total cost amounts of the miscellaneous expense account items for Construction A to Construction E as "90,000 yen + 10,000 yen + 9,000 yen + 8,000 yen + 7,000 yen + 4,000 yen + 2,000 yen = 130,000 yen." Then, the detection unit 33 stores the calculated total amount of miscellaneous expenses (cost amount) in the cluster-by-cluster cost composition ratio storage unit 20 shown in FIG. 23.

[0077] As a result, as shown in FIG. 23, the cluster-specific cost composition ratio storage unit 20 stores the total amount (cost amount) of each account item for each construction project in each cluster.

[0078] Next, the detection unit 33 calculates the ratio (cost composition ratio) of the total amount of each account item to the total amount obtained by adding up the total amounts of each account item. In the example of Figure 23, the total amount obtained by adding up the total amounts of each account item in cluster A is "10.2 million yen + 10.7 million yen + 11.6 million yen + 130,000 yen = 32.63 million yen." Therefore, the detection unit 33 calculates the cost composition ratio of material costs by calculating "(10.2 million yen ÷ 32.63 million yen) × 100 ≒ 31.26%," and stores this in the cluster-specific cost composition ratio storage unit 20, as shown in Figure 23.

[0079] Similarly, the detection unit 33 calculates the cost composition ratio of miscellaneous expenses by calculating "(130,000 yen ÷ 32,630,000 yen) × 100 ≒ 0.4%" and stores it in the cluster-specific cost composition ratio storage unit 20 as shown in Figure 23.

[0080] By calculating the cost composition ratio of each cost account item in this way, the detection unit 33 detects cost composition ratios that are equal to or less than the cost determination ratio threshold of "1%" specified in the abnormality determination definition storage unit 11. In the example of Figure 23, the cost composition ratio of "miscellaneous expenses" is "0.4%," which is equal to or less than the cost determination ratio threshold of "1%". Therefore, the detection unit 33 recognizes the account item "miscellaneous expenses" as a rare cost item.

[0081] Note that while this example is one in which rare cost items are determined by focusing on the "amount" of the recorded costs, rare cost items can also be determined based on the "number of recorded cost details (number of recorded cost items)" shown in Figure 24. In this case, the method of calculating the cost composition ratio is the same as above, where the total number of recorded cost items for each account item is calculated by adding up the number of recorded cost items for each account item, and then the cost composition ratio of the number of recorded cost items for each account item to this total number is calculated. Then, cost account items with a cost composition ratio that is "1% or less," which is the cost determination ratio threshold, are recognized as rare cost items.

[0082] Next, the detection unit 33 refers to the construction cost data table 16 as shown in Figure 25 and extracts the cost amount of the rare cost item for each construction project in each cluster. In this example, the rare cost item is the account item "miscellaneous expenses." Therefore, the detection unit 33 extracts the cost amount of "miscellaneous expenses" for each construction project in each cluster and stores it in the construction-specific rare cost item cost amount storage unit 21 as shown in Figure 26.

[0083] Next, the abnormality determination unit 34 uses the abnormality determination algorithm (interquartile range) and quartile multiplier (1.5) set in the abnormality determination definition memory unit 11 shown in Figure 11 (b) to determine abnormal cost amounts (abnormal values) among the cost amounts of rare cost items for each construction project in each cluster stored in the construction-specific rare cost item cost amount memory unit 21.

[0084] Specifically, the anomaly determination unit 34 detects an abnormal value using the interquartile range (IQR) of the data, as shown in FIG. 27. The anomaly determination unit 34 sets the threshold for detecting an abnormal value as follows, using the first quartile (Q1) and the third quartile (Q3). The quartiles refer to the division values ​​(25%, 50%, 75%) obtained by dividing the data into four equal parts when the data is sorted in ascending order, and these values ​​are the first quartile (Q1), median (Q2), and third quartile (Q3). The value "1.5" is a parameter value for defining an outlier, and is the value of the quartile magnification set in the anomaly determination definition storage unit 11. Increasing the value of the quartile magnification increases the normal value range shown in FIG. 27, and decreases the number of values ​​detected as abnormal.

[0085] IQR=Q3-Q1 Upper threshold = Q3 + (1.5 × IQR)

[0086] That is, in the example of Figure 26, the cost amounts for each construction project are "6,000 yen," "7,000 yen," "8,000 yen," "9,000 yen," and "100,000 yen," in ascending order. In this case, the division values ​​for "6,000 yen," "7,000 yen," "8,000 yen," "9,000 yen," and "100,000 yen" are 0%, 25%, 50%, 75%, and 100%, respectively. Therefore, the median (second quartile) is "8,000 yen," and the first quartile of "Q1" and the third quartile of "Q3" mentioned above are "7,000 yen" and "9,000 yen," respectively.

[0087] Therefore, the abnormality determination unit 34 calculates the interquartile range (IQR) by calculating "IQR = Q3 - Q1 = 9,000 yen - 7,000 yen = 2,000 yen." Furthermore, the abnormality determination unit 34 calculates the upper threshold of the normal range (12,000 yen) shown in Figure 27 by calculating "upper limit threshold = Q3 + (1.5 x IQR) = 9,000 yen + (1.5 x 2,000 yen) = 12,000 yen." Then, the abnormality determination unit 34 stores the calculated upper limit threshold of "12,000 yen" in the construction-specific rare cost item cost amount storage unit 21 as the upper limit threshold for the cost amount of miscellaneous expenses for each construction project.

[0088] Next, the abnormality determination unit 34 determines that, among the cost amounts of miscellaneous expenses for each construction project, cost amounts that are equal to or less than the upper threshold are normal cost amounts, and enters information of "normal" in the normal / abnormal category as shown in Figure 28. On the other hand, among the cost amounts of miscellaneous expenses for each construction project, cost amounts that exceed the upper threshold are abnormal cost amounts, and enters information of "abnormal" in the normal / abnormal category as shown in Figure 28. In the example of Figure 28, the abnormality determination unit 34 determines that the cost amount of "100,000 yen" for construction A, which exceeds the upper threshold of "12,000 yen," is an abnormal recorded amount.

[0089] Next, the result data update unit 35 stores the abnormality determination result data, including the definition ID, detection target, cost account item, cluster, cost amount, upper limit value, etc. corresponding to the cost amount determined to be "abnormal" by the abnormality determination unit 34, in the abnormality determination result table 13 as shown in Figure 29(a). The example in Figure 29(a) shows that the cost amount of "100,000 yen" recorded as miscellaneous expenses (rare cost item) for construction A to which cluster A belongs exceeded the upper limit value of "12,000 yen" and was therefore detected as an "abnormal" recorded amount.

[0090] Furthermore, the result data update unit 35 generates abnormality determination result message data including the construction work determined to be abnormal, the cost account item determined to be abnormal, and a detailed message (a message verbalized so that it is easy for people to understand) based on the abnormality determination result data, and stores this in the abnormality determination result message table 14, as shown in Figure 29(b). The example shown in Figure 29(b) shows an abnormality determination result message indicating that the cost amount recorded as miscellaneous expenses for construction work A in cluster A has exceeded the upper limit of 12,000 yen.

[0091] (Step S2: Analysis screen initial display process) Next, the result data acquiring unit 36 ​​acquires abnormality determination result data from the abnormality determination result table 13 shown in Fig. 29(a), and also acquires abnormality determination result message data from the abnormality determination result message table 14 shown in Fig. 29(b). The list screen display unit 38 of the display control unit 37 displays the acquired abnormality determination result data and abnormality determination result message data in a list via the output device 7, which is an example of a display unit, as shown in Fig. 30.

[0092] Next, the accounting staff designates desired abnormality determination result data and abnormality determination result message data from the displayed list of abnormality determination result data and abnormality determination result message data via the input device 6, as shown in FIG.

[0093] When this designation operation is performed, the list screen display unit 38 displays the acquired abnormality determination result data and abnormality determination result message data in, for example, the topmost display area, as shown in FIG.

[0094] Furthermore, the list screen display unit 38 reserves a display area for chart (1), a display area for chart (2), a display area for chart (3), a display area for chart (4), and a display area for chart (5) from the remaining display area excluding the display areas for the abnormality determination result data and the abnormality determination result message data, as shown in Fig. 32. Then, the list screen display unit 38 displays graphs or data, etc., as described below, in the display area for each chart, to facilitate the analysis of fraudulent accounting.

[0095] When displaying graphs or data in the display area for each chart, the list screen display unit 38 acquires data corresponding to the extraction conditions entered by the accounting staff via the extraction condition setting screen shown in Fig. 33(a) from the construction data table 15, construction cost data table 16, construction budget data table 17, etc. Then, the list screen display unit 38 stores each acquired data in the parameter storage unit 22 as shown in Fig. 33(b), and displays the graphs or data in the display area for each chart based on the various data stored in this parameter storage unit 22.

[0096] 33(a) and 33(b) show the state in which the data for "meeting expenses" set on the extraction condition setting screen is obtained from the construction cost data table 16 shown in FIG. 13 and the construction budget data table 17 shown in FIG. 9 and stored in the parameter memory unit 22.

[0097] Specifically, the list screen display unit 38 stores the business establishment, department, person in charge, cost account item, contract amount, construction period (months), and type of work input via the extraction condition screen shown in Figure 33(a) in the parameter storage unit 22. Then, the list screen display unit 38 references the construction data table 15 shown in Figure 15 based on the business establishment, department, and person in charge parameters stored in the parameter storage unit 22, and acquires data for each item of business establishment, department, and person in charge name. Furthermore, the list screen display unit 38 references the construction cost data table 16 shown in Figure 13 and the construction budget data table 17 shown in Figure 9 based on the cost account items stored in the parameter storage unit 22, and acquires data for each item of cost account items.

[0098] Furthermore, the list screen display unit 38 references the construction data table 15 shown in FIG. 15 and acquires data for each item of construction type, construction period (months), and contract amount based on the parameters of the contract amount, construction period, and construction type stored in the parameter storage unit 22. Then, the list screen display unit 38 generates graphs or data, etc., described below, based on the acquired various data, and displays them in the display areas for each chart. The accounting staff analyzes the fraudulent accounting based on the displayed charts.

[0099] (Display operation of Figure (1)) Next, how to display each chart will be described. First, when displaying chart (1), the list screen display unit 38 refers to the clustering result table shown in Fig. 17 based on "construction A," which is the "detection target" of the abnormality determination data shown in Fig. 29(a), and detects cluster A as the cluster to which construction A belongs, and also detects constructions A to E that belong to this cluster A, as shown in Fig. 34.

[0100] Furthermore, the list screen display unit 38 refers to the construction cost data table 16 shown in Figure 25 based on the "cost account item" of "miscellaneous expenses" in the abnormality determination data shown in Figure 29(a), and detects the cost amounts recorded as "miscellaneous expenses" for construction A to construction E, as shown in Figure 34. Furthermore, the list screen display unit 38 refers to the construction budget data table 17 shown in Figure 9 based on the "cost account item" of "miscellaneous expenses" in the abnormality determination data shown in Figure 29(a), and detects the budget amounts recorded as "miscellaneous expenses" for construction A to construction E, as shown in Figure 34.

[0101] Furthermore, the list screen display unit 38 detects the "upper limit value (see FIG. 28)" of "12,000 yen" in the abnormality determination data shown in FIG. 29(a), and detects the "upper limit value" of "12,000 yen" set for works A to E as shown in FIG. 34. Then, based on the various data shown in FIG. 34, the list screen display unit 38 generates chart (1) in which the cost amounts and budget amounts for works A to E are shown in a bar graph as shown in FIG. 44, and displays the upper limit values ​​on this bar graph in the display area of ​​chart (1) shown in FIG. 32.

[0102] When chart (1) is displayed in this way, the alert details display unit 39 displays the bar graph of the cost amount of construction A in which an abnormality was detected in a display format such as a color or pattern that is different from the bar graphs of other constructions, for example, red. This makes it possible to highlight the cost amount of construction A in which an abnormality was detected, making it easier to perform the analysis work described below.

[0103] In addition, instead of the bar graph shown in Figure 44, the list screen display unit 38 may display the cost amount, budget amount, and upper limit value of works A to E shown in Figure 34 in the display area of ​​chart (1) shown in Figure 32.

[0104] (Display operation of Figure (2)) Next, the list screen display unit 38 refers to the abnormality determination result table 13 shown in Figure 29(a) and obtains the business establishment, department name, person in charge name, payee name, and supplier name for "Construction A" in which the abnormality was detected from the construction data table 15 shown in Figure 15. The list screen display unit 38 also refers to the abnormality determination result table 13 and obtains the cost accounting voucher number, cost accounting detail number, accounting date, cost account item, cost amount, product name, detail notes, and transaction content corresponding to "miscellaneous expenses," which is the cost account item in which the abnormality was detected, from the construction cost data table 16 shown in Figure 7.

[0105] Next, the list screen display unit 38 generates data to be used to display Chart (2) based on the various acquired data, as shown in Fig. 35. In this data, the voucher number and row number correspond to the cost appropriation voucher number and cost appropriation detail number acquired from the construction cost data table 16, respectively.

[0106] Next, based on the generated data of Fig. 35, the list screen display unit 38 displays the "List of incurred cost details related to miscellaneous expenses for construction A" shown in Fig. 45 in the display area of ​​chart (2) shown in Fig. 32. This allows the accounting staff to recognize the cost account item, posting date, detailed notes, transaction content, etc. of the "miscellaneous expenses" for which an abnormality was detected.

[0107] When chart (2) is displayed in this way, the alert details display unit 39 displays the characters for the construction work A in which the abnormality was detected and the characters for "miscellaneous expenses," which is the rare cost item in which the abnormality was detected, in the title of chart (2), "List of detailed costs related to miscellaneous expenses for construction work A," in a display format that is different from the other characters, for example, in red, or displays each of the above characters in a frame of red or blue, etc. This makes it possible to prominently display that chart (2) is detailed data for construction work A in which an abnormality was detected, making it easier to perform the analysis work described below.

[0108] (Display operation of Figure (3)) Next, the list screen display unit 38 refers to the clustering result table 19 shown in Fig. 17 and detects each construction work belonging to cluster A to cluster D. In addition, the list screen display unit 38 refers to the construction work cost data table 16 shown in Fig. 7 and detects the cost account item and cost amount recorded for each construction work detected.

[0109] Next, the list screen display unit 38 aggregates the cost account items and cost amounts recorded for each construction project for each cluster based on the following rules.

[0110] (1) Rare cost items in which abnormalities were detected (2) Within the cluster to which the detected construction work belongs, the top three cost account items with the highest cost amounts (excluding rare cost items) (3) Cost accounts that do not fall under either rule (1) or rule (2) above shall be classified as "Other."

[0111] Specifically, the list screen display unit 38 first tally up the cost details of material costs, labor costs, outsourcing costs, meeting costs, communication costs, and miscellaneous costs recorded for each construction project in a specific cluster, as shown in Fig. 36(a). That is, in the case of cluster A, the material costs of construction projects A to E are tallied up, the labor costs are tallied up, the outsourcing costs are tallied up, the meeting costs are tallied up, the communication costs are tallied up, and the miscellaneous costs are tallied up.

[0112] Next, if the rare cost item in which an abnormality was detected is "miscellaneous expenses," the list screen display unit 38 will retain the rare cost item of "miscellaneous expenses" and the aggregated cost amount as shown in Figure 36(b) based on the rule (1) above.

[0113] Next, based on the rule (2) above, the list screen display unit 38 stores the top three cost amounts of the cost account items shown in Figure 36(a) as shown in Figure 36(b). In this example, the cost amounts of the cost account items of material cost, labor cost, and subcontract cost are stored.

[0114] Next, based on the rule (3) described above, the list screen display unit 38 further aggregates the aggregated cost amounts of the meeting expenses and communication expenses shown in Figure 36(a), which are cost account items that do not fall under either rule (1) or rule (2), and sets this as the cost amount of the "Other" cost account item, as shown in Figure 36(b).

[0115] This generates data that aggregates the cost amounts of each cost account item for each cluster, as shown in Figures 37 and 38. Based on this data, the list screen display unit 38 generates a bar graph of the aggregated amounts, which is an aggregation of the cost amounts of each cost account item for each cluster, as shown in Figure 46, and displays this in the display area of ​​chart (3) shown in Figure 32. This allows the accounting staff to recognize the cost amounts of each cost account item for each cluster.

[0116] When chart (3) is displayed in this way, the alert details display unit 39 displays the bar graph of "miscellaneous expenses," a rare cost item for which an abnormality was detected, in a display format such as a color or pattern, such as red, that differs from the bar graphs of other account items. This makes it possible to highlight the cost amount of miscellaneous expenses for construction A, for which an abnormality was detected, and makes it easier to perform the analysis work described below.

[0117] (Display operation of Figure (4)) Next, the list screen display unit 38 refers to the abnormality determination result table 13 shown in Fig. 4 and the clustering result table 19 shown in Fig. 17 to detect all of the works A to E that are assigned to cluster A to which the work A in which the abnormality was detected belongs. Furthermore, the list screen display unit 38 refers to the abnormality determination result table 13 shown in Fig. 4 to detect the cost account item of "miscellaneous expenses," which is the cost account item in which the abnormality was detected. Furthermore, based on the definition ID of "AB001" for works A to E, the list screen display unit 38 refers to the abnormality determination definition storage unit 11 shown in Fig. 11(b) to acquire the clustering explanatory variables (quantitative variables) in which "contract amount, construction period" is set and the clustering explanatory variables (qualitative variables) in which "type of work" is set.

[0118] Next, the list screen display unit 38 refers to the construction data table 15 shown in Fig. 12 based on the clustering explanatory variables (quantitative variables) and the clustering explanatory variables (qualitative variables), and acquires the construction type, construction period (months), and contract amount for each of constructions A to E. In addition, the list screen display unit 38 acquires, from the construction cost data table 16 shown in Fig. 13, the cost amount of the cost account item of "miscellaneous expenses" that is set in the abnormality determination result table 13 shown in Fig. 4 as the cost account item in which an abnormality has been detected.

[0119] Next, based on the acquired data, the list screen display unit 38 generates data to be used to display chart (4), including the contract amount, type of work, construction period (months), and miscellaneous cost amount for works A to E belonging to cluster A, as shown in Fig. 39. Furthermore, since "10 months" is the median of the construction periods of each of works A to E, which are "12 months, 11 months, 10 months, 9 months, and 8 months," the list screen display unit 38 adds this "10 months" to the data to be used to display chart (4), as the median of the construction periods (months) of each of works A to E. Furthermore, since "11,800,000 yen" is the median of the contract amounts for each of the works A to E, which are "11,800,000 yen, 11,818,000 yen, 11,600,000 yen, 9,214,000 yen, and 8,412,000 yen," the list screen display unit 38 adds this "11,160,000 yen" as the median of the contract amounts for each of the works A to E to the data that is the source of display in Figure (4).

[0120] Then, based on this data, the list screen display unit 38 displays the "List of classification indicators for construction projects classified into cluster A" shown in Fig. 47 in the display area for chart (4) shown in Fig. 32. This allows the accounting staff to recognize the cost amount of miscellaneous expenses that has been recorded against the contract amount for each construction project.

[0121] Furthermore, when chart (4) is displayed in this manner, the alert details display unit 39 displays the clustering explanatory variables (quantitative variables) and the clustering explanatory variables (qualitative variables), namely, the contract amount, type of work, and construction period (months), in a display format that is different from the other characters, for example, in red, or displays each of the above-mentioned characters in a frame that is a color such as red or blue, as shown in Fig. 47. This allows the clustering explanatory variables (quantitative variables) and clustering explanatory variables (qualitative variables) to be displayed prominently in chart (4), making it easier to perform the analysis work described below.

[0122] Here, the display example of this chart (4) is an example in which data related to construction A, which recorded the amount of "miscellaneous expenses (rare cost item)" for which an abnormality was detected, and cluster A to which construction A belongs, are displayed as chart (4). However, the business support device 1 of the embodiment can generate and display chart (4) using data related to the cost account item and cluster specified by the accounting staff.

[0123] In this case, the accounting staff specifies the cost account and cluster they wish to display as chart (4), as shown in Figure 40. In the example of Figure 40, "Miscellaneous Salaries" and "Cluster B" have been specified as the cost account and cluster they wish to display, respectively.

[0124] Based on the specified "cluster B," the list screen display unit 38 refers to the clustering result table 19 shown in FIG. 17 and recognizes work F to work J as work belonging to cluster B. Furthermore, the list screen display unit 38 refers to the work data table 15 shown in FIG. 12 based on the clustering explanatory variables (quantitative variables) and clustering explanatory variables (qualitative variables) set in the anomaly determination definition storage unit 11 shown in FIG. 11(b), and acquires the work type, construction period (months), and contract amount for each of work F to work J. Furthermore, the list screen display unit 38 acquires the cost amounts for work F to work J under the cost account item "miscellaneous salaries" specified by the accounting staff from the work cost data table 16 shown in FIG. 13.

[0125] Next, based on the acquired data, the list screen display unit 38 generates data to be used to display chart (4), including the contract amount, type of work, construction period (months), and cost amount of "various salaries" for construction works F to J belonging to cluster B, as shown in Fig. 41. Furthermore, since "17 months" is the median of the construction periods of each of the construction works F to J, which are "15 months, 17 months, 18 months, 21 months, and 15 months," the list screen display unit 38 adds this "17 months" to the data to be used to display chart (4), as the median of the construction periods (months) of each of the construction works F to J. Furthermore, since "30.5 million yen" is the median of the contract amounts for each of the works F to J, which are "30.1 million yen, 30.3 million yen, 30.5 million yen, 30.7 million yen, and 30.9 million yen," the list screen display unit 38 adds this "30.5 million yen" as the median of the contract amounts for each of the works F to J to the data that is the source of display in Figure (4).

[0126] This allows the chart (4) to be generated and displayed with data related to the cost accounts and clusters specified by the accountant.

[0127] (Display operation of Figure (5)) Next, the list screen display unit 38 generates a chart (5) that includes, for each cost detail recorded from works A to E of cluster A to which work A in which an abnormality was detected belongs, the added amount obtained by adding up the cost amounts of each cost account item, and the cost composition ratio (%) of each cost account item that corresponds to the added amount of each cost account item.

[0128] Specifically, the list screen display unit 38 refers to the abnormality determination result table 13 shown in FIG. 29(a) to recognize the cluster A to which the work A, which has recorded the cost amount of the cost accounting item in which an abnormality was detected, belongs. The list screen display unit 38 also refers to the clustering result table 19 to recognize the works A to E to which the cluster A belongs. The list screen display unit 38 also refers to the work cost data table 16 shown in FIG. 13 to acquire each cost account item and cost amount recorded for the works A to E.

[0129] Next, the list screen display unit 38 performs a compilation process on the cost details recorded for the constructions A to E of the cluster A based on the following rules.

[0130] (1) Rare cost items in which abnormalities were detected (2) Within cluster A to which construction A, in which an abnormality was detected, belongs, the top three cost accounts with the highest cost amounts (excluding rare cost items) (3) Cost accounts that do not fall under either rule (1) or rule (2) above shall be classified as "Other."

[0131] Specifically, the list screen display unit 38 first aggregates the material costs for each of the construction projects A to E belonging to cluster A, then aggregates the labor costs, aggregates the outsourcing costs, aggregates the meeting costs, aggregates the communication costs, and aggregates the miscellaneous costs.

[0132] Next, if the rare cost item in which an abnormality was detected is "miscellaneous expenses," the list screen display unit 38 will retain the rare cost item of "miscellaneous expenses" and the aggregated cost amount as shown in Figure 42, based on the rule (1) above.

[0133] Next, based on the rule (2) above, the list screen display unit 38 stores the cost amounts of the cost account items of material costs, labor costs, and subcontract costs, which are the top three cost amounts among the cost amounts of each cost account item, as shown in Figure 42.

[0134] Next, based on the above-mentioned rule (3), the list screen display unit 38 aggregates the cost amounts of cost account items that do not fall under either rule (1) or rule (2), as shown in Figure 42, and sets this as the cost amount of the "Other" cost account item.

[0135] Then, the list screen display unit 38 generates data that will be used to display chart (5), as shown in Figure 42, which includes the cost amounts of the rare cost items of "miscellaneous expenses" recorded for projects A to E of cluster A, the cost amounts of the cost account items of material costs, labor costs, and subcontracting costs, and the cost amounts of the cost account item of "other."

[0136] The list screen display unit 38 also calculates, for each cost detail, a cost composition ratio, which is the ratio of the cost amount of each cost account item corresponding to the added amount obtained by adding up the material costs, labor costs, subcontract costs, miscellaneous costs, and other cost amounts shown in Fig. 42, and adds this to the data that is the source of display in chart (5). Then, based on this data that is the source of display, the list screen display unit 38 displays the data of "amount ratio of each cost account item in cluster A" shown in Fig. 48 in the display area of ​​chart (5) shown in Fig. 32. This allows the accounting staff to recognize the cost amount of each cost account item recorded from works A to E of cluster A to which work A, in which an abnormality was detected, belongs, and the cost composition ratio (%) of each cost account item.

[0137] Next, the display example of this chart (5) is an example in which the cost amounts and cost composition ratios of works A to E of cluster A to which work A belongs, which includes the amount of "miscellaneous expenses (rare cost item)" for which an abnormality was detected, are displayed as chart (5). However, the business support device 1 of the embodiment can generate and display a selective chart (5) corresponding to the cost account item and cluster specified by the accounting staff.

[0138] In this case, the accounting staff specifies the cost account and cluster they wish to display as chart (5), as shown in Figure 40. In the example of Figure 40, "Miscellaneous Salaries" and "Cluster B" are specified as the cost account and cluster they wish to display, respectively.

[0139] Based on the specified "cluster B," the list screen display unit 38 refers to the clustering result table 19 shown in Fig. 17 and recognizes works F to J as works that belong to cluster B. In addition, the list screen display unit 38 acquires the cost amounts of works F to J in the cost account item of "various salaries" specified by the accounting staff from the work cost data table 16 shown in Fig. 13.

[0140] Next, the list screen display unit 38 calculates the added cost amount by adding up the cost amounts of "Miscellaneous Salaries" for the acquired works F to J. Then, as shown in Fig. 43, the list screen display unit 38 generates data that will be the display source for chart (5), including this added cost amount, and displays chart (5) in the display area for chart (5) shown in Fig. 32. This allows the accounting staff to specify the cluster and cost account item and recognize the cost amount.

[0141] (Step S3: Switching analysis, re-aggregation and calculation, chart redrawing process) Next, Fig. 49 shows an example of displaying charts (1) to (5) based on clusters, construction projects, and scarce cost items in which abnormalities have been detected. The display of charts based on clusters in which abnormalities have been detected is, so to speak, the initial display state of the charts. In contrast, Fig. 50 shows an example of displaying charts (1) to (5) based on cost account items and clusters specified as extraction conditions by the accounting staff, as explained using Fig. 40 and other figures. The business support device 1 of the embodiment can perform the analysis exemplified below by switching between the charts displayed in this way.

[0142] (Analysis based on Figure (1)) Specifically, in Figure 44, in chart (1), the accounting staff compares the cost amount of the rare cost item (the cost amount of miscellaneous expenses for construction A) with the other constructions B to E in the same cluster and their medians. Furthermore, the accounting staff checks whether the recorded cost amount exceeds the upper limit. This allows the accounting staff to check the degree of anomaly (how abnormal it is compared to the normal value) of the detected construction cost amount relative to the normal value.

[0143] In addition, in Figure (1), by comparing the budgeted amounts and cost amounts for Works A to E, it is possible to check the degree to which the recorded amounts deviate from the original budget. Even if an abnormal value is detected, if there is no significant deviation from the budgeted amount, it can be determined that "the cost amounts have been recorded as originally planned, and there is no problem."

[0144] In the case of Figure (1) shown in Figure 44, the cost amount recorded for Project A is significantly higher than the upper limit and is significantly different from the budgeted amount, so it can be determined that it is necessary to check the details of the miscellaneous expenses recorded for Project A.

[0145] (Analysis based on Chart (2)) Next, the accounting staff checks the details of the rare cost items recorded for the construction work in which the abnormality was detected based on chart (2) shown in Figure 45. The accounting staff also checks the details of the recorded data (remarks on the details or transaction details) using chart (2).

[0146] In addition, in the case of Chart (2) shown in Figure 45, the cost amount is large in the first line of the statement, and no remarks or transaction details are entered. In this case, the accounting staff will recognize that the first line of the statement may be a statement that has been fraudulently recorded.

[0147] (Analysis based on Chart (3)) Next, the accounting staff checks the extent to which the rare cost item in which the abnormality was detected is recorded in each cluster based on chart (3) shown in Figure 46. It can be confirmed whether the rare cost item in which the abnormality was detected is rare in a specific cluster. In the case of chart (3) shown in Figure 46, it can be confirmed that the miscellaneous expense account item in which the abnormality was detected is a rare cost item in clusters A, B, and C, but is not a rare cost item in cluster D.

[0148] (Analysis based on Chart (4)) Next, the accounting staff checks the list of construction projects that belong to one cluster based on Chart (4) shown in Figure 47. They also check the explanatory variables for the clustering of each construction project. By checking the construction projects that belong or the basis for the clustering, they can grasp the characteristics of the cluster.

[0149] In the case of chart (4) in Figure 47, it can be seen that cluster A contains construction projects with a contract amount of around 10 million yen, a construction period of around 10 months, and construction type A. By comparing these characteristics with other clusters and gaining insight, a person can assign meaning to the cluster, such as "Cluster A: small-scale, short-term construction projects."

[0150] (Analysis based on Chart (5)) Next, the accounting staff checks the composition ratio of costs in one cluster based on Chart (5) shown in Figure 48. They can check how rare the rare cost item in which the abnormality was detected is. Furthermore, they can check the trend in the recording of construction costs within the cluster.

[0151] (Display behavior of chart (2) corresponding to the graph operated in chart (1)) Next, when the bar graph of the desired construction work is operated on the chart (1) shown in Figure 51(a), the list screen display unit 38 displays a list of incurred cost details of rare cost items corresponding to the construction work of the bar graph operated on the chart (1) on the chart (2) shown in Figure 51(b).

[0152] Specifically, when the bar graph for Work B is selected by operating it on Chart (1), which displays the cost amounts of miscellaneous expenses, as shown in Figure 51(a), the list screen display unit 38 retrieves the business name, department name, person in charge name, payee name, and supplier name for Work B from the work data table 15 shown in Figure 15. The list screen display unit 38 also references the work cost data table shown in Figure 13 and retrieves the cost accounting slip number, cost accounting detail number, accounting date, cost account item, cost amount, product name, detail notes, and transaction details for Work B. Then, as shown in Figure 35, it generates data for displaying Chart (2) for Work B, and displays it in the display area for Chart (2), as shown in Figure 51(b).

[0153] This allows accounting personnel to easily check the cost details of rare cost items recorded in other construction projects (such as Construction B).

[0154] (Display operation of charts (4) and (5) corresponding to the graph operated in chart (3)) Next, when a bar graph of a desired cluster is operated on chart (3) shown in Fig. 52(a), the list screen display unit 38 displays a list of classification indicators of each construction project belonging to the cluster of the bar graph operated on chart (3) on chart (4) shown in Fig. 52(b). The list screen display unit 38 also displays the cost amount ratio of each cost account item of each construction project belonging to the cluster of the bar graph operated on chart (3) on chart (5) shown in Fig. 52(c).

[0155] Specifically, when the bar graph of cluster B is operated and designated on chart (3) displaying a bar graph of the cost amount of each account item for each cluster as shown in Figure 52(a), the list screen display unit 38 acquires the construction type, construction period (months), and contract amount of construction F to construction J belonging to cluster B from the construction data table 15 shown in Figure 12. In addition, the list screen display unit 38 acquires the cost amount of "miscellaneous expenses" of construction F to construction J belonging to cluster B from the construction cost data table 16 shown in Figure 13.

[0156] Then, the list screen display unit 38 displays the cost amount, contract amount, type of work, and construction period (months) of miscellaneous expenses for works F to J belonging to the acquired cluster B, along with the median contract amount and median construction period (months), in chart (4) as shown in Figure 52(b).

[0157] Furthermore, when the bar graph of cluster B is operated and designated on chart (3), the list screen display unit 38 refers to the construction cost data table 16 shown in Fig. 13 and acquires the cost amount of each account item of construction F to construction J belonging to cluster B. Then, as shown in Fig. 52(c), the list screen display unit 38 displays the cost amount of each account item of construction F to construction J belonging to cluster B together with the cost composition ratio on chart (5).

[0158] Furthermore, when a cost account item in the bar graph of Chart (3) is operated, the list screen display unit 38 displays Chart (3), Chart (4), and Chart (5) corresponding to the operated cost account item. For example, when the personnel cost portion of the bar graph of Cluster B shown in Figure 52(a) is operated, the list screen display unit 38 displays Chart (3), Chart (4), and Chart (5), which correspond to personnel costs. This allows the display of a chart of a desired cost account item to be used for analysis, etc.

[0159] (Effects of the embodiment) As is clear from the above explanation, the business support device 1 of the embodiment automatically performs the following steps: "grouping of constructions with similar properties" → "determining rare cost items in each group" → "detecting constructions for which the amount of rare cost items is recorded as abnormally high compared to other constructions in the same group." In addition, detailed data of the constructions for which the amount of rare cost items recorded as abnormally high is displayed on the screen, enabling analysis of the detected constructions.

[0160] This makes it possible to detect construction work for which account items may have been switched without relying on the experience of accounting personnel, and it is possible to speed up the work of monitoring fraudulent accounting by switching account items and prevent it from becoming dependent on one person.As a result, it is possible to support the work of monitoring fraudulent accounting by switching account items.

[0161] Furthermore, when detecting similar construction projects, the business support device of the embodiment automatically groups similar construction projects together using a "clustering method," which is a machine learning algorithm that groups each construction project data based on the similarity between the construction project data. This reduces the burden on accounting staff of detecting similar construction projects and prevents the work from becoming dependent on individual skills.

[0162] Furthermore, the business support device of the embodiment uses a "clustering method" to group construction work, and then calculates the component ratio of the cost amount for each group. Then, it determines the rare cost items in each group (each cluster) based on the rule that "account items whose component ratio of the calculated cost amount is below a certain value are considered rare cost items." This allows the determination of account items that are rare cost items to be automated without relying on the experience of accounting staff, preventing the inconvenience of this determination task becoming dependent on individual skills.

[0163] [Contribution to the United Nations-led Sustainable Development Goals (SDGs)] This invention can contribute to improving business efficiency and promoting appropriate management decisions by companies, thereby contributing to the achievement of goals "8" and "9" of the SDGs.

[0164] Furthermore, this invention can contribute to reducing waste and promoting paperless and electronic systems, thereby contributing to the achievement of SDGs goals 12, 13, and 15.

[0165] Furthermore, the present invention can contribute to strengthening control and governance, thereby contributing to the achievement of Goal 16 of the SDGs.

[0166] [Other embodiments] The present invention can be implemented in various different forms other than the above-described embodiments within the scope of the technical concept described in the claims.

[0167] For example, among the processes described in the embodiments, all or part of the processes described as being performed automatically may be performed manually, or all or part of the processes described as being performed manually may be performed automatically using a known method or the like.

[0168] Furthermore, the processing procedures, control procedures, specific names, registered data for each process, information including parameters such as search conditions, screen examples, and database configurations shown in the specification or drawings may be changed as desired unless otherwise specified.

[0169] Furthermore, the components of the business support device 1 shown in the figure are conceptual functional components and do not necessarily have to have the physical configuration shown in the figure. For example, all or any part of the processing functions of the business support device 1, particularly the processing functions performed by the control unit 3, may be realized by a program interpreted and executed by the control unit 3 (CPU: Central Processing Unit), or may be realized by hardware using wired logic.

[0170] The program is recorded on a non-transitory computer-readable recording medium containing programmed instructions for causing the information processing device to execute the processes described in the embodiments, and is mechanically read by the business support device 1 as needed. That is, a computer program is recorded in the storage unit 2, such as a ROM or HDD, for working with an OS (Operating System) to give instructions to a control unit 3 (CPU) and perform various processes. The computer program is loaded into RAM, expanded, and executed by the control unit 3 as appropriate.

[0171] In addition, the business support program (fraudulent accounting monitoring program) of this business support device 1 may be stored in another server device connected to the business support device 1 via any network, and all or part of it may be downloaded and executed as needed.

[0172] In addition, the business support program (fraudulent accounting monitoring program) for executing the processing described in the embodiment may be stored on a non-temporary computer-readable recording medium, or may be configured as a program product.

[0173] Here, the "recording medium" can be any "portable physical medium" such as a memory card, a USB (Universal Serial Bus) memory, an SD (Secure Digital) card, a flexible disk, a magneto-optical disk, a ROM, an EPROM (Erasable Programmable Read Only Memory), an EEPROM (registered trademark) (Electrically Erasable and Programmable Read Only Memory), a CD-ROM (Compact Disk Read Only Memory), an MO (Magneto-Optical Disk), a DVD (Digital Versatile Disk), and a Blu-ray (registered trademark) Disc.

[0174] Furthermore, a "program" is a data processing method written in any language or description method, regardless of the format, such as source code or binary code.

[0175] It should be noted that a "program" is not necessarily limited to a single structure, but includes a structure that is distributed as multiple modules or libraries, and a structure that achieves its function by working together with other programs, such as an OS.

[0176] Furthermore, the specific configuration for reading the recording medium in the task support device 1 of the embodiment, the reading procedure, and the installation procedure after reading can be any known configuration or procedure.

[0177] The memory unit 2 is a storage means such as a memory device such as RAM or ROM, a fixed disk device such as a hard disk, a flexible disk, or an optical disk, and stores various programs, tables, databases, web page files, etc. used for various processes or providing websites.

[0178] The business support device 1 may be configured as an information processing device such as a known personal computer or a workstation, or may be configured as an information processing device connected to any peripheral device. The information processing device may be implemented with software (including programs or data) that realizes the processes described in the embodiments.

[0179] Furthermore, the specific forms of distribution and integration of the devices are not limited to those shown in the drawings, and all or part of them can be functionally or physically distributed or integrated in any unit depending on various additions or functional loads. In other words, the above-mentioned embodiments can be selectively implemented by combining them in any way. [Industrial Applicability]

[0180] The present invention is suitable for application to accounting work in various industries, and is particularly suitable for application to monitoring fraudulent accounting through the substitution of account items.

[0181] Specifically, the present invention can be applied to other industries, such as the IT (Information Technology) media industry, where project or case management is required. Furthermore, the present invention groups projects with similar characteristics and detects projects that exhibit abnormal values ​​within each group. For example, in the IT media industry, projects with similar characteristics can be grouped, abnormal values ​​can be detected within each group, and cost trends within each group can be analyzed. [Explanation of symbols]

[0182] 1 Business support equipment 2 Storage section 3. Control Unit 4. Communication interface section 5 Input / Output Interface Section 6 Input Devices 7 Output Devices 11 Definition storage section for abnormality judgment 12 Clustering results table 13 Abnormality judgment result table 14 Abnormality judgment result message table 15 Construction Data Table 16 Construction cost data table 17 Construction Budget Data Table 18 Original data storage unit for clustering processing 19 Clustering Results Table 20 Cluster-specific cost composition ratio storage section 21 Construction-specific rare cost item cost amount memory section 22 Parameter storage section 30 Abnormal Value Definition Execution Unit 31 Alert list confirmation section 32 Cluster generation unit 33 Detection unit 34 Abnormality determination section 35 Result data update section 36 Result data acquisition section 37 Display control unit 38 List screen display section 39 Alert details display section

Claims

1. a cluster generation unit that assigns a plurality of pieces of recorded data, including at least account items and recorded amounts recorded for each task, to a plurality of clusters, which are groups, according to the similarity between the recorded data; A detection unit that detects, for each cluster, rare cost items that are account items whose amount composition ratio, which is the ratio of the accounted amount of each of the accounted data to the total amount of the accounted amounts of the accounted data of the entire cluster, is below a cost determination ratio threshold; an abnormality determination unit that determines, as an abnormality, when the recorded amount of the scarce cost item is outside a predetermined amount range; a display control unit that controls the display on a display unit of abnormality determination result data including the recorded amount determined to be abnormal, the work, and the cluster to which the work belongs; A business support device having the above.

2. the cluster generation unit selects, from the plurality of recorded data, recorded data that serve as a basis for generating each of the clusters, the number of which corresponds to the number of clusters to be generated, as selected data, and assigns each of the other recorded data to a cluster of the selected data that has a close value, thereby assigning the recorded data to any of the clusters; 2. The business support device according to claim 1, wherein:

3. The display control unit displays, on the display unit, a predetermined number of the corresponding operations and the recorded amounts in descending order of the recorded amounts of the rare cost items within the cluster to which the recorded data corresponding to the recorded amounts determined to be abnormal belongs, together with the abnormality determination result data; 3. The business support device according to claim 2, wherein:

4. the display control unit reads out from the storage unit detailed data including the work date, the work entity, the account item, the recorded amount, and the purpose of use of the recorded data corresponding to the recorded amount determined to be abnormal, and displays the data on the display unit; 4. The business support device according to claim 3, wherein:

5. the display control unit displays, on the display unit, the total amount of the recorded amounts for each of the account items for each of the tasks of each of the clusters; 5. The business support device according to claim 4, wherein:

6. The display control unit controls the display unit to display the recorded amount of the rare cost item of each of the operations in the cluster to which the operation including the recorded amount of the rare cost item determined to be abnormal belongs, 6. The business support device according to claim 5,

7. the display control unit displays at least one or more of the work, the account item, or the recorded amount corresponding to the abnormality determination result data in a display format different from a display format of other objects; 7. The business support device according to claim 3, wherein:

8. when a desired account item and a desired cluster are designated, the display control unit reads out from the storage unit the recorded amount of the designated account item for each of the operations belonging to the designated cluster and displays it on the display unit; 8. The business support device according to claim 7, wherein:

9. The display control unit displays on the display unit the total amount of each of the recorded amounts for each of the account items of each of the operations in the cluster to which the operation belongs, including the recorded amount of the rare cost item determined to be abnormal, and the amount composition ratio for each of the total amounts, which is the ratio of each of the total amounts to the added amount of each of the total amounts; 9. The business support device according to claim 8, wherein:

10. when a desired account item and a desired cluster are designated, the display control unit displays on the display unit the total amount of the recorded amounts of the designated account items of each of the operations belonging to the designated cluster, and an amount composition ratio which is the ratio of the recorded amount of the designated account item of each of the operations to the total amount; 10. The business support device according to claim 9,

11. a cluster generation step in which a cluster generation unit assigns a plurality of recorded data items, each including at least an account item and a recorded amount, recorded for each task to a plurality of clusters, which are groups, according to the similarity between the recorded data items; A detection step in which the detection unit detects, for each cluster, rare cost items that are account items whose amount composition ratio, which is the ratio of the accounted amount of each of the accounted data to the total amount of the accounted amounts of the accounted data of the entire cluster, is below a cost determination ratio threshold; an abnormality determination step in which an abnormality determination unit determines that the recorded amount of the scarce cost item is abnormal when the recorded amount is outside a predetermined amount range; a display control step in which a display control unit controls the display of abnormality determination result data including the recorded amount determined to be abnormal, the work, and the cluster to which the work belongs, on a display unit; A business support method having the above.

12. Computer, a cluster generation unit that assigns a plurality of pieces of recorded data, including at least account items and recorded amounts recorded for each task, to a plurality of clusters, which are groups, according to the similarity between the recorded data; A detection unit that detects, for each cluster, rare cost items that are account items whose amount composition ratio, which is the ratio of the accounted amount of each of the accounted data to the total amount of the accounted amounts of the accounted data of the entire cluster, is below a cost determination ratio threshold; an abnormality determination unit that determines, as an abnormality, when the recorded amount of the scarce cost item is outside a predetermined amount range; a display control unit that controls the display on a display unit of the abnormality determination result data including the recorded amount determined to be abnormal, the work, and the cluster to which the work belongs; A business support program that functions as a

Images