Recommendation method for examination and research colleges

Abstract

The invention relates to the technical field of college recommendation information processing, in particular to an examination college recommendation method. Comprising the following steps of S1, acquiring an intention specialty of a user; s2, acquiring a college specialty candidate set, wherein the college specialty candidate set comprises a plurality of college specialty samples; s3, calculating feature scores corresponding to the college specialty samples, and sorting the college specialty samples in the college specialty candidate set according to the feature scores; the process of calculating the feature score corresponding to the college specialty sample is that weight coefficients are preset for a plurality of feature elements of the college specialty sample and differentiated score assignment is carried out, and the feature score corresponding to the college specialty sample is equal to the sum obtained by multiplying the assignment of each feature element by the corresponding weight coefficient. And S4, according to the sorting result of the step S3, taking a preset number of college specialty samples from the front i (1-p) of the sorting to the back, and recommending the college specialty samples to the user. Recommendable college specialties can be expanded, and college specialties with concentrated user registration and mismatched registration capabilities are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of college recommendation information processing, and in particular to a method for recommending colleges for postgraduate entrance examinations. Background Art

[0002] Postgraduate examinations are an important way to improve academic qualifications in higher education. Educational training institutions often rely on experience or judgment or simply on the users' own decisions to guide users in choosing majors and schools when applying for postgraduate studies. The limitation is that there is no overall statistics on the current postgraduate enrollment schools and majors, and there is no objective assessment of the users' abilities. This will make it difficult for users to choose schools and majors, and they will overestimate or underestimate their own abilities, resulting in poor matching effects. There will also be a situation where users concentrate on applying for a certain school and major, resulting in fierce competition among users of the same educational training institution. In particular, the small number of schools and majors in some regions makes it easier for users who intend to apply to the region to concentrate on applying for a certain school and major. Summary of the invention

[0003] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method for recommending postgraduate entrance examination institutions, which can expand the recommended institutions and majors and avoid users from applying for concentrated applications and applying for institutions and majors that do not match their abilities.

[0004] In order to achieve the above object, the present invention is implemented through the following technical solutions: A method for recommending a postgraduate school comprises the following steps S1 to S4: S1. Obtain the user's intended major; S2. Obtain a candidate set of college majors, which includes several college major samples, including the following steps S21-S22: S21. Add basic college major samples to the college major candidate set, wherein the basic college major samples are college major samples that fully meet the user's personalized information and intended majors, and the user's personalized information includes target regions, learning forms, and test subject exclusion characteristics; S22, adding the school major samples that meet the user's personalized information and have a high similarity with the user's intended major to the school major candidate set, the specific process includes steps S221-S223: S221, obtaining a direction subject set corresponding to each college professional sample, wherein the direction subject set includes a research direction element and an examination subject element corresponding to the college professional sample; S222, calculate the similarity coefficient J(A,B)=|A∩B| / |A∪B| between each basic college professional sample and other college professional samples, where A∩B is the number of elements in the intersection between the direction subject set of college major A and the direction subject set of college major B, and A∪B is the number of elements in the union between the direction subject set of college major A and the direction subject set of college major B; S223, adding a preset number of university major samples with the highest similarity coefficients among other university major samples to the university major candidate set; S3, calculating the characteristic scores corresponding to the university and major samples, and sorting the university and major samples in the university and major candidate set according to the characteristic scores; The process of calculating the characteristic scores corresponding to the university professional samples is as follows: preset weight coefficients for several characteristic elements of the university professional samples and perform differentiated scoring. The characteristic scores corresponding to the university professional samples = the sum of the scores of each characteristic element multiplied by the corresponding weight coefficients; S4. According to the ranking result of step S3, a preset number of school and major samples are taken from i(1-p) before the ranking from high score to low score and recommended to the user, where i is the total number of school and major samples in the school and major candidate set, p is the user's ability coefficient, , Score u Score is the current user ability evaluation score. min Score is the lowest score of user ability evaluation within the statistical range. max The highest user capability assessment score within the statistical scope.

[0005] A postgraduate entrance examination institution recommendation system, used to implement the postgraduate entrance examination institution recommendation method, comprising: A user major intention acquisition unit, used to acquire the user's intended major; A college and major candidate set generation unit is used to screen college and major samples to generate a college and major candidate set; The college and major sample ranking unit is used to calculate the characteristic scores corresponding to the college and major samples in the college and major candidate set, and to sort the college and major samples in the college and major candidate set according to the characteristic scores; A user capability coefficient calculation unit is used to obtain the user's capability elements to calculate the user capability evaluation score, count the user capability evaluation scores in the system, and calculate the user capability coefficient of the current user; The school recommendation generation unit is used to select a number of school and major samples according to the user's ability coefficient and the ranking of the school and major samples in the school and major candidate set to generate corresponding recommendation information to inform the user.

[0006] It can be seen from the above technical solution that the present invention has the following beneficial effects: 1. The present invention provides a method for recommending postgraduate entrance examination institutions, which adds samples of institutions and majors that meet the personalized information of users and have a high similarity with the user's intended majors to a candidate set of institutions and majors to expand the samples of institutions and majors in the candidate set of institutions and majors, thereby avoiding the situation where the number of recommended institutions and majors is too small and user registrations are concentrated.

[0007] 2. The present invention provides a method for recommending postgraduate schools, which can match appropriate schools and majors to users based on the user's ability coefficient and the characteristic scores corresponding to the school and major samples, thereby preventing users from overestimating or underestimating their own abilities when selecting schools and majors, and also preventing user groups from concentratedly registering for a certain school and major, resulting in unbalanced registration. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is a schematic diagram of the corresponding content of the structured table data in the embodiment of the present application. DETAILED DESCRIPTION

[0009] Example 1 A method for recommending a postgraduate school comprises the following steps S1 to S4: S1. Obtain the user's intended major. In this embodiment, if the user does not clearly specify the intended major to apply for in step S1, a major recommendation module is also included to recommend a major to the user according to the user's needs.

[0010] In this embodiment, the major recommendation module includes an original major guidance recommendation module, which is used to make a major recommendation based on the user's original major, including the following steps S111-S113: S111. Obtain the corresponding postgraduate major category according to the user's original major; S112. Calculate the priority index of all enrollment majors under the postgraduate major category; Priority index Pri=ω1*Pm+ω2*n dir +ω3*n sub , where ω1, ω2, and ω3 are preset weight values, Pm is a preset constant parameter, ndir is the total number of research directions of the current major in all colleges and universities to correspond to the popularity of the major, and n sub is the total number of examination subjects for the current major in all colleges and universities. The more examination subjects there are, the greater the probability that the major contains examination subjects that the user has studied or is good at. S113. Recommend several majors with the highest priority indexes to the user. Specifically, in this embodiment, recommend two majors with the highest priority indexes to the user.

[0011] In this embodiment, the major recommendation module also includes an admission rate-oriented recommendation module, which calculates the admission rate of each major in previous years and recommends several majors with the highest admission rates to the user.

[0012] S2. Obtain a candidate set of college majors, which includes several college major samples, including the following steps S21-S22: S21. Add basic college major samples to the college major candidate set, where basic college major samples are college major samples that fully meet the user's personalized information and intended majors. The user's personalized information includes the target region (i.e., the region where the target college is located, such as the province, municipality, or autonomous region), learning form (such as whether to take an in-service postgraduate exam), and test subject exclusion characteristics (such as whether to take a mathematics exam). S22, adding the school major samples that meet the user's personalized information and have a high similarity with the user's intended major to the school major candidate set, the specific process includes steps S221-S223: S221. Obtain the direction subject set corresponding to each professional sample of each institution. The direction subject set includes the research direction elements and examination subject elements corresponding to the professional sample of the institution. For example, the major of Marxist philosophy of Peking University, the corresponding research direction and examination subject set include {History of Marxist Philosophy, Contemporary Chinese Marxist Philosophy, Western Marxism, Principles of Marxist Philosophy, 101 Ideological and Political Theory, 201 English I, ...} S222. Calculate the similarity coefficient J(A,B) = |A∩B| / |A∪B| between each basic college professional sample and other college professional samples, where A∩B is the number of elements in the intersection between the direction subject set of college major A and the direction subject set of college major B, and A∪B is the number of elements in the union between the direction subject set of college major A and the direction subject set of college major B; for example, the direction subject set of college major A is {direction 1, direction 2, subject 1, subject 2}; the direction subject set of college major B is {direction 2, direction 3, subject 2, subject 3}, then the similarity coefficient between college major A and college major B is 2 / 6.

[0013] S223. Add a preset number of samples of college majors with the highest similarity coefficients among samples of other college majors to the candidate set of college majors; for example, when the user selects Xinjiang region and finance major, Xinjiang University-Theoretical Economics and Shihezi University-Political Economics, which have the highest similarity to Xinjiang University-Finance, can be added to the candidate set.

[0014] S3, calculating the characteristic scores corresponding to the university and major samples, and sorting the university and major samples in the university and major candidate set according to the characteristic scores; The process of calculating the characteristic scores corresponding to the sample of colleges and majors is as follows: preset weight coefficients for several characteristic elements of the sample of colleges and majors and perform differentiated scoring. The characteristic scores corresponding to the sample of colleges and majors = the sum of the differentiated scores of each characteristic element multiplied by the corresponding weight coefficients.

[0015] In this embodiment, in step S3, the characteristic elements of the professional samples of the institutions include at least one of the following: the grade of the institution, the ranking of the major, and the doctoral degree. In this embodiment, the weight values ​​corresponding to the above characteristic elements are 8, 8, and 2, respectively. The grades of institutions are 985, 211, ordinary double first-class, and ordinary undergraduate, and the corresponding scores are 30, 20, 15, and 0, respectively. The professional ranking is to sort the professional rankings of the professional samples of each institution from high to low and assign scores. The score assigned is 5 if there is a doctoral degree, and 0 if there is no doctoral degree. For example, if a professional sample of a certain institution is a 211 institution with a professional ranking score of 15 and has a doctoral degree, then its corresponding characteristic score = 8*20+8*15+2*5.

[0016] In this embodiment, the characteristic factors of the college and major sample also include the number of students enrolled. Specifically, the weight value of the number of students enrolled is 4, and the corresponding score is calculated through the following steps S31-S33; S31, count the number of students enrolled in each school major sample in the school major candidate set, and obtain the school major enrollment number set {n1, n2, n3, n4...,ni}, where ni is the number of students enrolled in the i-th school major sample. S32, calculate the mean value μ and standard deviation σ of {n1, n2, n3, n4...,ni}. If the standard deviation is greater than the preset value, reduce the excessive value in the set of enrollment numbers of college majors, recalculate the mean value and standard deviation to control the standard deviation within the preset range; S33, substitute {n1, n2, n3, n4...,ni} in step S31 into the formula Calculate the score assigned to each college’s professional sample corresponding to the number of students enrolled.

[0017] In this embodiment, the process of reducing the excessively large values ​​in the set of enrollment numbers of colleges and universities in step S32 is: replacing the values ​​in the set of enrollment numbers of colleges and universities that are greater than aμ with aμ, where a>1.5. In this embodiment, a=2.

[0018] In this embodiment, the ranking of the characteristic scores of the preset number of college major samples recommended in step S4 in the college major candidate set is an arithmetic progression, the tolerance is the rounded value of i*b% and the tolerance is not less than 1, i is the total number of college major samples in the college major candidate set, and b is the interval coefficient. Therefore, the recommended college majors can be further dispersed to prevent the college majors recommended to the user from having a high degree of overlap. If p=1, i=100, and b is 2, then the 6 college majors ranked 1, 3, 5, 7, 9, and 11 are recommended to the user.

[0019] S4. According to the ranking result of step S3, a preset number of school and major samples are taken from i(1-p) before the ranking of high scores to low scores and recommended to the user. In this embodiment, 6 school and major samples are taken and recommended to the user, where i is the total number of school and major samples in the school and major candidate set, p is the user's ability coefficient, , Score u Score is the current user ability evaluation score. min Score is the lowest score of user ability evaluation within the statistical range. max The highest score of the user's ability assessment within the statistical range, specifically, the statistical range is the users who use the software corresponding to the postgraduate entrance examination school recommendation method in this embodiment.

[0020] The calculation process of the user ability evaluation score in step S4 is to preset weight coefficients for several ability factors of the user and grade and assign scores. The user ability evaluation score = the sum of the scores of each ability factor multiplied by the corresponding weight coefficient.

[0021] In this embodiment, the ability factor includes at least one of the test preparation stage, school level, grade ranking, and English proficiency; among which: The classification of the preparation stage includes: not graduated, graduated but unemployed, and graduated and employed, which reflects the learning environment of the candidates. In this embodiment, the weight coefficient corresponding to the preparation stage is the highest, which is 10. In this embodiment, those who have not graduated are further divided into freshman, sophomore, junior, and senior. Combined with graduation but unemployed and graduation and employed, the points assigned to each classification are 12, 13, 14, 15, 10, and 8 respectively.

[0022] The weight coefficient of the institution level is 8. The levels include: 985, 211, ordinary double first-class (i.e. double first-class institutions that are not 985 or 211), ordinary undergraduate (i.e. undergraduate institutions that are not 985, 211 or double first-class), junior college, and adult education. The corresponding scores are 25, 15, 10, 8, 7, and 6, respectively.

[0023] The weight coefficient of the grade ranking is 8, and the classification includes the top 10%, 20%, 30% of the major... The corresponding score values ​​are 10, 9, 8...; The weight coefficient of English proficiency is 6, and the classification includes CET-4, CET-6, TEM-4, TEM-8, IELTS score of 6.0 or above, and TOEFL score of 80 or above. The corresponding scores are 5, 10, 10, 15, 10, and 10. For example, if the user is a senior student of an ordinary double first-class university, with a professional score in the top 20%, and English level 4, then the user's ability assessment score = 10*15+8*8+8*9+6*5.

[0024] Based on the above method, a postgraduate school recommendation method in this embodiment is as follows: adding school major samples that meet the user's personalized information and have a high similarity with the user's intended major to the school major candidate set to expand the school major samples in the school major candidate set, so as to avoid the situation where the number of recommended school majors is too small and the user registration is concentrated. And it can match the appropriate school major to the user according to the user's ability coefficient and the characteristic score corresponding to the school major sample, so as to avoid the user overestimating or underestimating his own ability when choosing a school major, and also avoid the user group from concentrating on a certain school major, resulting in uneven registration.

[0025] In other embodiments, the major recommendation module also includes an employment-oriented recommendation module, which recommends postgraduate majors based on the user's employment needs.

[0026] The limitation of traditional employment-oriented recommendation of graduate majors is that it relies on manual data collection, cleaning and integration, which is time-consuming, labor-intensive, inefficient and difficult to cope with large-scale data processing needs. It is difficult to handle huge amounts of data when helping candidates choose schools and majors.

[0027] The career-oriented recommendation modules include: Job information crawling module, used to obtain job information from the website; A data processing module, used to convert unstructured text data in job information into structured data that can be stored and retrieved in a database; Position and major mapping module, used to establish the relationship between positions and majors; Update and dynamic analysis module, set up timing scripts, regularly process newly released job information through job information capture module and data processing module, and add it to the database. And update the relationship between job and major through job and major mapping module, and provide decision support for candidates when choosing majors. It can process huge job data information, establish the connection between job data and graduate application majors, and recommend suitable graduate majors according to users' personalized needs for careers.

[0028] in: The principle of the job information crawling module is to use web crawlers technology to automatically access various target websites and recruitment announcement platforms by writing crawler scripts to obtain job information. The implementation steps include steps Sa1-Sa3 Sa1, URL list construction: First, collect the URLs of job posting pages of major websites through manual or automated Sitemap parsing (code parses the sitemap.xml file of the target website), and store these URL information in the MySQL database (containing four fields: the base64-encoded URL as the unique index, the URL string, the crawling status of this data, and the number of retries when crawling this data fails).

[0029] Sa2, running the crawler program at a preset frequency and time: including the following steps Sa21-Sa23 Sa21, read the URL data table stored in the database in the previous step, and read the data in the order of uncrawled state > less crawling failure retry times > more crawling failure retry times; Sa22, obtain web page information according to URL information, if successful, proceed to the next step, and change the crawling status of this data in the database to success; Sa23. If the acquisition fails, retry. The retry time calculation process is as follows: Set the initial retry time T 0 , Maximum retry time T max , cumulative failure threshold N max 、The number of failures in this round is M max , obtain the cumulative number of failures N and the number of failures in this round M from the database, according to the formula , where α and β are weight coefficients used to adjust the impact of the cumulative number of failures and the number of failures in this round on the retry time. max 、M max According to the empirical setting, the weight coefficients α and β are calculated using the hierarchical analysis method, i.e., the AHP method. If the current round of retry is successful, the crawling status of the URL data is changed to success. Otherwise, the total number of crawling failures is updated and the next round of tasks is waited for. And the cumulative number of failures greater than N is discarded. max data.

[0030] Sa3, web page parsing and data extraction, analyze the page structure of the job posting website, use Python's BeautifulSoup library combined with regular expressions to extract key information of the position including: position information, salary, employer information, position requirement information, and contact information.

[0031] The data processing module includes the following steps Sb1-Sb3 Sb1. Data preprocessing: The crawled data has the following problems: the crawled salary is a range, and the units are not uniform. The salary is processed into three columns: the highest salary, the lowest salary, and the average salary; there is a large amount of blank data and duplicate data, which need to be cleaned up in batches; the company name column contains geographic location information, which needs to be extracted separately. These problems need to be solved through data preprocessing. Including the following steps Sb11-Sb14 Sb11. Clean up duplicate data: The first step is to calculate the hash value of the entire data. The same hash value means duplication; the second step is to select key fields, such as company name, position title, educational requirements, and work location. When the key field values ​​are the same, the two data are considered duplicate data; the third step is to use string matching algorithms (such as Jaccard similarity and Levenshtein distance) to detect the duplication of data from different sources, set a similarity threshold, and delete duplicate data.

[0032] Sb12. Salary processing: First, unify the salary unit into RMB / month. If the salary is a range, use regular expressions to extract the left boundary value and assign it to the minimum salary, the right boundary value to the maximum salary, and calculate the average salary.

[0033] Sb13. Standardization of job titles: remove differences in upper and lower case (such as converting all job titles to lower case); remove special characters and extra spaces (such as punctuation, line breaks, extra spaces, etc.); manually build a standard job title dictionary (list all common job titles and manually create mapping rules); automatically build a mapping (by counting the frequency of job titles, selecting the most common job titles as the standard, and then mapping job titles with high similarity to these standard jobs).

[0034] Sb14. Localization: Many job descriptions posted by foreign companies on recruitment websites are in English. They can be translated into Chinese using the translation website interface.

[0035] Sb2, data structuring: converting unstructured text data into structured data that can be stored and retrieved in the database (such as relational database tables), ensuring that the data is stored in a unified format and easy to query. It includes the following steps Sb21-Sb23: Sb21. Design a data model and define fields such as job title, location, educational requirements, and job responsibilities based on the structure of job information.

[0036] Sb22. Field mapping: Use NLP techniques or rules to map the extracted text data to fields in the data model.

[0037] Sb23, data format conversion: convert the extracted information into a format according to the data model and generate Figure 1The structured tabular data shown.

[0038] Sb3, data storage: Use relational database or NoSQL database to store structured data to meet the needs of data retrieval and analysis. In this embodiment, the selected database is Alibaba Cloud PolarDB, and indexes are established for commonly used query fields (such as job title, location, release date, and professional requirements).

[0039] The position and profession mapping module includes the following steps Sc1-Sc4: Sc1, word segmentation, stop word removal: mainly use the precise mode of the jieba library in Python to perform Chinese word segmentation.

[0040] In the process of Chinese word segmentation of the target text information, every time a new entry is read, it will be determined whether it is the same as the content of the preset stop word list. If so, the entry will be deleted, and then the processed entry group will be retained until all the text information words are judged.

[0041] Sc2, weight value calculation: After the processing of step Sc1, the text information is converted into multiple entries, and each entry needs to be weighted. The TFTDF algorithm is used to count the word frequency and calculate the importance of a certain word to the text information.

[0042] For example, the vector space model maps text information into the following feature vector ν(d i ) = (t 1 , w i1 ; t 2 ,w i2 ...t k , w ik ), where t k (k=1,2,3...n) is a set of different entries, w ik t k In d i The weight value in The TFIDF weight calculation formula is as follows: where w ik The feature word t k In text d i The weight value in tf ik (t k ) indicates that the feature word tk is in the text d i The frequency of occurrence in idf(t k ) represents the characteristic word t k The inverse document frequency, n k Indicates that it contains the feature word tk The number of texts.

[0043] Sc3. Text similarity calculation In the vector space model, two texts are represented by cosine values. i and d j The calculation formula for content correlation is: , The vector space model can transform unstructured text content into a structured form that is easy to perform data operations in the form of simple vectors.

[0044] Sc4. Mutual recommendation. Mutual recommendation means that in the matching process, an optimal matching solution should be given under the premise of meeting the requirements of both users to the greatest extent, so as to optimize the matching effect as much as possible. The mutual similarity Rsim(u, v) refers to the probability of successful matching between the two parties, that is, the probability of successful recommendation of professional u and position v by the other party at the same time. sim(u, v) is the similarity between u's preference and v's feature, while sim(v, u) is the similarity between u's feature and v's preference. The calculation formula is as follows: ; Calculate the reciprocity index between postgraduate entrance examination majors and positions, obtain the correlation between majors and positions through manually set thresholds, and store them in the database.

[0045] Example 2 A postgraduate entrance examination institution recommendation system, used to implement the postgraduate entrance examination institution recommendation method described in Example 1, characterized in that it includes: A user major intention acquisition unit, used to acquire the user's intended major; A college and major candidate set generation unit is used to screen college and major samples to generate a college and major candidate set; The college and major sample ranking unit is used to calculate the characteristic scores corresponding to the college and major samples in the college and major candidate set, and to sort the college and major samples in the college and major candidate set according to the characteristic scores; A user capability coefficient calculation unit is used to obtain the user's capability elements to calculate the user capability evaluation score, count the user capability evaluation scores in the system, and calculate the user capability coefficient of the current user; The school recommendation generation unit is used to select a number of school and major samples according to the user's ability coefficient and the ranking of the school and major samples in the school and major candidate set to generate corresponding recommendation information to inform the user.

[0046] The technical principles of the present invention are described above in conjunction with specific embodiments. These descriptions are only for explaining the principles of the present invention and cannot be interpreted as limiting the scope of protection of the present invention in any way. Based on the explanations herein, those skilled in the art can associate other specific implementations of the present invention without creative work, and these methods will fall within the scope of protection of the present invention.

Claims

1. A method for recommending colleges and universities for postgraduate entrance examinations, characterized in that: The method comprises the following steps S1 to S4: S1. Obtain the user's intended major; S2. Obtain a candidate set of college majors, which includes several college major samples, including the following steps S21-S22: S21. Add basic college major samples to the college major candidate set, wherein the basic college major samples are college major samples that fully meet the user's personalized information and intended majors, and the user's personalized information includes target regions, learning forms, and test subject exclusion characteristics; S22, adding the school major samples that meet the user's personalized information and have a high similarity with the user's intended major to the school major candidate set, the specific process includes steps S221-S223: S221, obtaining a direction subject set corresponding to each college professional sample, wherein the direction subject set includes a research direction element and an examination subject element corresponding to the college professional sample; S222, calculate the similarity coefficient J(A,B)=|A∩B| / |A∪B| between each basic college professional sample and other college professional samples, where A∩B is the number of elements in the intersection between the direction subject set of college major A and the direction subject set of college major B, and A∪B is the number of elements in the union between the direction subject set of college major A and the direction subject set of college major B; S223, adding a preset number of university major samples with the highest similarity coefficients among other university major samples to the university major candidate set; S3, calculating the characteristic scores corresponding to the university and major samples, and sorting the university and major samples in the university and major candidate set according to the characteristic scores; The process of calculating the characteristic scores corresponding to the university professional samples is as follows: preset weight coefficients for several characteristic elements of the university professional samples and perform differentiated scoring. The characteristic scores corresponding to the university professional samples = the sum of the scores of each characteristic element multiplied by the corresponding weight coefficients; S4. According to the ranking result of step S3, a preset number of school and major samples are taken from i(1-p) before the ranking from high score to low score and recommended to the user, where i is the total number of school and major samples in the school and major candidate set, p is the user's ability coefficient, , Score u Score is the current user ability evaluation score. min Score is the lowest score of user ability evaluation within the statistical range. max The highest user capability assessment score within the statistical scope.

2. A method for recommending postgraduate colleges and universities according to claim 1, characterized in that: If the user does not clearly specify the intended major to apply for in step S1, a major recommendation module is also included to recommend a major to the user according to the user's needs.

3. A method for recommending postgraduate colleges and universities according to claim 2, characterized in that: The professional recommendation module includes an original professional guidance recommendation module, which is used to make professional recommendations based on the user's original major, including the following steps: S111. Obtain the corresponding postgraduate major category according to the user's original major; S112. Calculate the priority index of all enrollment majors under the graduate major category; Priority index Pri=ω1*Pm+ω2*n dir +ω3*n sub , where ω1, ω2, ω3 are preset weight values, Pm is a preset constant parameter, n dir is the total number of research directions of the current major in all institutions, n sub The total number of examination subjects for the current major in all colleges and universities; S113. Recommend several majors with the highest priority indexes to the user.

4. A method for recommending postgraduate colleges and universities according to claim 2, characterized in that: The major recommendation module includes an admission rate-oriented recommendation module, which calculates the admission rates of various majors in previous years and recommends several majors with the highest admission rates to users.

5. The method for recommending postgraduate colleges and universities according to claim 1, characterized in that: The calculation process of the user ability assessment score is to preset weight coefficients for several ability factors of the user and grade and assign scores. The user ability assessment score = the sum of the scores of each ability factor multiplied by the corresponding weight coefficient; the ability factors include at least one of the preparation stage, school level, grade ranking, and English level.

6. A method for recommending postgraduate colleges and universities according to claim 1, characterized in that: In step S3, the characteristic elements of the school and major sample include at least one of the school grade, major ranking, and doctoral program status.

7. A method for recommending postgraduate colleges and universities according to claim 7, characterized in that: The characteristic elements of the school major sample also include the number of students enrolled, and the corresponding scoring values ​​are calculated through the following steps S31-S33; S31. Count the number of students enrolled in each school major sample in the school major candidate set, and obtain the school major enrollment number set {n1, n2, n3, n4...,ni}, where ni is the number of students enrolled in the i-th school major sample. S32, calculating the average value μ and standard deviation σ of {n1, n2, n3, n4...,ni}. If the standard deviation is greater than a preset value, reducing the excessive value in the set of enrollment numbers of the college major, recalculating the average value and standard deviation to control the standard deviation within a preset range; S33, substitute {n1, n2, n3, n4...,ni} in step S31 into the formula Calculate the score assigned to each college’s professional sample corresponding to the number of students enrolled.

8. A method for recommending postgraduate colleges and universities according to claim 8, characterized in that: The process of reducing the excessively large values ​​in the set of enrollment numbers of the colleges and universities in step S32 is as follows: replacing the values ​​in the set of enrollment numbers of the colleges and universities that are greater than aμ with aμ, where a>1.

5.

9. A method for recommending postgraduate colleges and universities according to claim 1, characterized in that: The ranking of the characteristic scores of the preset number of college and major samples recommended in step S4 in the college and major candidate set is an arithmetic progression, the tolerance is the rounded value of i*b% and the tolerance is not less than 1, i is the total number of college and major samples in the college and major candidate set, and b is the interval coefficient.

10. A postgraduate entrance examination institution recommendation system, used to implement the postgraduate entrance examination institution recommendation method according to claim 1, characterized in that: include: A user major intention acquisition unit, used to acquire the user's intended major; A college and major candidate set generation unit is used to screen college and major samples to generate a college and major candidate set; The college and major sample ranking unit is used to calculate the characteristic scores corresponding to the college and major samples in the college and major candidate set, and to sort the college and major samples in the college and major candidate set according to the characteristic scores; A user capability coefficient calculation unit is used to obtain the user's capability elements to calculate the user capability evaluation score, count the user capability evaluation scores in the system, and calculate the user capability coefficient of the current user; The school recommendation generation unit is used to select a number of school and major samples according to the user's ability coefficient and the ranking of the school and major samples in the school and major candidate set to generate corresponding recommendation information to inform the user.

Images