JP7147254B2 - Project Extraction Support Device, Project Extraction Support Method and Program - Google Patents

Description

The present invention relates to a case extraction support device, a case extraction support method, and a program.

In recent years, an electronic bidding system has been widely used, which is said to have effects such as improvement of competitiveness, reduction of cost, efficiency of office work, and the like.

US Pat. No. 6,200,000 discloses a bid monitoring system that can monitor bids and facilitate better analysis of bids. According to this document, in this bid monitoring system, an analysis unit 2 performs predetermined analysis based on various data relating to bids stored in a DB (database) 1 . Then, it is described that this bid monitoring system generates and displays a screen for displaying the analysis results so as to display the analysis results as appropriate visual information.

Patent Literature 2 discloses a data analysis system that enables users to efficiently find necessary information.

JP 2009-032103 A JP 2017-201543 A

The following analysis is given by the present invention. Although the above-mentioned electronic bidding system has achieved effects such as improvement of competitiveness, reduction of costs, and efficiency of clerical work, there is a problem that it is difficult to detect fraudulent bidding such as bid-rigging. The reasons for this are that bid rigging has become more sophisticated, that sufficient data has not been accumulated and shared between government agencies and local governments, and cross-sectional analysis has not been conducted over a certain period of time. is also conceivable.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a project extraction support device, a project extraction support method, and a program that can assist in the extraction of projects for inspecting whether bidding using the above-described electronic bidding system or the like has been properly conducted. do.

According to the first aspect, in a specific industry, a successful bid history analysis unit that calculates the number of successful bids or the total amount of successful bids for a predetermined period in the past of a plurality of traders who have successful bid results, and analysis results by the successful bid history analysis unit. and an output unit that outputs a case extraction support device.

According to a second aspect, in a specific industry, a step of calculating the number of successful bids or the total amount of successful bids for a predetermined period in the past of a plurality of traders who have a successful bid record, and outputting the analysis result obtained by the calculation. A case extraction support method including steps is provided.

According to a third aspect, a computer is provided with a process of calculating the number of successful bids or the total amount of successful bids for a predetermined period in the past of a plurality of traders who have a track record of successful bids in a specific industry, and the analysis result obtained by the calculation. A process for outputting and a program for executing are provided. This program can be recorded in a computer-readable (non-transitory) storage medium. That is, the present invention can also be embodied as a computer program product.

According to the present invention, there is provided means for confirming whether or not bidding by multiple vendors has been properly conducted.

1 is a diagram showing the configuration of an item extraction support device according to the first embodiment of the present invention; FIG. FIG. 4 is a diagram for explaining the operation of the item extraction support device according to the first embodiment of this invention; It is a figure which shows the structure of the matter extraction support apparatus of the 2nd Embodiment of this invention. It is a figure which shows the structure of the matter extraction assistance apparatus of the 3rd Embodiment of this invention. FIG. 12 is a diagram for explaining the operation of the successful bid record analysis unit of the item extraction support device according to the third embodiment of the present invention; FIG. 12 is a diagram for explaining the operation of the successful bid record analysis unit of the item extraction support device according to the third embodiment of the present invention; FIG. 13 is a flowchart for explaining the operation of the matter extraction support device according to the third exemplary embodiment of the present invention; FIG. FIG. 14 is a diagram for explaining the operation of a successful bid record analysis unit of the item extraction support device according to the fourth embodiment of the present invention; FIG. 14 is a diagram for explaining another operation of the successful bid record analysis unit of the item extraction support device according to the fourth embodiment of the present invention; It is a diagram showing the configuration of a computer that constitutes the case extraction support device of the present invention.

First, an outline of an embodiment of the present invention will be described with reference to the drawings. It should be noted that the drawing reference numerals added to this overview are added to each element for convenience as an example to aid understanding, and are not intended to limit the present invention to the illustrated embodiments. Also, connection lines between blocks in drawings and the like referred to in the following description include both bidirectional and unidirectional connections. The unidirectional arrows schematically show the flow of main signals (data) and do not exclude bidirectionality. Also, although there are ports or interfaces at input/output connection points of each block in the figure, they are omitted from the drawing.

[First embodiment]
In one embodiment of the present invention, as shown in FIG. 1, the item extraction support device 10 includes a successful bid record analysis unit 11 and an output unit 12 .

More specifically, the successful bid record analysis unit 11 calculates the total number of successful bids or the total amount of successful bids for a predetermined period in the past for a plurality of traders who have successful bid records in a specific industry. Then, the output unit 12 outputs the result of analysis by the successful bid record analysis unit.

For example, the item extraction support device 10 uses the bid data shown in FIG. 2 to calculate the sum of the successful bid prices for each successful bidder during a predetermined period in the past. In the example of FIG. 2, the total amount of winning bids of Company A, Company B, and Company C in the past two terms ("term" indicates an accounting term) is calculated. In the example of FIG. 2, these amounts are within the range of 100 million yen±5 million yen. Based on such data, the user can select the items for which Company A, Company B, and Company C won bids as items to be inspected (deliberated and examined) by a bid monitoring committee or the like.

In the actual bidding, in addition to cases where companies A, B, and C bid independently, cases where multiple companies form a JV (joint venture), and cases where affiliates of companies A, B, and C Bid cases are also envisioned. In the case of a JV (joint venture), it is preferable to perform processing such as apportioning the number of successful bids and the successful bid price. In addition, with respect to bidding for affiliated companies, it is sufficient to perform preliminary processing such as regarding them as the same company according to the strength of the capital relationship.

[Second embodiment]
Next, a detailed description will be given of a second embodiment in which a function of selecting a trader to be analyzed from among a plurality of traders who have successful bid records is added with reference to the drawings. FIG. 3 is a diagram showing the configuration of an item extraction support device 10a according to the second embodiment of the present invention. A difference from the case extraction support device 10 of the first embodiment shown in FIG. 1 is that a vendor designation unit 13 is added. Since other configurations are the same as those of the first embodiment, the differences will be mainly described below.

The trader designation unit 13 receives from the user a selection of traders to be tallied by the successful bid record analysis unit 11 .

The successful bid record analysis unit 11 calculates the total number of successful bids or the total amount of successful bids for a predetermined period in the past of the trader designated by the trader designating unit 13 . In the example of FIG. 3, company A and company D are selected from among a plurality of companies that have a track record of successful bids, and the sum of their successful bid prices is calculated. In the example of FIG. 3, the total amount of winning bids over the past 18 months is within the range of 100 million yen±3 million yen. The user can select the items for which Company A and Company D made successful bids as items to be inspected (deliberated and examined) by a bid monitoring committee or the like.

According to the present embodiment, it is possible to freely select vendors and compare the number of successful bids or the total amount of successful bids. Depending on the type of bid item and local government, there may be a wide variety of bidders who win bids, but in this embodiment, it is possible to select a specific bidder from among them and compare the number of successful bids and the amount of bids. For example, by selecting a vendor based on news reports or information from an information provider, it becomes possible to efficiently select items to be inspected (subjects to be deliberated and examined).

Although omitted in the above-described first and second embodiments, the user may be allowed to freely set the period (predetermined past period) during which the number of successful bids or the amount of successful bids is tallied. For example, in the comparison of the past two periods, it was not recognized that the number of successful bids or the successful bid price was within the predetermined range, but if the "past predetermined period" is changed, the number of successful bids or the successful bid price is within the predetermined range. It is said that there is By adding a function to change the "predetermined period in the past" to be aggregated to the case extraction support apparatuses 10 and 10a, it is possible to prevent cases from being overlooked.

Alternatively, the item extraction support apparatus 10, 10a automatically calculates the number of successful bids or the total amount of successful bids for a plurality of periods (for example, quarters to several years), and for each period thus calculated, each It is also possible to employ a configuration in which the number of successful bids and the successful bid price of each trader are presented to the user.

Further, a period having a certain width (range) may be set as the above-mentioned "predetermined period in the past". For example, if the user sets the predetermined period as 1 to 5 years, the total amount of successful bids is calculated for each year or every half year for 1 to 5 years, and there is a possibility that fraud such as collusion is being carried out. A period in which is high may be automatically extracted. By doing so, it is possible to more reliably prevent omissions in the detection of cases.

[Third embodiment]
Next, a detailed description will be given of a third embodiment in which a large amount of data obtained from an electronic bidding system can be analyzed using AI (Artificial Intelligence), with reference to the drawings. FIG. 4 is a diagram showing the configuration of an item extraction support device according to the third embodiment of the present invention.

Referring to FIG. 4, the item extraction support device 100 that can use the bid data stored in the item database 31 of the electronic bidding system 30 and the learning model 20 is shown. As the electronic bidding system 30, a package-type electronic bidding system introduced by public offices can be used.

The item extraction support device 100 also includes a successful bid record analysis unit 101 and an output unit 102 .

The successful bid record analysis unit 101 acquires and analyzes the bid data stored in the item database 31 of the electronic bidding system 30 . Also, the successful bid record analysis unit 11 analyzes the bid data using the learning model, and selects candidate deals to be inspected (discussed and examined) by the bid monitoring committee or the like.

The output unit 102 displays the analysis result of the bid data of the successful bid record analysis unit 101 and the information of the selected candidate item.

Here, a learning model used for selection of candidate items by the successful bid record analysis unit 101 will be described. In a preferred embodiment of the present invention, this learning model is created using cases (supervised learning data) of fraudulent bidding represented by "collusion" as described below.
・A case in which the total amount of successful bids (order amounts) for multiple vendors during a predetermined period in the past is within a certain range. Cases in which the total amount of successful bids (order amounts) by contract entity (government office, local government, etc.) of multiple contractors during a predetermined period in the past falls within a certain range. Cases in a predetermined period in the past. A case where the sum of at least one or more of the successful bid amounts (order amounts) for each business category, construction category, and type of work for multiple contractors is within a certain range.

The project categories, construction categories, and types of work can be organized as follows. For example, the business classification is a classification for business execution such as civil engineering work and landscaping work, the construction classification is a further subdivision of the business classification such as leveling, drainage, planting, etc. The type of work is ground leveling earthwork, retaining wall work, It can be positioned as a further subdivision of the construction categories such as planting foundation work and garden path plaza maintenance work.

A mechanism for selecting candidate items to be inspected (discussed and examined) by the successful bid record analysis unit 101 using the learning model as described above will be described. For example, as shown in FIG. 5, it is assumed that the amount of orders received (successful bid amount) of a trader for a fixed period in the past has been aggregated. Here, if the order amounts (successful bid amounts) of Company A, Company B, and Company C are within a certain range, the candidate items to be inspected (deliberated and examined) by the Bid Monitoring Committee, etc. are selected. can be done. Such cases can also be selected by the case extraction support devices 10 and 10a of the first and second embodiments.

On the other hand, if the order amounts (successful bid amounts) of Company A, Company B, and Company C do not fall within a certain range, there is a possibility that the projects in charge of Company A, Company B, and Company C will be excluded from the candidates. FIG. 6 is a diagram showing an example in which the order amounts (successful bid amounts) of Company A, Company B, and Company C are stratified by regional information (regions, prefectures, cities, towns and villages, etc.). In the example of FIG. 6, focusing on the order amount (successful bid amount) of Company A, Company B, and Company C in prefecture a, it can be said that the amounts are within a certain range. In such a case, the item extraction support device 100 of the present embodiment selects related items as candidate items to be inspected (examined and examined) by a bid monitoring committee or the like. Similarly, it is possible to extract candidate projects from multiple perspectives by comparing the project classification, construction classification, and construction type. Of course, it is also possible to compare the order amount (successful bid amount) by using the above-described area information and the business classification in combination. For example, the order amounts (successful bid amounts) of Company A, Company B, and Company C in prefecture a vary, but the order amounts (successful bid amounts) for specific construction categories in prefecture a are within a certain range. Sometimes. In this case, a series of related projects can be selected as candidate projects to be inspected (discussed and examined) by a bid monitoring committee or the like.

Since the determination is made using a learning model, the total amount of successful bids (accepted amount) in the specific region or contracting entity does not necessarily have to fall within a certain range. Specifically, AI is used to calculate the likelihood that indicates the possibility of fraudulent bidding and the score that indicates the suitability of a candidate project to be inspected (subject to deliberation and examination). If it exceeds the limit, it can be selected as an inspection object (deliberation examination object) by the bid monitoring committee, etc. Of course, it is also possible to adopt a configuration in which the case extraction support device 100 presents the likelihood and score to the user via the output unit 102 and allows the user to select the final candidate case.

Next, the operation of this embodiment will be described in detail with reference to the drawings. FIG. 7 is a flowchart for explaining the operation of the matter extraction support device according to the third embodiment of the present invention. Referring to FIG. 7, first, the item extraction support device 100 reads successful bid record data (acceptance record data) for the item extraction target period from the item database 31 of the electronic bidding system 30 (step S001, "Actual data input"). . If the item database 31 of the electronic bidding system 30 cannot be used, past successful bid record data (acceptance record data) of the government office or local government may be used.

Next, the item extraction support device 100 selects candidate items to be inspected (discussed and examined) by a bid monitoring committee or the like using the learning model described above (step S002). For example, when the analysis results shown in Figure 6 are obtained, the contracted projects of Company A, Company B, and Company C in prefecture a are selected as candidates for inspection (deliberation and examination) by the Bid Monitoring Committee, etc. be done. In addition, since a learning model is used, contracts with other prefectures that have a strong correlation with prefecture a and municipalities belonging to prefecture a are selected as candidates for inspection (subject to deliberation and examination) by the Bid Monitoring Committee, etc. Sometimes it is done.

Next, the item extraction support device 100 displays the information of the item selected as a candidate item to be inspected (discussed and examined) by the bid monitoring committee (step S003, "result output"). Here, the item extraction support device 100 may display, in addition to the information on the item selected as the candidate item, the above-described likelihood and score that are the reasons for the selection of the item. Further, the item extraction support device 100 may display the information of the trader who received the order for the selected item. This information may include the company profile of the trader, the punishment that the trader received in the past (in short, the content and cause of the bid suspension punishment), the past news about the trader, and the like.

As described above, according to the item extraction support device 100 of the present embodiment, AI is used to select candidate items to be inspected (deliberated and examined) by the bid monitoring committee or the like from various viewpoints. becomes possible.

Also, the learning model may be a single model, but it is also preferable to create it for each type of industry or size of local government. In this way, by creating a learning model for each type of industry and size of local government, it is possible to allow similar organizations that do not have sufficient bid data to extract deals using the learning model.

It is also preferable to use the results of whether or not the projects extracted using the project extraction support device of the present invention were actually fraudulent bids as new supervised learning data. By doing so, it is possible to create a learning model based on more cases and newer cases.

[Fourth embodiment]
Next, a fourth embodiment will be described in which a learning model is created using salesperson information included in successful bid record data (order record data). In the following, the sales person (bid person) will be taken as an example, but it is also possible to use the information of the person in charge of sales (for example, sales manager, business manager, etc.). Since the basic configuration of the fourth embodiment is similar to that of the third embodiment, the differences will be mainly described below.

FIG. 8 is a diagram showing an example in which the order amounts (successful bid amounts) of Company A, Company B, and Company C are stratified by sales persons (persons in charge of bidding). As in the third embodiment, if the order amounts (successful bid amounts) of Company A, Company B, and Company C do not fall within a certain range, the cases in charge of Company A, Company B, and Company C are excluded from the candidates. It may get lost. However, in the example of FIG. 8, focusing on the order amounts (successful bid amounts) of the persons in charge X, M, and V of Company A, Company B, and Company C, it can be said that the amounts are within a certain range. The item extraction support device 100 of the present embodiment selects such a case as a candidate item to be inspected (examined and examined) by a bid monitoring committee or the like.

Figure 9 shows the order amount (successful bid amount) of Company A, Company B, and Company C, stratified by regional information (region, prefecture, municipality, etc.), and then the amount of the sales person (bid person). It is a figure represented by color coding. In the example of FIG. 9, the order amounts (successful bid amounts) by region for Company A, Company B, and Company C do not fall within a certain range, so the projects handled by Company A, Company B, and Company C are excluded from the candidates. It may get lost. In addition, however, if we focus on sales representatives (bidding personnel) in each region, it can be said that the amounts are within a certain range. The item extraction support device 100 of the present embodiment selects such a case as a candidate item to be inspected (examined and examined) by a bid monitoring committee or the like.

In addition, since the judgment is based on a learning model, the total amount of successful bids (acceptance amounts) of the salesperson (bidder) or regional salesperson (bidder) does not necessarily fall within a certain range. you don't have to be Specifically, AI is used to calculate the likelihood that indicates the possibility of fraudulent bidding and the score that indicates the suitability of a candidate project to be inspected (subject to deliberation and examination). If it exceeds the limit, it can be selected as an inspection object (deliberation examination object) by the bid monitoring committee, etc. Of course, it is also possible to adopt a configuration in which the case extraction support device 100 presents the likelihood and score to the user via the output unit 102 and allows the user to select the final candidate case.

As described above, according to the item extraction support device 100 of the present embodiment, candidate items to be inspected (examined and examined) by the above-mentioned bid monitoring committee or the like are examined from a different point of view from the third embodiment. can be selected.

In addition, in the above-described fourth embodiment, the successful bid price (order price) is compared from the perspective of the sales person (bid person) or the regional sales person (bid person). You may stratify and compare the successful bid price (accepted price) from a viewpoint. For example, if information such as branches, sales offices, departments, etc. can be obtained for each project, a combination of these can be used instead of the salesperson information, and compared with the successful bid price (order price) stratified may be performed.

Although each embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiments, and further modifications, replacements, and adjustments can be made without departing from the basic technical idea of the present invention. can be added. For example, the network configuration, the configuration of each element, and the form of expression of messages shown in each drawing are examples for helping understanding of the present invention, and are not limited to the configurations shown in these drawings. Moreover, in the following description, "A and/or B" is used to mean at least one of A and B.

Further, in the above-described embodiment, it is assumed that candidates for inspection by a bid monitoring committee or the like are selected based on the total amount of successful bids for a predetermined period in the past from a plurality of companies that have a track record of successful bids. The number of cases may be used to select candidate cases. For example, if the number of successful bids by Company A, Company B, and Company C in the past 12 months is 10 each, these may be selected as candidates for inspection (deliberation and examination) by the Bid Monitoring Committee, etc. .

Further, in the above-described embodiment, it is explained that candidate items to be inspected by the Bid Monitoring Committee or the like are selected. For example, it is possible to use the information of the projects extracted for examination of the bidding method, examination criteria, and the like.

Further, in the above-described embodiment, the explanation was given on the assumption that the contract was decided by the contractor through competitive bidding. is to be confirmed. For this reason, the present invention can be suitably applied to cases in which a trader is determined by the comprehensive evaluation method.

Further, the procedures shown in the above-described first to fourth embodiments are executed in the computers (9000 in FIG. 10) functioning as the matter extraction support devices 10, 10a, and 100. It can be implemented by a program that implements the function. Such a computer is exemplified by a configuration including a CPU (Central Processing Unit) 9010, a communication interface 9020, a memory 9030, and an auxiliary storage device 9040 in FIG. That is, the CPU 9010 in FIG. 10 may execute a data analysis program or a determination program using a learning model to update each calculation parameter held in the auxiliary storage device 9040 or the like.

That is, each part (processing means, function) of the matter extraction support apparatus shown in the above-described first to fourth embodiments is implemented by a processor installed in the matter extraction support apparatus using the hardware. It can be realized by a computer program that executes the processing.

Finally, preferred forms of the invention are summarized.
[First form]
(Refer to the case extraction support device from the first viewpoint above)
[Second form]
The case extraction support device described above further
It is preferable to include a trader designation unit that receives designation of a trader for which the total is to be calculated from among the multiple traders who have a successful bid record.
[Third form]
The successful bid record analysis unit of the above-described item extraction support device,
A learning model using cases determined to have a high possibility of fraudulent bidding can be used to select the plurality of traders.
[Fourth mode]
The learning model used by the item extraction support device described above determines that a bid is fraudulent when the difference in the total number of successful bids or the total amount of successful bids for a predetermined item of a plurality of specific traders falls within a predetermined range. is preferably created using supervised learning data.
[Fifth form]
It is preferable that the predetermined project mentioned above is a project in a specific area.
[Sixth form]
It is preferable that the predetermined case described above is a combination of at least one or more of specific persons in charge, branches, sales offices, and departments of the plurality of traders.
[Seventh form]
It is preferable that at least one of the business category, construction category, and type of construction is the same in the above-described predetermined projects.
[Eighth mode]
In the case extraction support device described above,
A plurality of learning models are created for each type of industry and / and the scale of the local government,
It is preferable that the successful bid record analysis unit selects a learning model according to the type of industry or the size of the local government.
[Ninth form]
The case extraction support device described above is
From a packaged electronic bidding system introduced by at least one or more public offices, data on the number of successful bids or the amount of successful bids of a plurality of companies that have a successful bid record for a predetermined period in the past is acquired, and the extracted items are actually fraudulent bids. The result of whether or not it was can be used as new supervised learning data.
[Tenth mode]
The case extraction support device described above is
When a period having a predetermined width is set as the past predetermined period, each period belonging to the predetermined width can be analyzed.
[Eleventh form]
(Refer to the project extraction support method from the second viewpoint above)
[Twelfth form]
(Refer to the program from the third viewpoint above)
It should be noted that the eleventh to twelfth modes can be developed into the second to tenth modes in the same manner as the first mode.

The disclosures of the above patent documents and non-patent documents are incorporated herein by reference. Within the framework of the full disclosure of the present invention (including the scope of claims), modifications and adjustments of the embodiments and examples are possible based on the basic technical concept thereof. Also, within the framework of the disclosure of the present invention, various combinations or selections (partial (including targeted deletion) is possible. That is, the present invention naturally includes various variations and modifications that can be made by those skilled in the art according to the entire disclosure including claims and technical ideas. In particular, any numerical range recited herein should be construed as specifically recited for any numerical value or subrange within that range, even if not otherwise stated.

Reference Signs List 10, 10a, 100 item extraction support device 11, 101 successful bid record analysis unit 12, 102 output unit 13 vendor designation unit 20 learning model 30 electronic bidding system 31 item database 9000 computer 9010 CPU
9020 Communication interface 9030 Memory 9040 Auxiliary storage device

Claims (9)

Using the successful bid record data of a plurality of traders who have a track record of successful bids and a learning model generated using cases determined to have a high possibility of fraudulent bids as learning data, a successful bid performance analysis department that selects a deal ;
an output unit that outputs analysis results by the successful bid record analysis unit;
Item extraction support device. The learning model uses supervised learning data to indicate that a bid is fraudulent when the difference in the total number of successful bids or the total amount of successful bids for a given project from a plurality of specific traders falls within a predetermined range. The matter extraction support device according to claim 1 , which has been created. 3. A case extraction support device according to claim 2 , wherein said predetermined case is a case in a specific area. 3. The case extraction support device according to claim 2 , wherein said predetermined case is a combination of at least one of a specific person in charge, a branch, a sales office, and a department of said plurality of traders. 5. The project extraction support device according to any one of claims 2 to 4 , wherein the predetermined projects are projects in which at least one of the business division, construction division, and work type is the same. A plurality of learning models are created for each type of industry and / and the scale of the local government,
6. The item extraction support device according to any one of claims 1 to 5 , wherein said successful bid record analysis unit selects a learning model corresponding to a business type or a scale of a local government. Acquiring data on the number of successful bids or the amount of successful bids for a predetermined period in the past of a plurality of vendors who have a successful bid record from an electronic bidding system introduced by at least one or more public offices;
7. The item extraction support device according to any one of claims 1 to 6 , wherein a result as to whether or not the extracted item is actually a fraudulent bid is used as new supervised learning data. A computer uses bidding record data of a plurality of vendors who have a track record of winning bids and a learning model generated using cases determined to have a high possibility of fraudulent bidding as learning data, selecting a case to be inspected ;
a computer outputting the analysis result obtained by the selection ;
Project extraction support method including. to the computer,
Using the successful bid record data of a plurality of traders who have a track record of successful bids and a learning model generated using cases determined to have a high possibility of fraudulent bids as learning data, a process of selecting an item ;
a process of outputting the analysis result obtained by the selection ;
program to run.

Images