A method and system based on business relationship graph data mining

Abstract

A method and system based on business relationship graph data mining includes the following steps: obtaining information on all projects and the executive personnel associated with the projects; decomposing the projects according to the schedule to obtain sub-projects, and associating the executive personnel with the sub-projects; arranging the sub-projects in similar projects in an overlapping manner to obtain a project string, wherein the similarity judgment is related to the information of the executive personnel and the projects; and calculating the missing degree of the sub-projects in the project to be mined relative to the project string. This application decomposes existing projects to obtain sub-projects, and then groups and decomposes and arranges the sub-projects according to the similarity between the projects to obtain a project string, thereby providing a relatively objective comparison basis for comparing different projects, thereby discovering the missing points between similar projects, and thus obtaining possible new needs through mining existing projects.

Description

Technical Field

[0001] The present application relates to a method and system based on business relationship graph data mining. Background Art

[0002] Relationship graphs represent data in a graphical form, making connections between entities more distinct and thus more interpretable. Therefore, when data volumes are large, relationship graphs can fully leverage the power of data mining to achieve better data applications. However, it should be noted that the graphical representation of relationship graphs actually conceals a significant amount of underlying information. Relationship graphs, when used solely for their graphical nature, are not truly effective. Therefore, leveraging relationship graphs while avoiding oversimplification is crucial in relationship graph-based mining. Summary of the Invention

[0003] In order to solve the above problems, the present application discloses a method based on business relationship graph data mining, including the following steps: obtaining information on all projects and the executive personnel associated with the projects; decomposing the projects according to the progress to obtain sub-projects, and associating the executive personnel with the sub-projects; arranging the sub-projects in the projects with similarity to obtain a project string, wherein the similarity judgment is related to the information of the executive personnel and the projects; and calculating the missing degree of the sub-projects in the project to be mined relative to the project string. The present application decomposes the existing projects to obtain sub-projects, and then groups and decomposes and arranges the sub-projects according to the similarity between the projects to obtain a project string, thereby providing a relatively objective comparison basis for the comparison between different projects, so that the missing points between similar projects can be found. The missing points here include the missing points in the middle of the project, and also include the missing points in the later stage, that is, the missing points in the later stage of the project, so as to obtain possible new needs through the existing project mining.

[0004] Preferably, the similarity is determined as follows:

[0005] Obtain a project description, including the application field, design tasks, and design goals, and perform language normalization processing on the descriptions that have undergone language normalization processing. Perform similarity comparison on the descriptions that have undergone language normalization processing to obtain a preliminary similarity r;

[0006] The number of executive personnel in the two contrasting sub-projects is L1 and L2 respectively, L1≥L2, and the number of identical executive personnel in the two contrasting sub-projects is X;

[0007] Calculate P = X / L2,

[0008] If P=1, then the corrected similarity R=1.5*r;

[0009] If 1>P≥0.8, then the corrected similarity R=1.2*r;

[0010] If 0.8>P≥0.5, then the corrected similarity R=r;

[0011] If P < 0.5, then the corrected similarity R = 0.8*r;

[0012] If R is not less than 100%, they are considered identical; if R is greater than 60% but less than 100%, they are considered related; but if R is less than 60%, they are all considered unrelated. This application obtains natural language that can be used for similarity comparison by normalizing the description. Here, it is carried out through the entire project, and then for specific sub-projects, it is revised by the participating personnel. This is mainly because the consideration of personnel will only be meaningful if there is a certain degree of similarity. In the specific project analysis, it can also be seen that due to the characteristics of team work, different projects may also have the same people. However, if the overall tasks are similar, the same executors can handle similar tasks more easily.

[0013] Preferably, the item strings are arranged as follows:

[0014] Arrange the positions of sub-projects according to the overall progress of the project. If the sub-projects between different projects are the same, they will be merged and set. If the sub-projects between different projects are related, they will be set in parallel. The reason why this application is divided into merging and paralleling when setting up project strings is mainly to consider the specific analysis of missing degrees between sub-projects with large differences. When the system finds a missing point, it can consider which similar sub-project idea to adopt, providing more reference basis for business personnel or decision makers.

[0015] Preferably, the missing degree is obtained as follows:

[0016] The target project is analyzed in the project string in which it is located to obtain the number of all sub-projects in the project string. The number is calculated by combining identical sub-projects and counting them as 1, and combining parallel sub-projects and counting them as 1. The total number after the combined counting of the project strings is N. The number of sub-projects of the target project is n, and the missing degree λ = 1-n / N.

[0017] Preferably, the items are sorted by the missing degree value, and a mining threshold is set. For items exceeding the mining threshold, a reverse search is performed on the missing sub-items in the target item based on the item string content. The setting of the mining threshold in this application is mainly based on the severity of the mining. When the mining threshold is large, fewer items are mined, but the targeting is stronger. If the mining threshold is small, more items are mined, but the invalid mining is also increased.

[0018] Preferably, the method further includes clearing irrelevant sub-items, wherein the irrelevant sub-items refer to sub-items that are irrelevant to the main line information of the project, and the sub-items are deleted manually.

[0019] Preferably, it also includes the merging processing of executive personnel, which includes the replacement processing of resigned or transferred personnel or the merging processing of multiple people into one executive team.

[0020] Preferably, the project is broken down into sub-projects as follows:

[0021] Obtain project information, obtain the project's nodal requirements and set the nodal requirements in a unified manner, use the unified nodal requirements as a basis, decompose all projects into sub-projects, and obtain a decomposition template for the sub-project decomposition.

[0022] Preferably, in the process of decomposing a project into sub-projects, if a situation occurs in which the decomposition cannot be performed according to the nodal requirements, the situation will be separated automatically, and a new sub-project will be obtained based on the original decomposition template.

[0023] On the other hand, the present application also discloses a system based on business relationship graph data mining, including the following modules:

[0024] Data acquisition module, used to obtain information about all projects and the executives associated with the projects;

[0025] Decomposition module, used to decompose the project into sub-projects according to the schedule and associate the executors with the sub-projects;

[0026] A project string generation module is used to overlap and arrange sub-projects in projects with similarity to obtain a project string, wherein the similarity judgment is related to the information of the executor and the project;

[0027] The missing degree module is used to calculate the missing degree of sub-projects in the project to be mined relative to the project string, and obtain the potential demand of projects whose missing degree exceeds the mining threshold.

[0028] This application can bring the following beneficial effects:

[0029] 1. This application decomposes existing projects into sub-projects, and then groups, decomposes and arranges the sub-projects according to the similarity between the projects to obtain project strings, thereby providing a relatively objective comparison basis for the comparison between different projects, so that missing points between similar projects can be discovered. The missing points here include missing points in the middle of the project, as well as missing points in the later stage, that is, missing points in the later stage of the project, so as to obtain possible new needs through mining existing projects.

[0030] 2. This application obtains natural language that can be used for similarity comparison through the standardization of descriptions. This is done through the project as a whole, and then the specific sub-projects are revised by the participating personnel. This is mainly because the consideration of personnel will only be meaningful if there is a certain degree of similarity. In the specific project analysis, it can also be seen that due to the characteristics of team work, the same people may be used in different projects. However, if the overall tasks are similar, the same executors can handle similar tasks better.

[0031] 3. The reason why this application divides the project string into merging and parallel methods when setting it up is mainly to provide reference value when conducting specific analysis of the missing degree between sub-projects with great differences. When the system finds that a project is missing at this location, it can consider which similar sub-project idea to adopt, providing more reference basis for business personnel or decision makers. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0033] Figure 1 is a schematic diagram of Example 1;

[0034] Figure 2 This is a schematic diagram of Example 2. DETAILED DESCRIPTION

[0035] In order to clearly illustrate the technical features of this solution, this application is described in detail below through specific implementation methods.

[0036] In the first embodiment, if Figure 1 As shown, a method based on business relationship graph data mining includes the following steps:

[0037] S101 obtains information about all projects and the executives associated with the projects;

[0038] There is also a merging process for executive personnel, which includes replacing personnel who have resigned or transferred, or merging multiple personnel into one executive team.

[0039] S102 decomposes the project into sub-projects according to the progress, and associates the executors with the sub-projects;

[0040] The project is broken down into sub-projects as follows:

[0041] Obtain project information, obtain the project's nodal requirements and set the nodal requirements in a unified manner, use the unified nodal requirements as a basis, decompose all projects into sub-projects, and obtain a decomposition template for the sub-project decomposition.

[0042] Irrelevant sub-items are also cleared. Irrelevant sub-items refer to sub-items that are irrelevant to the main line information of the project. These sub-items are deleted manually.

[0043] In the process of decomposing a project into sub-projects, if there is a situation where the decomposition cannot be carried out according to the node requirements, the situation will be separated automatically and a new sub-project will be obtained based on the original decomposition template.

[0044] S103: overlapping and arranging sub-projects in projects with similarity to obtain a project string, wherein the similarity determination is related to the information of the executor and the project;

[0045] The similarity is determined as follows:

[0046] Obtain a project description, including the application field, design tasks, and design goals, and perform language normalization processing on the descriptions that have undergone language normalization processing. Perform similarity comparison on the descriptions that have undergone language normalization processing to obtain a preliminary similarity r;

[0047] The number of executive personnel in the two contrasting sub-projects is L1 and L2 respectively, L1≥L2, and the number of identical executive personnel in the two contrasting sub-projects is X;

[0048] Calculate P = X / L2,

[0049] If P=1, then the corrected similarity R=1.5*r;

[0050] If 1>P≥0.8, then the corrected similarity R=1.2*r;

[0051] If 0.8>P≥0.5, then the corrected similarity R=r;

[0052] If P < 0.5, then the corrected similarity R = 0.8*r;

[0053] If R is not less than 100%, they are considered identical; if R is greater than 60% and less than 100%, they are considered related; but if R does not exceed 60%, they are all considered unrelated.

[0054] The item strings are arranged as follows:

[0055] Arrange the positions of sub-projects according to the overall progress of the project. If the sub-projects between different projects are the same, merge the sub-projects. If the sub-projects between different projects are related, set the sub-projects in parallel.

[0056] S104 calculates the missing degree of the sub-items in the project to be mined relative to the project string.

[0057] The missing degree is obtained as follows:

[0058] The target project is analyzed in the project string in which it is located to obtain the number of all sub-projects in the project string. The number is calculated by combining identical sub-projects and counting them as 1, and combining parallel sub-projects and counting them as 1. The total number after the combined counting of the project strings is N. The number of sub-projects of the target project is n, and the missing degree λ = 1-n / N.

[0059] Sort by missing degree value, set mining threshold, and for those exceeding the mining threshold, reverse search for missing sub-items in the target project based on the project string content.

[0060] This method can be used to obtain potential missing content in some completed projects and projects that have the potential to be continued in the future.

[0061] like Figure 2 As shown, a system based on business relationship graph data mining includes the following modules:

[0062] The data acquisition module 201 is used to obtain information about all projects and the executive personnel associated with the projects;

[0063] Decomposition module 202, used to decompose the project into sub-projects according to the progress, and associate the executors with the sub-projects;

[0064] A project string generation module 203 is used to overlap and arrange sub-projects in projects with similarity to obtain a project string, wherein the similarity determination is related to the information of the executor and the project;

[0065] The missing degree module 204 is used to calculate the missing degree of sub-projects in the project to be mined relative to the project string, and obtain the potential demand of projects whose missing degree exceeds the mining threshold.

[0066] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method based on business relationship graph data mining, characterized by: The steps include: Get information about all projects and the executives associated with them; Decompose the project into sub-projects according to the progress, and associate the executors with the sub-projects; The sub-projects in the projects with similarity are overlapped and arranged to obtain a project string, wherein the similarity judgment is related to the information of the executor and the project; Calculate the missing degree of sub-projects in the project to be mined relative to the project string; The similarity is determined as follows: Obtain a project description, including the application field, design tasks, and design goals, and perform language normalization processing on the descriptions that have undergone language normalization processing. Perform similarity comparison on the descriptions that have undergone language normalization processing to obtain a preliminary similarity r; The number of executive personnel in the two contrasting sub-projects is L1 and L2 respectively, L1≥L2, and the number of identical executive personnel in the two contrasting sub-projects is X; Calculate P=X / L2, If P=1, then the corrected similarity R=1.5*r; If 1>P≥0.8, then the corrected similarity R=1.2*r; If 0.8>P≥0.5, then the corrected similarity R=r; If P < 0.5, the corrected similarity R = 0.8*r; If R is not less than 100%, they are considered identical; if R is more than 60% but less than 100%, they are considered related; but if R is less than 60%, they are all considered unrelated; The item strings are arranged as follows: Arrange the positions of sub-projects according to the overall progress of the project. If the sub-projects of different projects are the same, merge them together. If the sub-projects of different projects are related, place them in parallel. The missing degree is obtained as follows: Analyze the target project in the project string in which it is located to obtain the number of all sub-projects in the project string. The number is calculated by combining identical sub-projects and counting them as 1, and combining parallel sub-projects and counting them as 1. The total number of project strings after merging and counting is N. The number of sub-projects of the target project is n, and the missing degree λ=1-n / N.

2. The method according to claim 1, characterized in that: Sort by missing degree value, set mining threshold, and for those exceeding the mining threshold, reverse search for missing sub-items in the target project based on the project string content.

3. The method according to claim 1, wherein: It also includes the removal of irrelevant sub-items, which are sub-items that are irrelevant to the main line information of the project. The sub-items are deleted manually.

4. The method according to claim 1, wherein: It also includes the merging of executive personnel, including the replacement of personnel who have resigned or transferred, or the merging of multiple personnel into one executive team.

5. The method according to claim 1, wherein: The project is broken down into sub-projects as follows: Obtain project information, obtain the project's nodal requirements and set the nodal requirements in a unified manner, use the unified nodal requirements as a basis, decompose all projects into sub-projects, and obtain a decomposition template for the sub-project decomposition.

6. The method based on business relationship graph data mining according to claim 5, characterized in that: In the process of decomposing a project into sub-projects, if there is a situation where the decomposition cannot be carried out according to the node requirements, the situation will be separated automatically and a new sub-project will be obtained based on the original decomposition template.

7. A system for implementing the method based on business relationship graph data mining according to any one of claims 1 to 6, characterized in that: Includes the following modules: Data acquisition module, used to obtain information about all projects and the executives associated with the projects; Decomposition module, used to decompose the project into sub-projects according to the schedule and associate the executors with the sub-projects; A project string generation module is used to overlap and arrange sub-projects in projects with similarity to obtain a project string, wherein the similarity judgment is related to the information of the executor and the project; The missing degree module is used to calculate the missing degree of sub-projects in the project to be mined relative to the project string, and obtain the potential demand of projects whose missing degree exceeds the mining threshold.

Images