Project team establishment method and system based on knowledge graph

Abstract

The invention provides a project team recommendation building method and system based on a knowledge graph, and the method comprises the steps: collecting scientific researcher information, constructing a scientific researcher information database, creating a scientific researcher label library, constructing a scientific researcher knowledge graph, carrying out the vectorization of scientific researcher labels, and forming a scientific researcher exclusive label vectorization result; determining scientific research project demands, generating project demand tags, and establishing a project demand tag library; vectorizing and distributing weights for the project demand labels to form project demand label vectorization results; and according to the vectorization result, calculating the similarity between the scientific research personnel and the project demand, and carrying out personnel matching or project recommendation. According to the method, the scientific researchers and the project demands are subjected to tagging processing and then are processed through the knowledge graph, so that the scientific researchers and the project demands are converted into structured data which is easy to process and analyze, and weight selection is added, so that a project builder can realize personnel personalized recommendation according to own specific demands.

Description

Technical Field

[0001] The present invention belongs to the field of information matching technology, and in particular, relates to a project team building method and system based on knowledge graph. Background Art

[0002] At present, the process of forming a scientific research team mainly relies on manually screening information such as academic qualifications, professional experience and research achievements of researchers, and matching this information with specific project requirements. In addition, some scientific research management systems can also automatically recommend suitable candidates based on specific rules and algorithms by analyzing the personal information, research fields and achievements of researchers.

[0003] Although existing technologies have improved the efficiency of recommendations to a certain extent, they still have limitations in practical applications. First, traditional manual matching methods are not only inefficient, but also easily affected by personal bias, which affects the objectivity and fairness of matching. Second, although expert systems can provide assistance, the accuracy and credibility of their recommendations still need to be further improved, and personalized recommendations cannot be made. Finally, existing technologies usually only focus on the academic achievements and technical expertise of researchers, while ignoring their personalized characteristics, such as leadership, communication skills, etc., which are crucial to the overall performance of the team. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a project team building method based on knowledge graph, the method comprising:

[0005] Collect information on scientific researchers and build a scientific researcher information database;

[0006] Based on the scientific researcher information database, give scientific researchers exclusive labels, create a scientific researcher label library and build a scientific researcher knowledge map;

[0007] Vectorize the exclusive labels of researchers to form the exclusive label vectorization results of researchers;

[0008] Determine the needs of scientific research projects, generate project requirement tags, and establish a project requirement tag library;

[0009] Vectorize the project requirement labels and assign weights to form the project requirement label vectorization results;

[0010] Based on the scientific researcher knowledge graph, the vectorization results of the scientific researcher's exclusive labels and the vectorization results of the project requirement labels, the similarity between the scientific researcher and the project requirement is calculated to perform personnel matching or project recommendation.

[0011] Furthermore, the scientific researcher information includes: basic information, personality characteristics information, educational background information, professional experience information, scientific research information and scientific research results.

[0012] Furthermore, the creation of a researcher tag library and the construction of a researcher knowledge graph include:

[0013] The created researcher label library is divided into personnel label categories, and several personnel sub-label categories are set up to give researchers exclusive labels;

[0014] The construction of the knowledge graph of researchers includes entity extraction and relationship extraction;

[0015] The entity extraction is to extract the basic information of researchers as knowledge graph nodes, and store the exclusive labels of researchers as node attributes;

[0016] The relationship extraction is to extract the basic information of scientific researchers and the relationship between them as the edges of the knowledge graph.

[0017] Furthermore, the categories of personnel labels include: scientific research management, scientific research field, social attributes, and personal characteristics;

[0018] The exclusive labels for scientific researchers are identified and assigned by using data synchronization technology between the scientific researcher label library and the scientific researcher information database.

[0019] Furthermore, the researcher-specific labels are vectorized, including:

[0020] Define corresponding vector indexes for each person label category and the person sub-label dimension under each person label category;

[0021] The vector index of the personnel sub-label is concatenated with the vector index of the corresponding personnel label category dimension to form the vectorization result of the exclusive label for scientific researchers.

[0022] Furthermore, the weight allocation is to define the weight according to the usage and dynamically allocate it to the project requirement tag;

[0023] The project requirement tag library divides the requirement tag categories into several requirement sub-tag categories;

[0024] The demand label categories include scientific research management, scientific research fields, social attributes, and personal characteristics.

[0025] Furthermore, the project requirement tag vectorization includes: defining corresponding vector indexes for each requirement tag category and each requirement sub-tag under each requirement tag category dimension;

[0026] The vector index of the requirement sub-label is concatenated with the vector index of the corresponding requirement label category to form the project requirement label vectorization result.

[0027] Furthermore, the similarity between the scientific research project and the project requirement is calculated by combining the structural similarity algorithm and the weighted cosine similarity algorithm, including:

[0028] The structural similarity algorithm is based on the constructed scientific researcher knowledge graph and calculates the structural similarity of the scientific researcher knowledge graph. The formula is:

[0029]

[0030] Among them, A1 and A2 are the constructed knowledge graphs of researchers, N1 and N2 are the neighbor node sets of the ontologies A1 and A2, N1∩N2 represents the number of neighbor nodes shared by the two entities, and N1∪N2 represents their total number of neighbor nodes;

[0031] Introducing weights into the formula of weighted cosine similarity, the adjusted formula is expressed as,

[0032]

[0033] Among them: A i and B i are the vector representations of researchers and project requirements respectively; W i is the weight of the i-th dimension, reflecting the importance of this dimension in the matching process; n is the number of dimensions of the vector, that is, the number of attributes or characteristics of scientific research capabilities and project requirements that are subdivided;

[0034] Combining the structural similarity algorithm and the weighted cosine similarity algorithm, the proposed similarity calculation method is expressed as follows:

[0035] Sim=k1×Enty+k2×Strct

[0036] Among them, k1 and k2 are two adjustable weight coefficients, satisfying k1+k2=1; Sim is the calculated similarity value.

[0037] Furthermore, when the similarity value is larger, the similarity between the scientific researcher and the scientific research project is higher, and the matching degree is higher;

[0038] The smaller the similarity value is, the lower the similarity between the scientific researcher and the scientific research project is, and the lower the matching degree is.

[0039] A project team building system based on knowledge graph, which includes: data layer, business layer and application layer;

[0040] The data layer includes a database, a cache, a file server and an operating system; the data layer is used for data storage, data caching, data indexing and system operation;

[0041] The business layer includes a scientific researcher information collection module, a scientific researcher label library generation module, a scientific researcher knowledge graph generation module and a scientific research project demand label library generation module; the business layer is used to collect basic information of scientific researchers, and generate a scientific researcher label library and a scientific researcher knowledge graph, and at the same time, generate a scientific research project demand label library according to the needs of scientific research projects;

[0042] The application layer includes a scientific research project requirement label selection module, a scientific research project requirement weight selection module and a scientific researcher recommendation module; the application layer is used to select scientific research projects for which teams need to be formed, and to match and recommend scientific researchers based on the assigned scientific research project requirement weights.

[0043] Compared with the prior art, the present invention has the following advantages:

[0044] 1. The present invention labels the basic information of scientific researchers and scientific research data and constructs a knowledge graph. It can promote the integration and sharing of knowledge, accurate personnel matching and collaboration, accelerate the scientific research process and improve the quality of results, facilitate talent training and team building, and assist scientific research decision-making and strategic planning.

[0045] 2. By labeling and vectorizing scientific researchers and project requirements, scientific researchers and project requirements are transformed into structured data that is easy to process and analyze.

[0046] 3. The function of weighted selection has been added, so that project organizers can select the required scientific researchers according to their own specific needs, achieving highly personalized recommendations.

[0047] 4. The present invention provides a comprehensive, flexible and efficient matching and recommendation system and method, which is crucial for accurately matching scientific researchers, helping these organizations to better utilize talent resources and improve scientific research efficiency.

[0048] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 A flow chart of a project team building method based on knowledge graph of the present invention is shown.

[0051] Figure 2 A functional architecture diagram of a project team building system based on a knowledge graph of the present invention is shown. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] like Figure 1 The method for forming a scientific research project team based on a knowledge graph comprises:

[0054] S1. Collect personal information of scientific researchers, build a scientific researcher information database, fully understand the background, ability and personality characteristics of scientific researchers, and provide data support for subsequent labeling and recommendation.

[0055] Optionally, the scientific researcher information database ensures that the scientific researcher's personal information covers their latest research trends and achievements by connecting with data from other scientific and technological platforms, thereby ensuring that the scientific researcher information database can be updated in a timely manner.

[0056] The personal information of the scientific researchers includes basic information, personality characteristics information, educational background information, professional experience information, scientific research information and scientific research results.

[0057] Optionally, the basic information includes name, age, gender, etc.;

[0058] The personality characteristics information includes introversion / extroversion, carefulness / carelessness, etc.;

[0059] The educational background information includes degree, major, graduate school, etc.;

[0060] The professional experience information includes work unit, position, years of work, etc.;

[0061] The scientific research information includes academic research direction, research methods, awards received, teamwork ability, etc.;

[0062] The scientific research results include published papers, monographs, applied patents, software copyrights, participated scientific research projects, etc.

[0063] S2. Create a researcher tag library and build a researcher knowledge graph based on the researcher information database

[0064] S2.1. Create a detailed label library for researchers and divide the label categories into several sub-label categories.

[0065] Optionally, the divided label categories include scientific research management, scientific research fields, social attributes, personal characteristics, etc.

[0066] S2.2. According to the set sub-label categories, the various types of scientific researcher information collected in the scientific researcher information database in step S1 are analyzed and classified, and exclusive labels are assigned to the scientific researchers.

[0067] Optionally, data synchronization technology is used to synchronize information between the researcher tag library and the researcher information database, and to identify and assign the latest exclusive tags to researchers.

[0068] S2.3. Perform entity extraction and relationship extraction to construct knowledge graph nodes and edges.

[0069] Optionally, the entity extraction is to extract basic information of researchers as knowledge graph nodes.

[0070] The relationship extraction is to extract the basic information of scientific researchers and the relationship between them as the edges of the knowledge graph.

[0071] Optionally, the researcher-specific label is stored as a node attribute.

[0072] S2.4. Use the graph database to store the extracted entities and relationships in the form of a graph structure to form a knowledge graph for researchers.

[0073] S3. Vectorize the labels of scientific researchers to form vectorized results of scientific researcher labels, accurately describing the professional field, research expertise and unique energy of each scientific researcher.

[0074] S3.1. Define the dimension of the label category and define a unique basis vector index for each label category dimension.

[0075] S3.2. For each sub-label dimension under each label category dimension, define a unique vector index.

[0076] S3.3. Concatenate the vector index of the sub-label with the vector index of the corresponding label category dimension to form a vectorization result of the researcher label.

[0077] S4. Determine the scientific research project requirements and generate a scientific research project requirement label library.

[0078] Optionally, the scientific research project requirement label library, like the scientific researcher label library, divides label categories into categories such as scientific research management, scientific research fields, social attributes, personal characteristics, etc., and divides sub-label categories under each label category.

[0079] S5. Vectorize the demand labels, assign a weight to each demand label, and apply the weight to the corresponding label vector to form a demand label vectorization result.

[0080] S5.1. Similar to step S3, construct a vector index for the scientific research project requirement label category dimension, and construct a corresponding vector index for each sub-label dimension, concatenate the label category dimension and the sub-label dimension to form the vectorization result of the scientific research project requirement label.

[0081] S6. Based on the scientific researcher knowledge graph, the scientific researcher label vectorization results and the scientific research project requirement label vectorization results, calculate the similarity between scientific research projects and project requirements for accurate matching and recommendation.

[0082] The similarity between the scientific research project and the project requirements is calculated by combining the structural similarity algorithm and the weighted cosine similarity algorithm.

[0083] The structural similarity algorithm is based on the constructed scientific researcher knowledge graph and calculates the structural similarity of the scientific researcher knowledge graph. The formula is:

[0084]

[0085] Among them, A1 and A2 are the constructed knowledge graphs of scientific researchers, N1 and N2 are the neighbor node sets of ontologies A1 and A2, N1∩N2 represents the number of neighbor nodes shared by the two entities, and N1∪N2 represents their total number of neighbor nodes.

[0086] Introducing weights into the formula of weighted cosine similarity, the adjusted formula is:

[0087]

[0088] Among them: A i and B i are the vector representations of researchers and project requirements respectively; W iis the weight of the i-th dimension, reflecting the importance of this dimension in the matching process; n is the number of dimensions of the vector, that is, the number of attributes or characteristics of scientific research capabilities and project requirements that are subdivided.

[0089] The similarity calculation method proposed in the present invention combines the structural similarity algorithm and the weighted cosine similarity algorithm, and its formula is:

[0090] Sim=k1×Enty+k2×Strct

[0091] Among them, k1 and k2 are two adjustable weight coefficients, satisfying k1+k2=1, and Sim is the similarity value obtained after calculation.

[0092] Optionally, after similarity calculation, a similarity value can be obtained; when the value is larger, it means that the similarity between the scientific researcher and the scientific research project is higher, and the matching degree is better; conversely, when the value is smaller, it means that the similarity between the scientific researcher and the scientific research project is lower, and the matching degree is worse; so that the project organizer can select the most suitable scientific researcher by judging the similarity value.

[0093] like Figure 2 As shown, a project team building system based on knowledge graph includes a data layer, a business layer and an application layer.

[0094] The data layer includes a database, a cache, a file server and an operating system, which are used for data storage, data caching, data indexing and system operations.

[0095] The business layer includes a scientific researcher information collection module, a scientific researcher knowledge graph generation module and a scientific research project demand label library generation module; the business layer is used to collect basic information of scientific researchers, and generate a scientific researcher label library and a scientific researcher knowledge graph, and at the same time generate a scientific research project demand label library according to the needs of scientific research projects.

[0096] The application layer includes a scientific research project demand label selection module, a scientific research project demand weight selection module and a scientific researcher recommendation module.

[0097] Based on the above disclosed content, the present invention also provides an electronic device accordingly. The electronic device of an embodiment of the present invention includes at least one electrically connected processor and at least one storage medium, the storage medium is electrically connected to the processor, wherein the storage medium stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method as described above.

[0098] Based on the same inventive concept, the present invention also provides a storage medium, which stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method as described above.

[0099] The above description and the accompanying drawings fully illustrate the embodiments of the present invention so that those skilled in the art can practice them. Other embodiments may include structural and other changes. The embodiments represent only possible variations. Unless explicitly required, individual components and functions are optional, and the order of operations may vary. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. The embodiments of the present invention are not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A project team building method based on knowledge graph, characterized in that: The method comprises, Collect information on scientific researchers and build a scientific researcher information database; Based on the scientific researcher information database, give scientific researchers exclusive labels, create a scientific researcher label library and build a scientific researcher knowledge map; Vectorize the exclusive labels of researchers to form the exclusive label vectorization results of researchers; Determine the needs of scientific research projects, generate project requirement tags, and establish a project requirement tag library; Vectorize the project requirement labels and assign weights to form the project requirement label vectorization results; Based on the scientific researcher knowledge graph, the vectorization results of the scientific researcher's exclusive labels and the vectorization results of the project requirement labels, the similarity between the scientific researcher and the project requirement is calculated to perform personnel matching or project recommendation.

2. The project team building method based on knowledge graph according to claim 1 is characterized in that: The scientific researcher information includes: basic information, personality characteristics information, educational background information, professional experience information, scientific research information and scientific research results.

3. The project team building method based on knowledge graph according to claim 1 is characterized in that: The creation of a researcher tag library and construction of a researcher knowledge graph includes: The created researcher label library is divided into personnel label categories, and several personnel sub-label categories are set up to give researchers exclusive labels; The construction of the knowledge graph of researchers includes entity extraction and relationship extraction; The entity extraction is to extract the basic information of researchers as knowledge graph nodes, and store the exclusive labels of researchers as node attributes; The relationship extraction is to extract the basic information of scientific researchers and the relationship between them as the edges of the knowledge graph.

4. The project team building method based on knowledge graph according to claim 3 is characterized in that: The personnel label categories include: scientific research management, scientific research field, social attributes, and personal characteristics; The exclusive labels for scientific researchers are identified and assigned by using data synchronization technology between the scientific researcher label library and the scientific researcher information database.

5. The project team building method based on knowledge graph according to claim 1 is characterized in that: The researcher-specific label vectorization includes: Define corresponding vector indexes for each person label category and the person sub-label dimension under each person label category; The vector index of the personnel sub-label is concatenated with the vector index of the corresponding personnel label category dimension to form the vectorization result of the exclusive label for scientific researchers.

6. The project team building method based on knowledge graph according to claim 1 is characterized in that: The allocation weight is defined according to the usage and dynamically allocated to the project requirement tag; The project requirement tag library divides the requirement tag categories into several requirement sub-tag categories; The demand label categories include scientific research management, scientific research fields, social attributes, and personal characteristics.

7. The project team building method based on knowledge graph according to claim 1 is characterized in that: The project requirement tag vectorization includes: defining corresponding vector indexes for each requirement tag category and each requirement sub-tag under each requirement tag category dimension; The vector index of the requirement sub-label is concatenated with the vector index of the corresponding requirement label category to form the project requirement label vectorization result.

8. The project team building method based on knowledge graph according to claim 1 is characterized in that: The similarity between the scientific research project and the project requirement is calculated by combining the structural similarity algorithm and the weighted cosine similarity algorithm, including: The structural similarity algorithm is based on the constructed scientific researcher knowledge graph and calculates the structural similarity of the scientific researcher knowledge graph. The formula is: Among them, A1 and A2 are the constructed knowledge graphs of researchers, N1 and N2 are the neighbor node sets of the ontologies A1 and A2, N1∩N2 represents the number of neighbor nodes shared by the two entities, and N1∪N2 represents their total number of neighbor nodes; Introducing weights into the formula of weighted cosine similarity, the adjusted formula is expressed as, Among them: A i and B i are the vector representations of researchers and project requirements respectively; W i is the weight of the i-th dimension, reflecting the importance of this dimension in the matching process; n is the number of dimensions of the vector, that is, the number of attributes or characteristics of scientific research capabilities and project requirements that are subdivided; Combining the structural similarity algorithm and the weighted cosine similarity algorithm, the proposed similarity calculation method is expressed as follows: Sim=k1×Enty+k2×Strct Among them, k1 and k2 are two adjustable weight coefficients, satisfying k1+k2=1; Sim is the calculated similarity value.

9. The project team building method based on knowledge graph according to claim 8 is characterized in that: When the similarity value is larger, the similarity between the scientific researcher and the scientific research project is higher, and the matching degree is higher; The smaller the similarity value is, the lower the similarity between the scientific researcher and the scientific research project is, and the lower the matching degree is.

10. A project team building system based on knowledge graph, comprising: Data layer, business layer and application layer; The data layer includes a database, a cache, a file server and an operating system; The data layer is used for data storage, data caching, data indexing and system operations; The business layer includes a scientific researcher information collection module, a scientific researcher knowledge graph generation module and a scientific research project demand label library generation module; the business layer is used to collect basic information of scientific researchers, and generate a scientific researcher label library and a scientific researcher knowledge graph, and at the same time, generate a scientific research project demand label library according to the needs of scientific research projects; The application layer includes a scientific research project requirement label selection module, a scientific research project requirement weight selection module and a scientific researcher recommendation module; the application layer is used to select scientific research projects for which teams need to be formed, and to match and recommend scientific researchers based on the assigned scientific research project requirement weights.

Images