CN117312676B

CN117312676B - Intelligent reading recommendation and cooperation analysis method

Info

Publication number: CN117312676B
Application number: CN202311339176.5A
Authority: CN
Inventors: 李志义; 郭雅怡; 胡轶男
Original assignee: South China Normal University
Current assignee: South China Normal University
Priority date: 2023-10-16
Filing date: 2023-10-16
Publication date: 2024-06-04
Anticipated expiration: 2043-10-16
Also published as: CN117312676A

Abstract

The application provides an intelligent reading recommendation and cooperation analysis method, which comprises the following steps: searching research materials related to the scholars according to the determined cross research field, and providing recommended reading for the scholars; collecting information and published records of a proposed student, and judging professional depth and historical research data of the proposed student in the cross research field; comparing the professional depths of the scholars and the personification scholars, judging the potential cooperation degree of the two parties, and generating a cooperation proposal for the two parties; the cooperation will and the potential research direction of the scholars and the personification scholars are further evaluated by combining the activity degree of the scholars in the recommended materials; if both the learner and the proposed learner show the intention of cooperation and the predicted research directions are matched, automatically generating an intention-of-cooperation notification for the learner and the proposed learner; through academic influence indexes, the relative status and influence of both the scholars and the personification scholars in the potential cooperation field are comprehensively evaluated.

Description

Intelligent reading recommendation and cooperation analysis method

Technical Field

The invention relates to the technical field of information, in particular to an intelligent reading recommendation and cooperation analysis method.

Background

With the continued development and cross-convergence of the scientific research field, collaboration between scholars is becoming increasingly important. However, existing methods of learner collaboration have some problems. First, students often can only determine the collaboration object through their own published records and specialized fields, which has subjectivity and limitations. Since the published records reflect only the personal study results, the learner may miss opportunities to interdisciplinary work with experts in other areas. Furthermore, relying solely on specialized fields to find collaborative objects may limit the contact and exploration of new fields and views by students. Second, a learner often spends a significant amount of time and effort searching for relevant research materials after determining the collaborative object, which is a challenge for a busy learner. Existing learner partners lack efficient methods to collect and sort the learner's research efforts and interests and match them with potential partners. This makes it time consuming for the learner to search and screen between the various databases, academic papers, and research projects, which search results tend to be incomplete or inaccurate. In addition, the degree of cooperation and the intention of cooperation between the scholars are difficult to evaluate accurately. Existing learner cooperation methods tend to match based on the professional field and research emphasis of individuals, and lack comprehensive consideration of factors such as cooperation style, communication ability and team cooperation experience among individuals. Such one-sided matching may lead to dissonance of the partnership, affecting progress and outcome of the partnership project. Finally, existing methods lack comprehensive assessment of the relative status and impact of a learner and a proposed learner in the potential collaboration area.

Disclosure of Invention

The invention provides an intelligent reading recommendation and cooperation analysis method, which mainly comprises the following steps:

Through publication records of students and professional fields, the core research direction of the students and the potential cross research fields thereof are determined by utilizing knowledge maps of the cross research fields; searching research materials related to the scholars according to the determined cross research field, and providing recommended reading for the scholars; collecting information and published records of a proposed student, and judging professional depth and historical research data of the proposed student in the cross research field; comparing the professional depths of the scholars and the personification scholars, judging the potential cooperation degree of the two parties, and generating a cooperation proposal for the two parties; if the professional fitness of the learner and the proposed learner reaches a preset standard, further analyzing the activity degree of the learner in the recommended cross study reading material; the cooperation will and the potential research direction of the scholars and the personification scholars are further evaluated by combining the activity degree of the scholars in the recommended materials; if both the learner and the proposed learner show the intention of cooperation and the predicted research directions are matched, automatically generating an intention-of-cooperation notification for the learner and the proposed learner; comprehensively evaluating the relative status and influence of both a scholars and a personification scholars in the potential cooperation field through academic influence indexes; after confirming that both the scholars and the personel have cooperative will, the system automatically constructs a cooperative framework, initializes the cross research projects for both sides and provides a recommended path in the subsequent research direction.

Further optionally, the determining, by the publication record and the professional field of the learner and the knowledge graph of the intersecting research field, the core research direction of the learner and the potential intersecting research field thereof includes:

Obtaining time information of all published records of a learner through database query; classifying contents published by a learner by using a random forest algorithm to obtain corresponding topic labels; matching other authors by using the theme label set, and establishing a collaboration relation graph to obtain a collaboration author list; counting the publication times of each field according to the topic label set; matching the concept in the knowledge graph by using a cosine similarity method to obtain a matching degree score; predicting the collaboration opportunity by combining the collaboration author list and the matching degree score to obtain a collaboration opportunity assessment report; using a knowledge graph query tool to obtain related nodes and weights thereof in the cross research field according to the matching degree score; according to the time distribution data of all publication records of the scholars, the random forest algorithm is used for prediction in combination with the publication frequency of the field, so that the research trend of the scholars in the research field is obtained; and (5) integrating data to obtain the core research direction and potential crossing research field of the scholars.

Further optionally, searching research materials related to the learner according to the determined cross research field to provide recommended reading for the learner, including:

Specific interests of a learner are obtained through certainty and range indexes of the cross research field; acquiring historical record data of a material read before a learner and research association degree data of the learner and other learner; carrying out time sensitivity and quality scoring on each research material by combining with release date, content quality and importance index evaluation of the material; obtaining a preliminary matching degree result according to background knowledge of a scholars, a research stage of the scholars and audience positioning of materials; analyzing feedback data and historical reading records of the former recommended materials by a learner, and optimizing the matching degree; analyzing the inherent logic association and the reference chain between materials by using ItemCF recommendation algorithm to obtain association scores; combining the matching degree result and the relevance score to generate a recommended reading list; when the study dynamics or study materials of the scholars are updated, the matching degree and the relevance score are recalculated by utilizing ItemCF recommendation algorithm, and the recommended reading list is updated.

Further optionally, the collecting information and published records of the proposed candidate, judging the professional depth and the historical research data in the cross research field, includes:

Acquiring a structured storage path of historical research data of a proposed student, and primarily sorting the data; importing standardized data of a recognized certification authority in the cross research field through a professional database, and primarily comparing the standardized data with the tidied scholars data; weight analysis is carried out on the research topics and keywords in the scholars published records based on a TF-IDF algorithm, and research topic distribution data is output; importing patent application and authorization records of a learner in the cross research field through a patent database, and integrating the patent application and authorization records into cross research patent data of the learner; obtaining the data link of the cooperation study of the scholars and other scholars, and calculating the cooperation frequency and depth through the co-occurrence matrix; the name or the identifier of the cooperator is used as input, the cooperation frequency between the cooperator and each cooperator is judged through the data of the cooperation frequency, and the cooperation depth between the cooperator and each cooperator is judged through the data of the cooperation depth; generating a cooperative network according to the cooperative relationship, the cooperative frequency and the cooperative depth among the scholars, wherein the connection among the scholars represents the cooperative relationship; importing external fund source data of cross research projects participated by the students through a project database, and analyzing the diversity of fund sources; analyzing academic conference participation data in the field of student cross study based on TF-IDF algorithm, and outputting conference participation degree data; importing the quotation times of articles published in the cross research field by a learner through an academic database to generate a quotation frequency report; importing the cross research domain ranking of the research institutions where the proposed students are located through a research institution ranking database, and comparing the global same-domain research institution ranking; the policy and regulation data related to the field of the student cross study is imported from the policy and regulation database, and key policy points related to the student study are extracted.

Further optionally, the comparing the professional depths of the scholars and the personification scholars, judging the potential cooperation degree of the two parties, and generating a cooperation proposal for the potential cooperation degree, including:

Acquiring the historical depth of the research field of the learner through the publication record of the researcher and the historical depth of the research field of the target personification learner; the method comprises the steps of obtaining cooperation research data of a scholars and other scholars through an academic search engine, and establishing a cooperation relation database; judging historical cooperation frequency and success degree of the two parties through a cooperation relation database; adopting a cosine similarity algorithm to analyze the research overlapping degree of a scholars and a personification scholars in the same field and judge the possibility of cooperation of the scholars and the personification scholars; analyzing the professional background diversity of the partner of the learner through the historical partner information, and judging the universality of the partner of the learner; analyzing cross-domain research experience of the personification according to the historical cooperation data; and scoring the influence and reputation of the two parties in the academic community by using a factoring machine, and generating a cooperation proposal for the two students in the same field, wherein the two students are higher than a preset score.

Further optionally, if the professional fitness of the learner and the proposed learner reaches the preset standard, further analyzing the activity of the learner in the recommended cross-study reading material includes:

if the research depth and width of the professional field of the scholars are higher than the preset values, acquiring all papers of the scholars in the cross research field by using an academic search engine; if the matching degree between the professional research direction of the proposed scholars and the current scholars exceeds a set threshold value, utilizing CNKI database to acquire the cooperation records of the two parties in the cross research field; judging the innovativeness and contribution degree of the learner in the recommended cross study reading material through the citation rate, and if the citation rate exceeds a set standard, considering that the innovativeness and contribution degree of the learner are high; the study depth and contribution of the analyst in the material with the number of cooperations in the cross study reading material exceeding the set value are analyzed; analyzing the update date of the recommended cross study reading material, and acquiring the contribution of a learner in the recently updated material; judging whether the student is a leading-edge researcher of the field according to the innovativeness and contribution degree ranking of the student in the cross research reading material; and obtaining the activity degree of the scholars in the recommended cross study reading material according to the citation rate, the study depth and contribution in the cross study reading material, the contribution in the recently updated material and the introduction study degree of the scholars by using a weighted average method.

Further optionally, the step of further evaluating the cooperative wish and the potential research direction of the learner and the proposed learner by combining the activity level of the learner in the recommended material includes:

Obtaining literature citation frequency in a material recommended by a learner by adopting a TF-IDF algorithm; using literature content and using TF-IDF algorithm to make frequency statistics on key words of scholars in recommended materials; acquiring the number of cooperation history records of a scholars and a personification scholars through a research history database; according to the common works or projects of the scholars and the personification scholars, counting the academic achievement sharing rate of the scholars and the personification scholars; acquiring a cooperation history record, and judging the overlapping degree of the research field related to the recommended material and the history research field of a scholars; acquiring research quality evaluation indexes of students mentioned in recommended materials through literature citation data; comparing the research methods of the scholars and the personification scholars by using recommended materials, and obtaining the similarity of the research methods by using a cosine similarity algorithm; using literature citation data to count past research results of a learner and obtain innovation indexes and research depth of the learner; and obtaining the activity degree of the scholars and the personification according to the research quality evaluation index, innovation index and research depth of the scholars, and obtaining the final evaluation result of the cooperative wish.

Further optionally, if both the learner and the proposed learner show a willingness to collaborate and the predicted study direction matches, automatically generating an intentional notification of collaboration for the same, including:

Acquiring the cooperation history of a learner and the research field label of a proposed learner; carrying out matching degree evaluation on the study field labels of the scholars and the personification scholars through a KMP algorithm; according to the detailed description of the predicted research direction, obtaining the research direction prediction results of both parties; determining whether the two parties have cooperative wishes or not by using a decision tree algorithm according to the matching degree evaluation and the research direction prediction; if both parties show the intention of collaboration and the research directions are matched, generating an intention-of-collaboration notification; sending the notification to both parties according to the content format and the sending mechanism of the intentional notification; using accuracy evaluation indexes of cooperation direction matching, using regression analysis and calculation to further evaluate the accuracy of the cooperation will; according to the evaluation result, if the matching degree of the cooperative direction is higher than a preset value, marking the cooperative intention as high matching degree in a database; if the cooperation willingness is marked as high matching degree, storing the cooperation records of the two parties in a cooperation history information storage system; the cooperation history of the learner and the personification is updated periodically, and the next matching and evaluation is adjusted.

Further optionally, the comprehensively evaluating the relative status and influence of both the scholars and the personification scholars in the potential cooperation field through academic influence indexes includes:

Obtaining published records of a scholars and published records of a proposed scholars from a plurality of academic databases; calculating the frequency of introduction and the frequency of publication of the literature to obtain academic liveness of a learner; determining contribution and a leading edge theme of the system in a specific research field by using a TF-IDF algorithm, screening journal influence factors and comparing the publication of the two parties; comparing the sponsor of the project sponsor database and the text search, and collecting academic rewards and honor records obtained by the two parties; the TF-IDF algorithm is used to extract and compare the academic impact index of both parties and the change trend of the academic impact index with time from the academic impact index database.

Further optionally, after confirming that both the learner and the proposed learner have a cooperative wish, the system automatically constructs a cooperative framework, initializes a crossover study item for both parties, and provides a recommended path for a subsequent study direction, including:

Acquiring a history cooperation record, a paper publishing direction and a research direction of a learner through a CNKI database to obtain a cooperation wish evaluation value of each learner; when the cooperative intention evaluation value of the two parties reaches or exceeds a preset threshold, the system determines that the two parties have the cooperative intention; calculating the matching degree score of the cooperation will of the two scholars in the research direction by using a VSM matching algorithm; when the matching degree score is higher than a threshold value set by a system, judging whether to generate a cooperation frame, and entering a construction stage of the cooperation frame; analyzing the study histories of the two scholars, judging the intersection of the study fields, and determining the initial study field of the cross study project; analyzing the current research trend of the market, and providing the latest research direction advice for the scholars according to the current research trend and the research field of the scholars; using a VSM matching algorithm to analyze the correlation and overlapping property of past researches of the two parties; comprehensively considering all the analysis results, and automatically generating guidance and suggestion of the cooperative research; matching degree analysis is carried out on the recommended research direction and the characteristics of a scholars, so that the pertinence and the effectiveness of the research are ensured; further comprises: and calculating the matching degree score of the cooperative wishes of the two parties by using a VSM matching algorithm.

The calculating the matching degree score of the cooperative will of the two parties by using the VSM matching algorithm specifically comprises the following steps:

And acquiring keywords or information of the research field according to the research directions of the two scholars. Keywords of study direction, paper title and abstract are converted into vector representation by TF-IDF algorithm. According to academic results of scholars, published papers and obtained patent information are obtained. And acquiring the quoted times and the h index information to obtain academic influence of the scholars. And normalizing the index of the academic influence to obtain a standardized academic influence score. A determination is made as to whether the learner has a record of team cooperation, including papers published in cooperation with other students and participating cooperation projects. According to the study plan description of the scholars, study targets, study contents and study method information are acquired. Each term is represented as a vector by TF-IDF algorithm, the value of which is determined by the frequency of the term in the text and the inverse document frequency in the entire document collection, and the study plan description is converted to a vector representation. From the vector representation, the similarity between the scholars is calculated. And determining the cooperative wish score among the students according to the similarity and the attributes of other students. And outputting the cooperative willingness scores among the scholars.

The technical scheme provided by the embodiment of the invention can have the following beneficial effects:

the invention discloses a learner cooperation method based on a knowledge graph in the cross research field. According to the method, through publication records of students and the professional field, the core research direction of the students and the potential cross research field of the students are determined by utilizing the knowledge graph of the cross research field. And searching research materials related to the scholars according to the determined cross research fields, and further providing recommended reading for the scholars. Meanwhile, information and published records of the proposed scholars are collected, and professional depth and historical research data of the proposed scholars in the cross research field are judged. And then, comparing the professional depths of the scholars and the personification scholars, judging the potential cooperation degree of the scholars and the personification scholars by using a recommendation algorithm, and generating a cooperation proposal for the potential cooperation degree. If the professional fitness of the learner and the proposed learner reaches the preset standard, further analyzing the activity level of the learner in the recommended cross-study reading material. And further evaluating the cooperative wish and potential research direction of the scholars and the personification according to the activity degree of the scholars in the recommended materials. If both the learner and the proposed learner show a willingness to collaborate and the predicted directions of study match, an intentional notification of collaboration is automatically generated for them. Through academic influence indexes, the relative status and influence of both the scholars and the personification scholars in the potential cooperation field are comprehensively evaluated. After confirming that both the scholars and the personel have strong cooperative will, the system automatically constructs a cooperative framework, initializes the cross research project for both sides and provides a recommended path for the subsequent research direction.

Drawings

FIG. 1 is a flow chart of a method for intelligent reading recommendation and collaboration analysis according to the present invention.

FIG. 2 is a schematic diagram of an intelligent reading recommendation and collaboration analysis method according to the present invention.

FIG. 3 is a schematic diagram of a smart reading recommendation and collaboration analysis method according to the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and specifically described below with reference to the drawings in the embodiments of the present invention. The described embodiments are only a few embodiments of the present invention.

The intelligent reading recommendation and collaboration analysis method in this embodiment specifically may include:

Step S101, determining the core research direction of a learner and the potential cross research field thereof by using the knowledge graph of the cross research field through the publication record of the learner and the professional field.

And obtaining time information of all published records of the scholars through database query. And classifying contents published by a scholars by using a random forest algorithm to obtain corresponding topic labels. And matching other authors by using the theme label set, and establishing a collaboration relation graph to obtain a collaboration author list. And counting the publication times of each field according to the topic label set. And matching the concept in the knowledge graph by using a cosine similarity method to obtain a matching degree score. And forecasting the collaboration opportunity by combining the collaboration author list and the matching degree score to obtain a collaboration opportunity assessment report. And obtaining related nodes and weights thereof in the cross research field according to the matching degree scores by using a knowledge graph query tool. And according to the time distribution data of all publication records of the scholars, the random forest algorithm is used for prediction in combination with the publication frequency of the domain, so that the research trend of the scholars in the research domain is obtained. And (5) integrating data to obtain the core research direction and potential crossing research field of the scholars. For example, the time information of all published records of the learner a is obtained by database query as follows, published record 1, published record 2 in 2018, published record 03 in 2018, published record 3 in 2018, published record 4 in 2018, published record 02 in 2019, published record 5 in 2019, and published record 06 in 2019. Then, classifying the content published by the learner A by using a random forest algorithm to obtain corresponding topic labels, wherein the topic label published in the record 1 is machine learning, the topic label published in the record 2 is artificial intelligence, the topic label published in the record 3 is data mining, the topic label published in the record 4 is machine learning, and the topic label published in the record 5 is natural language processing. According to the topic label set, matching other authors, and establishing a collaboration relation graph to obtain a collaboration author list as follows: partner 1, learner B, partner 2, learner C, partner 3, and learner D. According to the topic label set, counting the publishing times of each field to obtain: number of machine learning domain publication, number of 2 times, number of artificial intelligence domain publication, number of 1 time, number of data mining domain publication, number of natural language processing domain publication, number of 1 time. And matching the artificial intelligence concepts with concepts in the knowledge graph by using a cosine similarity method to obtain a matching degree score, wherein the matching degree score with the machine learning concepts is 85, the matching degree score with the artificial intelligence concepts is 75, the matching degree score with the data mining concepts is 80 and the matching degree score with the natural language processing concepts is 70. And predicting possible cooperation opportunities by combining the cooperation author list and the matching degree score to obtain a cooperation opportunity assessment report, wherein the cooperation opportunity comprises a cooperation opportunity 1, a student B, the matching degree score, 85, a cooperation opportunity 2, a student C, the matching degree score, 75, the cooperation opportunity 3, a student D and the matching degree score: 80. obtaining relevant nodes and weights thereof in the cross research field according to matching degree scores by using a knowledge graph query tool, and obtaining the cross research field 1: machine learning data mining, weight 85, cross research field 2, artificial intelligence natural language processing and weight 75. According to the time distribution data of all publication records of a scholars A, the random forest algorithm is used for prediction in combination with the publication frequency of the field, so as to obtain the research trend of the scholars A in the research field, and the research trend of the scholars A in the research field is obtained: the study of the machine learning field is emphasized from 07 in 2019 to 12 in 2019, and the study of the artificial intelligence field is emphasized from 01 in 2020 to 06 in 2020. Finally, data integration is carried out, the core research direction of a learner A is machine learning, potential research directions in the cross research field are data mining, and a cooperation opportunity assessment report and a research trend prediction result are provided.

And step S102, searching research materials related to the scholars according to the determined crossed research fields, and providing recommended reading for the scholars.

Specific interests of the learner are obtained through certainty and range indexes of the cross research field. Historical record data of the material read before the scholars and study association data of the scholars and other scholars are obtained. Time sensitivity and quality scores were performed for each study material in combination with release date, content quality and importance index evaluations of the material. And obtaining a preliminary matching degree result according to the background knowledge of the scholars, the research stage of the scholars and the audience positioning of the materials. And analyzing feedback data and historical reading records of the former recommended materials by a learner, and optimizing the matching degree. Using ItemCF recommendation algorithm, the inherent logical associations between materials and the reference chains are analyzed to obtain an association score. And combining the matching degree result and the relevance score to generate a recommended reading list. When the study dynamics or study materials of the scholars are updated, the matching degree and the relevance score are recalculated by utilizing ItemCF recommendation algorithm, and the recommended reading list is updated. For example, a list of recommended reads needs to be provided for a learner. First, it is necessary to determine research areas that the learner may be interested in, including artificial intelligence, natural language processing, computer vision, by crossing certainty and scope indicators of the research areas. Then, the history record data of the material read before the scholars and the research association degree data of the scholars and other scholars need to be obtained, and related papers and research results of the scholars in the artificial intelligence field can be found through an academic search engine. Next, time sensitivity and quality scores are performed for each study material, requiring evaluation in combination with the release date, content quality and importance index of the material. Each study material was scored according to factors such as publication time, number of citations, study value, etc. Then, according to background knowledge of a learner, a research stage in which the learner is located and audience positioning of materials, a preliminary matching degree result is obtained, and if the learner is familiar with the research in the artificial intelligence field and the paper is aimed at a beginner or a middle level learner, the paper can be preliminarily judged to be matched with the research direction of the learner. Finally, feedback data and historical reading records of the past recommended materials are analyzed by a learner, the matching degree is optimized, and if the learner is satisfied after reading the paper, the paper can be considered to be matched with the research direction of the learner. And using ItemCF recommendation algorithm, analyzing internal logic association and reference chains between materials to obtain association scores, and finding the association with other papers by analyzing keywords, research methods, experimental results and the like mentioned in the papers. And generating a recommended reading list by combining the matching degree result and the relevance score, and recommending a paper which is matched with the research direction of the scholars and has higher relevance with other papers according to the research direction and the reading interest of the scholars. When the study dynamics or study materials of the scholars are updated, the matching degree and the relevance score are recalculated by utilizing ItemCF recommendation algorithm, and the recommended reading list is updated. If a learner finds a new research hotspot when researching the artificial intelligence field, the matching degree and the relevance score can be recalculated according to the new research hotspot to generate a new recommended reading list.

And step S103, collecting information and published records of the proposed students, and judging the professional depth and historical research data of the proposed students in the cross research field.

And obtaining a structured storage path of the history study data of the proposed scholars, and carrying out preliminary arrangement on the data. And importing standardized data of a recognized certification authority in the cross research field through a professional database, and performing preliminary comparison with the tidied scholars data. And carrying out weight analysis on the study subjects and keywords in the scholars published records based on the TF-IDF algorithm, and outputting study subject distribution data.

The patent application and the authorized record of the scholars in the cross research field are imported through the patent database and integrated into the cross research patent data of the scholars. And obtaining the data link of the cooperation study of the scholars and other scholars, and calculating the cooperation frequency and the cooperation depth through the co-occurrence matrix. The name or the identifier of the cooperator is used as input, the cooperation frequency between the cooperator and each cooperator is judged through the cooperation frequency data, and the cooperation depth between the cooperator and each cooperator is judged through the cooperation depth data. And generating a cooperative network according to the cooperative relationship, the cooperative frequency and the cooperative depth among the scholars, wherein the connection among the scholars represents the cooperative relationship. The diversity of funding sources is analyzed by importing external funding source data of intersecting study projects in which the learner participates through the project database. And analyzing academic conference participation data in the field of student cross study based on the TF-IDF algorithm, and outputting conference participation data. And importing the article quoted times in the cross research field by a student through an academic database to generate a quote frequency report. The ranking of the cross research fields of the research institutions where the proposed students are located is imported through the research institution ranking database, and the ranking of the global research institutions in the same field is compared. The policy and regulation data related to the field of the student cross study is imported from the policy and regulation database, and key policy points related to the student study are extracted. For example, suppose that standardized data of a certification authority is imported through a professional database, in which there are 100 scholars' information. The names, research fields, academic achievements and the like of the scholars are organized into a structured storage path which is stored in a database. For one of the students, he published 30 papers in the last five years, assuming that his study topics included artificial intelligence, machine learning, and data mining. And (3) carrying out weight analysis on the published records by using a TF-IDF algorithm to obtain the study subjects of the scholars, wherein the artificial intelligence weight is 4, the machine learning weight is 6, the data mining weight is 3, and the deep learning weight is 5, the algorithm weight is 4 and the big data weight is 3 in the studied keywords. This represents a more prominent study by the learner in terms of machine learning, and relatively less in terms of artificial intelligence and data mining. In addition, patent application and authorization records of the scholars in the cross research field are also imported. He filed a total of 10 patents in the last five years and successfully granted 5 of them. A link of the collaboration study data of the learner with other students is imported, and the collaboration frequency and depth of each collaboration learner are calculated through the co-occurrence matrix. The co-occurrence matrix is used to represent the collaboration relationship of the scholars as a two-dimensional matrix, with each element in the matrix representing the frequency of collaboration between two scholars. By calculating the sum of the rows or columns in the matrix, the frequency of collaboration of each learner with the other students can be obtained. One of the cooperatives, the cooperatives published 5 papers with a frequency of 5 and a depth of 4. External funding source data for the crossover study in which the learner participated is imported. He participated in 3 projects, wherein the fund sources were 50% of national natural science foundation, 30% of enterprise sponsor and 20% of other research institutions, respectively, indicating that the study fund sources of the scholars are diversified. Academic conference participation data for a learner who participated in 10 academic conferences in the past five years, his conference participation degree can be analyzed according to TF-IDF algorithm. He has more published times in some important conferences and higher participation, while other conferences have fewer published times and lower participation. Article cited times data of the scholars in the cross research field are imported. His article is cited 200 times in total, and a report of the frequency of the citations can be generated to obtain the influence degree of his research results in the field. The cross study domain ranking of the study where the learner is located is imported from the study ranking database and compared to the global co-domain study. The study institution where the scholars are located is 10 times in the global ranking, which shows that the institution has higher reputation and influence in the cross study field. In the policy and regulation database, policy and regulation data related to the field of the student's cross study is imported, and key policy points related to the student's study are extracted. A policy has important significance for protecting the data privacy in the field of artificial intelligence, and the research result of the learner is closely related to the policy.

Step S104, the professional depths of the scholars and the personification scholars are compared, the potential cooperation degree of the scholars and the personification scholars is judged, and a cooperation proposal is generated for the potential cooperation degree.

And acquiring the historical depth of the research field of the learner through the published record of the researcher and the historical depth of the research field of the target personification learner. And acquiring cooperation research data of the scholars and other scholars through an academic search engine, and establishing a cooperation relation database. And judging the historical cooperation frequency and the success degree of the two parties through the cooperation relation database. And analyzing the research overlapping degree of the scholars and the personification scholars in the same field by adopting a cosine similarity algorithm, and judging the possibility of cooperation of the scholars and the personification scholars. And analyzing the professional background diversity of the partner of the learner through the historical partner information, and judging the universality of the partner of the learner. And analyzing cross-domain research experience of the personification according to the historical cooperation data. And scoring the influence and reputation of the two parties in the academic community by using a factoring machine, and generating a cooperation proposal for the two students in the same field, wherein the two students are higher than a preset score. For example, researcher A published 50 papers on the field of artificial intelligence over the last 10 years, indicating that he has a deeper research history in this field. He now wishes to work with researcher B who published 20 papers on artificial intelligence in the last 5 years, indicating that he also has a certain research history. Through academic search engine CNKI, obtaining the data links of the cooperation study of the scholars and other scholars finds that the researcher A and the researcher B commonly published 5 papers in the last 3 years, and these papers get higher scores in peer reviews, which indicates that their cooperation success is higher. And (3) calculating literature similarity by adopting a cosine similarity algorithm, and analyzing research overlapping degree of researchers A and B in the artificial intelligence field. They have published a total of 100 papers in the last 10 years, 30 of which relate to similar topics of research, indicating that they have a high interest and overlap in the same field. Through historical partner information analysis, the partner of the researcher A is found to come from different professional backgrounds, including fields of computer science, data science, psychology and the like, which shows that the partner of the researcher A has higher cooperative universality. Based on historical collaboration data, it was found that researcher B has been collaborating with partners of different domains, including machine learning, computer vision, and natural language processing, over the last 5 years, indicating that he has experience in cross-domain research. The influence of researchers a and B in the academic community was scored using a factoring machine. The academic dynamic update frequency index of the scholars is evaluated according to the number of the recently published papers of the scholars, the cited conditions of the papers and the conditions of the academic conferences and lectures of the participations or organizations. In the past year, the academic dynamic update frequency index of the academic da is 5+3+2+1=11, wherein 3 sheets are cited for more than 10 times and take part in 2 international academic conferences and organize one academic lecture. The influence score for researcher A was 8, while that for researcher B was 6. Meanwhile, according to their published records for the past year, the academic dynamic update frequency of researcher a was 3 papers per month, and the academic dynamic update frequency of researcher B was 2 papers per month. The academic dynamic update frequency index of the scholars A is 11, the scoring result is 8, and the academic dynamic update frequency index of the proposed scholars B is 10, and the scoring result is 7. The scoring result of the learner A and the personification learner B in the artificial intelligence field is higher than a preset value 6, and the system generates a collaborative proposal to suggest that the researcher A and the researcher B together research an item about the artificial intelligence field to fully utilize the research history and experience of the researcher A and the personification learner B in the field.

In step S105, if the professional fitness of the learner and the proposed learner reaches the preset standard, the activity of the learner in the recommended cross-study reading material is further analyzed.

If the research depth and width of the professional field of the scholars are higher than the preset values, all papers of the professional field of the scholars are acquired through an academic search engine. If the matching degree between the professional research direction of the proposed scholars and the current scholars exceeds a set threshold value, the CNKI database is utilized to acquire the cooperation records of the two parties in the cross research field. And judging the innovativeness and contribution degree of the scholars in the recommended cross study reading material through the citation rate, and if the citation rate exceeds a set standard, considering the innovativeness and contribution degree of the scholars to be high. The analyst's study depth and contribution in the material where the number of cooperations in the cross-study reading material exceeds the set point. The update date of the recommended cross-study reading material is analyzed to obtain the contribution of the learner in the recently updated material. Whether the student is a leading-edge researcher of the field is judged according to the innovativeness and contribution degree ranking of the student in the cross research reading material. And obtaining the activity degree of the scholars in the recommended cross study reading material according to the citation rate, the study depth and contribution in the cross study reading material, the contribution in the recently updated material and the introduction study degree of the scholars by using a weighted average method. For example, the preset value is a study depth of 5 for the learner and a study width of 3. The study depth of scholars a was 7, the study width was 4, and exceeded the preset value. All papers of the scholars A in the cross research field are obtained through an academic search engine, and 30 papers are obtained. At present, the learner B is an anthropomorphic learner, and the matching degree between the professional research direction of the learner B and the learner A is 8, and the matching degree exceeds a set threshold. And obtaining cooperation records of the scholars A and B in the cross research field through CNKI databases, wherein the total number of the cooperation records is 10. The innovation and contribution degree of the learner A in the recommended cross-study reading material are judged through the citation rate. Among them, 5 papers are cited by other scholars, the cited rate is 50%, and the setting standard is exceeded. The depth of investigation and contribution of learner a in materials with a number of cooperations exceeding the set point were analyzed. There are 3 collaborative papers with depths of investigation of 8, 9 and 10, respectively, contributing 6, 7 and 8, respectively. Analysis recommended date of update of cross-study reading material, assuming that learner a was recently updated material was a paper published in the past year, indicating that he contributed recently. Whether or not student A is the leading researcher of the field is judged according to the innovativeness and contribution degree ranking of student A in the cross research reading material. He ranks 2 nd, 3 rd and 2 nd, respectively, in terms of citation rate, study depth and contribution, contribution in recently updated material, etc. Using a weighted average approach, the degree of activity of learner A in these materials is derived from his reference rate, depth of investigation and contribution in the recommended cross-study reading materials, contribution in recently updated materials, and the degree of introduction study. The reference rate weight is 4, the research depth and contribution weight are 3, the contribution weight of the recently updated material is 2, the leading-edge research degree weight is 7, and the activity degree of the learner A is 16.

Step S106, the cooperation will and the potential research direction of the scholars and the personification scholars are further evaluated by combining the activity degree of the scholars in the recommended materials.

And obtaining literature citation frequency in the material recommended by the scholars by adopting a TF-IDF algorithm. Using literature content, frequency statistics are performed on the keywords of the learner in the recommended materials using TF-IDF algorithm. And acquiring the number of the cooperation histories of the scholars and the personification scholars through the research history database. And counting the academic achievement sharing rate of the scholars and the personification according to the common works or projects of the scholars and the personification. And acquiring a cooperation history record, and judging the overlapping degree of the research field related to the recommended material and the history research field of the scholars. The study quality evaluation index of the scholars mentioned in the recommended materials is obtained by literature reference data. And comparing the research methods of the scholars and the personification scholars by using recommended materials, and obtaining the similarity of the research methods by using a cosine similarity algorithm. And using literature citation data to count past research results of the scholars and obtain innovation indexes and research depth of the scholars. And obtaining the activity degree of the scholars and the personification according to the research quality evaluation index, innovation index and research depth of the scholars, and obtaining the final evaluation result of the cooperative wish. For example, there is a learner A who is referenced 5 times in the recommended materials. Through the TF-IDF algorithm, the literature reference frequency of learner A in the recommended materials is calculated. In addition, scholars a have 10 keywords in the recommended material, and the TF-IDF algorithm is used to count the frequency of each keyword. Keyword machine learning occurs 20 times in the recommended material with a TF-IDF value of 20. Through CNKI databases, the number of collaboration history records for learner A and proposed learner B may be obtained. They have completed 3 projects together, then their number of collaboration histories is 3. According to the common works or projects of the scholars A and the personification scholars B, the academic achievement sharing rate of the scholars A and the personification scholars B can be counted. Their common work amount is 2, then their academic achievement sharing rate is 2/3. After the cooperation history record is acquired, the overlapping degree of the research field related to the recommended material and the history research field of the scholars A is judged. The number of documents in which the recommended material relates to the study area overlapping with the scholars a history study area is 10, and the overlapping degree is 10/20. The study quality evaluation index of the scholars mentioned in the recommended material can be obtained by literature reference data. The study quality evaluation index of the learner a was 8, and the study quality evaluation index of the learner mentioned in the recommended material was 8. And comparing the research methods of the scholars A and the personification scholars B by using recommended materials, and obtaining the similarity of the research methods by using a cosine similarity algorithm. The similarity of the study methods of scholars A and personification scholars B is 9. Through literature citation data, past research results of a student A are counted, and innovation indexes and research depths of the student A are obtained. Among the recommended materials, 10 papers of scholars a were mentioned, of which 7 papers were cited more often by others, each cited 50 times on average, and the other 5 papers were cited less often, each cited 10 times on average. The study quality evaluation index of the learner a was 70%. It was found that learner A published a total of 20 papers related to computer science in the last 5 years, and mentioned 10 papers related to artificial intelligence. The degree of overlap of the historic study area of the learner a with the study area related to the recommended material is 50%. The innovation index of the scholar A is 7, and the research depth is 5. And obtaining the activity degree of the scholars A and the personification scholars B according to the research quality evaluation index, the innovation index and the research depth of the scholars, and obtaining the final evaluation result of the cooperative will. The activities of the scholars a and the personification scholars B are 8 and 9 respectively, and the final evaluation result of the cooperative will is 8*9 =72.

Step S107, if both the learner and the proposed learner show a intention to cooperate and the predicted study directions match, an intention to cooperate notification is automatically generated for the same.

And acquiring the cooperation history of the scholars and the research field labels of the personification scholars. And carrying out matching degree evaluation on the study field labels of the scholars and the personification scholars through a KMP algorithm. And obtaining the research direction prediction results of the two parties according to the detailed description of the prediction research direction. And determining whether the two parties have cooperative wishes or not by using a decision tree algorithm according to the matching degree evaluation and the research direction prediction. If both parties show willingness to collaborate and the research directions match, an intentional notification of collaboration is generated. And sending the notification to both sides according to the content format and the sending mechanism of the intentional notification. And further evaluating the accuracy of the match intention by using regression analysis calculation by using the accuracy evaluation index of the match of the cooperative directions. And according to the evaluation result, if the matching degree of the cooperative direction is higher than a preset value, marking the cooperative intention as high matching degree in the database. And if the cooperation willingness is marked as high matching degree, storing the cooperation records of the two parties in a cooperation history information storage system. The cooperation history of the learner and the personification is updated periodically, and the next matching and evaluation is adjusted. For example, there is a student library containing personal information of student a, for example, the research area is mathematics and computer science. There is also a library of proposed authors, containing research domain labels for the proposed authors, e.g. the research domain is artificial intelligence. First, the KMP algorithm is used to calculate the matching degree of the study area labels among the study area labels of the learner a and the personifier. LSTM is then used to predict likely directions of study for learner a and the proposed person in the future based on a detailed description of the predicted directions of study. And then, judging whether the learner A and the persuader have cooperative wishes or not by using a decision tree algorithm according to the matching degree evaluation and the research direction prediction. If both parties show willingness to collaborate and the research directions match, an intentional notification of collaboration is generated. And sending the notification to both sides according to the content format and the sending mechanism of the intentional notification. And finally, using accuracy evaluation indexes of cooperative direction matching, and using regression analysis to calculate the accuracy of cooperative will of the learner A and the personification when the cooperative direction matching degree is high. And according to the evaluation result, if the matching degree of the cooperation direction is high, marking the cooperation intention as high matching degree in the database. And if the cooperation willingness is marked as high matching degree, storing the cooperation records of the two parties in a cooperation history information storage system. The cooperation history of the learner and the personification is updated periodically, and the next matching and evaluation is adjusted.

And S108, comprehensively evaluating the relative status and influence of both the scholars and the personification scholars in the potential cooperation field through academic influence indexes.

The published records of the scholars and the published records of the proposed scholars are obtained from a plurality of academic databases. And calculating the frequency of introduction and the frequency of publication of the literature to obtain the academic liveness of the scholars. The TF-IDF algorithm is used for determining the contribution and the leading edge theme of the TF-IDF algorithm in the specific research field, and journal influence factors are screened and published by the two parties are compared. And comparing the sponsors of the two through a project sponsoring database by using text retrieval, and collecting academic rewards and honor records obtained by the two parties. The TF-IDF algorithm is used to extract and compare the academic impact index of both parties and the change trend of the academic impact index with time from the academic impact index database. For example, two scholars A and B are selected for comparison. The acquisition of student a published 10 papers in the database, with a total of 50 cited frequencies, with a publication frequency of 2 papers per year. Scholar B published 15 papers, with a total of 80 cited frequencies and 3 papers per year. By calculating the frequency of introduction and the frequency of publication, the academic activity of the scholars A is 25, and the academic activity of the scholars B is 26.7. Using the TF-IDF algorithm, the contribution and leading edge topic of scholars a, scholars B in a particular study area are determined. The contribution degree of the scholars A is 8, the leading-edge theme is artificial intelligence, the contribution degree of the scholars B is 7, and the leading-edge theme is machine learning. By screening the journal influence factor, it was found that the published paper average influence factor of scholar a was 5, and that of scholar B was 0. This suggests that learner a is publishing more papers on high impact journals, and may have more impact. In the project funding database, it was found that learner A obtained 3 project funding for a total of $100 ten thousand, while learner B obtained 5 project funding for a total of $150 ten thousand. This illustrates that learner B was more successful in subsidizing the project. And extracting and comparing academic impact indexes of two scholars from the academic impact index database through a TF-IDF algorithm. The academic impact index of scholars a is 9 and the academic impact index of scholars B is 8. It is concluded that the relative position and academic impact of learner a in the potential collaboration area is high, while the relative position and academic impact of learner B in the potential collaboration area is low.

Step S109, after confirming that both the scholars and the personification scholars have cooperative will, the system automatically constructs a cooperative framework, initializes the cross research projects for both sides and provides a recommended path of the subsequent research direction.

And acquiring a history cooperation record, a paper publishing direction and a research direction of the scholars through CNKI databases to obtain a cooperation wish evaluation value of each scholars. When the cooperative intention evaluation value of the two parties reaches or exceeds a preset threshold, the system determines that the two parties have the intention of cooperation. And calculating the matching degree score of the cooperation intention of the two scholars in the research direction by using a VSM matching algorithm. And when the matching degree score is higher than a threshold value set by the system, judging whether to generate a cooperation frame, and entering a construction stage of the cooperation frame. And analyzing the study histories of the two scholars, judging the intersection of the study fields, and determining the initial study field of the cross study project. Analyzing the current research trend of the market, and providing the latest research direction advice for the scholars according to the current research trend and the research field of the scholars. The correlation and overlap of the past studies of both parties were analyzed using a VSM matching algorithm. And (3) automatically generating guidance and advice of the collaborative research by comprehensively considering all the analysis results. And carrying out matching degree analysis on the recommended research direction and the characteristics of the scholars, and ensuring the pertinence and the effectiveness of the research. For example, there are two scholars a and B who have published 10 papers in the past in collaboration, and the frequency of collaboration is high. By analyzing the history cooperation records of the scholars a and B, the evaluation value of their cooperation will is obtained, the evaluation value of the scholars a is 8, and the evaluation value of the scholars B is 7. According to the preset threshold value of 6, the two parties can be judged to have the intention of being in charge. Next, a matching degree score of the cooperative wishes of both parties is calculated using a VSM matching algorithm. Their matching degree score is 9, which is higher than the threshold value 8 set by the system, and it can be judged that the collaboration frame is generated and the construction stage of the collaboration frame is entered. When analyzing the study history of two scholars, a certain intersection exists in the study field of the scholars, and particularly in the artificial intelligence field. Thus, the initial field of research that determines the crossover study is artificial intelligence. Meanwhile, the current research trend of the market is analyzed, and the machine learning is found to be widely applied in the field of artificial intelligence. According to the current research trend and the research fields of scholars a and B, the latest research direction suggestions are provided for the scholars, such as deep learning application in the natural language processing field. Then, the correlation and overlap of both past studies were analyzed using a VSM matching algorithm. Their relevance and overlap score is 8, and by comprehensively considering all the previous analysis results, guidance and advice of collaborative research are automatically generated, for example, in the artificial intelligence field, they can jointly study the application of deep learning in the natural language processing field. And finally, carrying out matching degree analysis on the recommended research direction and the characteristics of the scholars, and ensuring the pertinence and the effectiveness of the research. His degree of matching in the direction of the study can be assessed by analyzing the professional expertise of learner a and his study experience in the field of natural language processing.

And calculating the matching degree score of the cooperative wishes of the two parties by using a VSM matching algorithm.

And acquiring keywords or information of the research field according to the research directions of the two scholars. Keywords of study direction, paper title and abstract are converted into vector representation by TF-IDF algorithm. According to academic results of scholars, published papers and obtained patent information are obtained. And acquiring the quoted times and the h index information to obtain academic influence of the scholars. And normalizing the index of the academic influence to obtain a standardized academic influence score. A determination is made as to whether the learner has a record of team cooperation, including papers published in cooperation with other students and participating cooperation projects. According to the study plan description of the scholars, study targets, study contents and study method information are acquired. Each term is represented as a vector by TF-IDF algorithm, the value of which is determined by the frequency of the term in the text and the inverse document frequency in the entire document collection, and the study plan description is converted to a vector representation. From the vector representation, the similarity between the scholars is calculated. And determining the cooperative wish score among the students according to the similarity and the attributes of other students. And outputting the cooperative willingness scores among the scholars. For example, there are two students, a and B, whose study directions are computer vision and natural language processing, respectively. Based on their academic results, student a has published 15 papers and obtained 2 patents, while student B published 12 papers and obtained 1 patent. One paper of a scholars A is entitled to study on image recognition technology based on deep learning, and abstract provides an image recognition algorithm based on deep learning for the text, and accurate recognition of objects in images is achieved through a large amount of training data and convolutional neural networks. Keywords in the headlines and abstracts are converted into vector representations according to frequencies in texts and inverse document frequencies in the whole document set by a TF-IDF algorithm, and the vector representations comprise deep learning, image recognition and convolutional neural networks. According to the 15 papers of the scholars A, the cited times are more than 10 times, the h index is 10, and the academic influence of the scholars is obtained. And normalizing the index of the academic influence to obtain a standardized academic influence score. The normalized chemical influence score for scholar a was obtained as 8. Determining whether learner A has a record of team cooperation may see if he has published papers in cooperation with other students or participated in a collaborative project. The student A cooperates with the other two students to release three papers and participate in two cooperation projects. And acquiring information such as study targets, study contents, study methods and the like according to the study plan description of the learner A. The study objective of the learner a is to study an image processing algorithm based on deep learning, wherein the study contents comprise image recognition, object detection and image generation, and the study method mainly uses a convolutional neural network and generates an countermeasure network. The study plan description of scholar a is converted into a vector representation by TF-IDF algorithm. Keywords in the study plan description include deep learning, image processing and image recognition. The vector of each keyword is determined by its frequency in text and the inverse document frequency in the entire document collection. From the vector representation, the similarity between learner a and learner B may be calculated. The keyword vector representations of the scholars a and B in terms of study direction, academic achievement, study plan description, and the like are [ a1, a2, a3] and [ B1, B2, B3], respectively. The cosine similarity can be used to calculate the similarity between them, which is 9. Based on the similarity and other learner's attributes, a collaborative willingness score between learner a and learner B may be determined. The similarity is 9, the academic impact score of the scholars a is 8, and the academic impact score of the scholars B is 6, and then the cooperative willingness score can be calculated as the similarity multiplied by the sum of the academic impact scores of the two scholars, i.e., 9×8+6=126. The final output, student a and student B, have a collaborative intent score of 126, indicating that they have a higher collaborative intent.

It will be evident to those skilled in the art that the application is not limited to the details of the foregoing illustrative embodiments, and that the present application may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.

Claims

1. A method for intelligent reading recommendation and collaboration analysis, the method comprising:

Through publication records of students and professional fields, the core research direction of the students and the potential cross research fields thereof are determined by utilizing knowledge maps of the cross research fields; searching research materials related to the scholars according to the determined cross research field, and providing recommended reading for the scholars; collecting information and published records of a proposed student, and judging professional depth and historical research data of the proposed student in the cross research field; comparing the professional depths of the scholars and the personification scholars, judging the potential cooperation degree of the two parties, and generating a cooperation proposal for the two parties; if the professional fitness of the learner and the proposed learner reaches a preset standard, further analyzing the activity degree of the learner in the recommended cross study reading material; the cooperation will and the potential research direction of the scholars and the personification scholars are further evaluated by combining the activity degree of the scholars in the recommended materials; if both the learner and the proposed learner show the intention of cooperation and the predicted research directions are matched, automatically generating an intention-of-cooperation notification for the learner and the proposed learner; comprehensively evaluating the relative status and influence of both a scholars and a personification scholars in the potential cooperation field through academic influence indexes; after confirming that both the scholars and the personification scholars have cooperative will, the system automatically constructs a cooperative framework, initializes the cross research projects for both sides and provides a recommended path of the subsequent research direction; wherein the searching research materials related to the scholars according to the determined cross research field, and providing recommended reading for the scholars comprises the following steps:

Determining a specific research interest of a learner; acquiring a history reading material record of a learner and research association data of the learner; evaluating the time sensitivity and quality of each study material; obtaining a preliminary matching degree result; analyzing the history reading record and optimizing the matching degree; analyzing the association between materials by using ItemCF recommendation algorithm to obtain association scores; and generating a recommended reading list by combining the matching degree result.

2. The method of claim 1, wherein the determining the core study direction of the learner and the potential cross-study area thereof by the published records of the learner and the professional area using the knowledge graph of the cross-study area comprises:

Obtaining time information of all published records of a learner through database query; classifying the publication content of a scholars by using a random forest algorithm to obtain a theme tag; matching other authors by using the theme label set, and establishing a cooperation relation graph; counting the publication times of each field; matching the cosine similarity with concepts in the knowledge graph to obtain a matching degree score; predicting a collaboration opportunity in combination with a collaboration author list; and obtaining the related node weight in the cross research field by using a knowledge graph query tool.

3. The method of claim 1, wherein the collecting information and published records of the proposed candidate, determining its expertise in the cross-study area and historical study data, comprises:

Acquiring and sorting historical research data of a proposed student; the standardized data imported into the cross research field is compared with the scholars data; analyzing the student to release records by using the TF-IDF algorithm weight; importing the crossed research patent data of the scholars; calculating the frequency and depth of cooperative research of a learner; generating a collaboration network of the learner; importing the data of the crossed research fund sources of the scholars, and analyzing the fund sources; analyzing academic conference participation data in the field of crossed study of scholars; the article of the importer is quoted for times to generate a quote frequency report; comparing the research institution ranking of the fitting scholars; policy and regulatory data related to the study of the learner is imported.

4. The method of claim 1, wherein comparing the professional depths of the learner and the proposed learner, determining potential collaboration agreements of the two parties, and generating a collaboration proposal therefor, comprises:

Acquiring the historical depth of the research field of a learner and the historical depth of the research field of a target simulation learner; establishing a cooperation relation database; judging historical cooperation frequency of the two parties through a cooperation relation database; adopting a cosine similarity algorithm to judge the cooperation of the two parties; analyzing the professional background diversity of the partner of the scholars; scoring both parties using a factorizer; collaborative proposals are generated for both scholars with high scores.

5. The method of claim 1, wherein if the professional fitness of the learner with the proposed learner meets a preset standard, further analyzing the activity of the learner in the recommended cross-study reading material comprises:

All papers of the scholars in the cross research field are acquired through an academic search engine; utilizing CNKI databases to obtain cooperation records of both parties; judging the innovativeness of a learner through the citation rate; analyzing the research depth of a learner in the recommended material; analyzing the update date of the recommended material, and obtaining the contribution of the scholars; the weighted average method is used to obtain the activity level of the scholars in the cross research material.

6. The method of claim 1, wherein the further evaluating the intention of the learner to cooperate with the proposed learner and the potential direction of the study in combination with the activity of the learner in the recommended material comprises:

Obtaining literature citation frequency by adopting a TF-IDF algorithm; counting keyword frequencies by using a TF-IDF algorithm; the method comprises the steps of obtaining the number of cooperation history records of a learner and a proposed learner; counting the sharing rate of academic achievements; acquiring the overlapping degree of the research field related to the recommended material and the history research field of the scholars; statistics of the study quality evaluation index of the scholars; obtaining similarity of a research method by using a cosine similarity algorithm; innovation index and research depth of statistics scholars; and obtaining an evaluation result of the cooperative wish.

7. The method of claim 1, wherein if both the learner and the proposed learner exhibit a willingness to collaborate and the predicted study direction matches, automatically generating an intentional notification of collaboration for the same, comprising:

Acquiring the cooperation history of a learner and the research field label of a proposed learner; evaluating the research field matching degree and cooperation willingness of the two parties by utilizing a KMP algorithm and a decision tree algorithm; generating and sending an intentional notice of collaboration to both parties; carrying out regression analysis and evaluation on the matching degree, and marking the cooperation intention in a database according to the result; if the mark is high in matching degree, the cooperation records are stored and the cooperation history of the scholars is updated periodically.

8. The method of claim 1, wherein the comprehensively evaluating the relative status and influence of both the learner and the proposed collaboration learner in the potential collaboration area by the academic impact index comprises:

Obtaining published records of both parties from an academic database; calculating the frequency of the cited documents and the published frequency; determining contribution and leading edge topics of the research field by using a TF-IDF algorithm, and screening journal influence factors; comparing the sponsored conditions and academic rewards of the two from the project sponsored database by using text retrieval; and extracting and comparing academic impact indexes of the two parties.

9. The method of claim 1, wherein the system automatically builds a collaboration framework after confirming that both the learner and the proposed collaboration learner have a will to collaborate, initializes the crossover study for both parties, and provides a recommendation path for the subsequent study direction, comprising:

Utilizing CNKI databases to obtain historical cooperation records and research directions of the two parties; when the evaluation value of the cooperative intention reaches a preset value, the system determines the cooperative intention; calculating the matching degree of the research direction by using a VSM matching algorithm; analyzing research histories of both parties to determine the field of the crossed research project; providing research direction suggestions according to market trends, and synthesizing the previous analysis to generate guidance and suggestions of collaborative research; further comprises: calculating the matching degree score of the cooperative will of the two parties by using a VSM matching algorithm;

The calculating the matching degree score of the cooperative will of the two parties by using the VSM matching algorithm specifically comprises the following steps: acquiring keywords or information of the research field according to the research directions of the two scholars; converting keywords of research directions, paper titles and abstracts into vector representations through a TF-IDF algorithm; according to academic achievements of scholars, published papers and obtained patent information are obtained; acquiring the quoted times and h index information to obtain academic influence of a scholars; normalizing the index of the academic influence to obtain a standardized academic influence score; judging whether the scholars have records of team cooperation, including papers published in cooperation with other scholars and participated cooperation projects; acquiring research targets, research contents and research method information according to the research plan description of a learner; each word is expressed as a vector through a TF-IDF algorithm, the value of the vector is determined by the frequency of the word in the text and the inverse document frequency in the whole document set, and the study plan description is converted into a vector expression; calculating the similarity between the scholars according to the vector representation; determining cooperative willingness scores among the students according to the similarity and the attributes of other students; and outputting the cooperative willingness scores among the scholars.