CN112818254A - Farmer and civil industry personalized employment recommendation method and system based on intelligent logic collaborative filtering - Google Patents

Farmer and civil industry personalized employment recommendation method and system based on intelligent logic collaborative filtering Download PDF

Info

Publication number
CN112818254A
CN112818254A CN202110148315.0A CN202110148315A CN112818254A CN 112818254 A CN112818254 A CN 112818254A CN 202110148315 A CN202110148315 A CN 202110148315A CN 112818254 A CN112818254 A CN 112818254A
Authority
CN
China
Prior art keywords
user
membership
information
employment
users
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110148315.0A
Other languages
Chinese (zh)
Other versions
CN112818254B (en
Inventor
李鹏
韩伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Niuniu Intelligent Technology Co ltd
Original Assignee
Guangdong Niuniu Intelligent Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Niuniu Intelligent Technology Co ltd filed Critical Guangdong Niuniu Intelligent Technology Co ltd
Priority to CN202110148315.0A priority Critical patent/CN112818254B/en
Publication of CN112818254A publication Critical patent/CN112818254A/en
Application granted granted Critical
Publication of CN112818254B publication Critical patent/CN112818254B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9536—Search customisation based on social or collaborative filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G06Q10/105—Human resources

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

A rural industry personalized employment recommendation method and system based on intelligent logic collaborative filtering belongs to the technical field of artificial intelligence and is used for solving the problem that the existing employment recommendation method cannot accurately and reasonably recommend the employment requirements of rural industry users. According to the method, historical information of background similar groups is screened out by uploading a labor tool or a working environment picture and information of the country of the year, the intelligent logic is introduced into a collaborative filtering algorithm, and employment preference of rural workers is reflected more truly by a scoring means of conversion of three membership degrees of true membership degree, uncertain membership degree and unreal membership degree, so that recommendation is more accurate, and the problems that rigid evaluation is adopted in a scoring process in the prior art and incomplete, complete or inconsistent evaluation cannot be well measured due to the fact that subjective uncertainty of people is not considered are solved. The invention can be applied to the farmer to quickly obtain employment recommendations, thereby improving the efficiency of the farmer to find work.

Description

Farmer and civil industry personalized employment recommendation method and system based on intelligent logic collaborative filtering
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a farmer and civil worker personalized employment recommendation method and system based on intelligent logic collaborative filtering.
Technical Field
China is a world big labor country, and the employment recommendation system has a very important meaning for solving the employment of the labor in China, however, the current employment recommendation system mainly aims at most of the users of college students or groups suffering from high education and the like, and rarely aims at the employment system of the agricultural workers. The conventional employment recommendation system needs to fill in a large amount of information when users register and use, is particularly unfriendly to 25 percent of rural workers over 50 years old and 72 percent of rural workers in junior high school and below cultural degree, so that a plurality of rural workers cannot operate well to omit important information and cannot recommend proper employment opportunities. Therefore, the prior art does not solve the following problems: 1) the agricultural and civil worker population is large in the number of the elderly population, low in education degree, more centralized in employment distribution industry and embodies the employment characteristic of 'same country centralized', and the existing recommendation system product is complex to operate and has no targeted screening process; 2) in a collaborative filtering process of the conventional recommendation system, rigid evaluation is adopted in a scoring process, subjective uncertainty of a person is not considered, and incomplete, inconsistent evaluations cannot be well measured, so that the accuracy of a recommendation result is influenced.
Disclosure of Invention
In view of the above problems, the invention provides a rural power industry personalized employment recommendation method and system based on the intelligent logic collaborative filtering, which are used for solving the problem that the existing employment recommendation method and system cannot accurately and reasonably recommend the employment requirements of rural power industry users.
According to one aspect of the invention, the invention provides a farmer personalized employment recommendation method based on the intelligent logic collaborative filtering, which comprises the following steps:
the method comprises the steps of firstly, obtaining pictures and user information uploaded by a user; wherein, the picture comprises a working tool or working environment for the work of the agricultural workers; the user information comprises the information of the place where the user is located and the historical scoring information of each working post by the user;
step two, judging the type of the belonging work category according to the picture, and preliminarily screening out an enterprise object group corresponding to the type of the belonging work category;
thirdly, secondary screening is carried out in the enterprise object group according to the information of the user home location, and the enterprise object group associated with the information of the user home location is screened out;
step four, according to a preset intelligent logic scoring conversion formula with uncertainty processing capability and historical scoring information of the user on each working post, converting the historical scoring information in the enterprise object group obtained after secondary screening into rural worker user and post scoring information based on the intelligent number;
calculating according to the rural and civil worker users based on the Chinese wisdom number and the post scoring information to obtain the similarity between the users and other users, arranging the similarity calculation scores in a descending order, and obtaining a plurality of other users with the similarity calculation scores arranged in the front as nearest neighbor users;
and step six, calculating and acquiring a predicted user employment recommendation list according to the rural worker user and post scoring information based on the noon data acquired in the step four and the nearest neighbor user information acquired in the step five.
Further, the category of the industry in the second step includes six categories of manufacturing industry, construction industry, wholesale and retail industry, transportation and postal storage industry, lodging and catering industry, residential service repair and other service industry.
Further, in the second step, the class of the work is judged by an image classification method based on deep learning.
Further, the enterprise object group associated with the information of the place where the user belongs in the third step includes an enterprise object group to which employment personnel in the same place or a place near the place where the user belongs.
Further, the conversion formula of the intelligent logic scoring in the fourth step is as follows:
Figure BDA0002931563400000021
wherein the content of the first and second substances,
Figure BDA0002931563400000022
c is 1,2, …, n is grade number;
Figure BDA0002931563400000023
representing true value membership;
Figure BDA0002931563400000024
representing an uncertain membership;
Figure BDA0002931563400000025
representing a non-true degree of membership; delta represents the step size between the levels,
Figure BDA0002931563400000026
m-0.5 represents the median value; n is the number of levels.
Further, the historical scoring information in the fourth step includes the expected conformity scoring information of the signing post and the interested post.
Further, in the fifth step, the similarity between the user and other users is obtained by calculating the weighted similarity of the reinforced true membership proportion, wherein the similarity calculation formula is as follows:
Figure BDA0002931563400000027
wherein, wjRepresents the attribute weight, p represents the number of attributes, and
Figure BDA0002931563400000031
Figure BDA0002931563400000032
and
Figure BDA0002931563400000033
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u;
Figure BDA0002931563400000034
and
Figure BDA0002931563400000035
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u'; mu, lambda and gamma are weights of true value membership, uncertain membership and unreal membership respectively, and satisfy that mu + lambda + gamma is 1.
Further, in the sixth step, the obtained predicted user employment recommendation list is calculated according to the following score prediction value calculation formula:
Figure BDA0002931563400000036
wherein the content of the first and second substances,
Figure BDA0002931563400000037
representing the predicted score of the user u for the position i;
Figure BDA0002931563400000038
the average value of the scores of the user u for the interesting positions is represented; neigh represents the nearest neighbor number of the user u; sim (u, u ') represents the similarity of users u and u'; r isu′,iRepresenting the rating of the user u' for the position i;
Figure BDA0002931563400000039
represents the average value of the scores of the user u' for the interesting positions.
According to another aspect of the invention, a farmer personalized employment recommendation system based on the intelligent logic collaborative filtering is provided, which comprises:
the information acquisition module is used for acquiring the picture uploaded by the user and the user information; wherein, the picture comprises a working tool or working environment for the work of the agricultural workers; the user information comprises the information of the place where the user is located and the historical scoring information of each working post by the user;
the primary screening module is used for judging the type of the work and the seed according to the picture and preliminarily screening out an enterprise object group corresponding to the type of the work and the seed; wherein the work category comprises six categories of manufacturing industry, construction industry, wholesale and retail industry, transportation and postal storage industry, lodging and catering industry, and residential service repair and other service industry; judging the class of the work by an image classification method based on deep learning;
the secondary screening module is used for carrying out secondary screening in the enterprise object group according to the information of the user home location and screening out the enterprise object group associated with the information of the user home location; the enterprise object group associated with the information of the place where the user hometown belongs comprises an enterprise object group to which employment personnel in the same hometown or a region near the hometown of the user belong;
the intelligent conversion module is used for converting historical scoring information in the enterprise object group obtained after secondary screening into rural workers user and post scoring information based on the intelligent number according to a preset intelligent logic scoring conversion formula with uncertain processing capacity and the historical scoring information of the user on each work post; the historical scoring information comprises signing post and expected conformity scoring information of interested posts and the signing post;
the similarity calculation module is used for calculating and obtaining the similarity between the user and other users according to the rural labor user and the post scoring information based on the Chinese wisdom number, arranging the similarity calculation scores in a descending order and obtaining a plurality of other users with the similarity calculation scores arranged in the front as nearest neighbor users; the similarity between the user and other users is obtained by calculating the weighted similarity of the membership proportion of the reinforced truth value;
the prediction recommendation module is used for calculating and acquiring a predicted user employment recommendation list according to the acquired rural and civil users and post scoring information based on the Chinese wisdom number and the nearest neighbor user information; the user employment recommendation list for obtaining the prediction is calculated according to the following scoring prediction value calculation formula:
Figure BDA0002931563400000041
wherein the content of the first and second substances,
Figure BDA0002931563400000042
representing the predicted score of the user u for the position i;
Figure BDA0002931563400000043
the average value of the scores of the user u for the interesting positions is represented; neigh represents the nearest neighbor number of the user u; sim (u, u ') represents the similarity of users u and u'; r isu′,iRepresenting the rating of the user u' for the position i;
Figure BDA0002931563400000044
represents the average value of the scores of the user u' for the interesting positions.
Further, the conversion formula of the zhongzhi logic score in the zhongzhi conversion module is as follows:
Figure BDA0002931563400000045
wherein the content of the first and second substances,
Figure BDA0002931563400000046
c is 1,2, …, n is grade number;
Figure BDA0002931563400000047
representing true value membership;
Figure BDA0002931563400000048
representing an uncertain membership;
Figure BDA0002931563400000049
representing a non-true degree of membership; delta represents the step size between the levels,
Figure BDA00029315634000000410
m-0.5 represents the median value;n is the number of levels.
Further, the similarity calculation formula in the similarity calculation module is as follows:
Figure BDA00029315634000000411
wherein, wjRepresents the attribute weight, p represents the number of attributes, and
Figure BDA00029315634000000412
Figure BDA00029315634000000413
and
Figure BDA00029315634000000414
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u;
Figure BDA00029315634000000415
and
Figure BDA00029315634000000416
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u'; mu, lambda and gamma are weights of true value membership, uncertain membership and unreal membership respectively, and satisfy that mu + lambda + gamma is 1.
The beneficial technical effects of the invention are as follows:
according to the method, historical information of background similar groups is screened out by uploading labor tools or working environment pictures and information of the country of the year, the intelligent logic is introduced into a collaborative filtering algorithm, the employment preference of agricultural and civil users is reflected more truly by a scoring means of conversion of three membership degrees of true membership degree, uncertain membership degree and unreal membership degree, the traditional similarity formula is improved, the intelligent similarity in weighting of the strengthened true logic is used for measurement, the similar users are further found out on the basis of the background similar groups, recommendation is more accurate, and the problems that rigid evaluation is adopted in the scoring process in the prior art and incomplete, complete or inconsistent evaluation cannot be measured well due to the fact that subjective uncertainty of people is not considered are solved.
Drawings
The invention may be better understood by referring to the following description in conjunction with the accompanying drawings, in which like reference numerals are used throughout the figures to indicate like or similar parts. The accompanying drawings, which are incorporated in and form a part of this specification, illustrate preferred embodiments of the present invention and, together with the detailed description, serve to further explain the principles and advantages of the invention.
FIG. 1 shows a schematic flow chart of the individual employment recommendation method for farmers based on the intelligent logic collaborative filtering.
FIG. 2 is a schematic structural diagram of the individual employment recommendation system for the farmer based on the intelligent logic collaborative filtering.
Detailed Description
Exemplary embodiments of the present invention will be described hereinafter with reference to the accompanying drawings. In the interest of clarity and conciseness, not all features of an actual implementation are described in the specification. It should be noted that, in order to avoid obscuring the present invention with unnecessary details, only the device structures and/or processing steps closely related to the solution according to the present invention are shown in the drawings, and other details not so relevant to the present invention are omitted.
Detailed description of the preferred embodiment
With reference to fig. 1, the invention provides a rural folk worker personalized employment recommendation method based on intelligent logic collaborative filtering, which comprises the following specific steps:
the first step is to upload the photos of the labor tools used by the rural folk workers, carry out the classification and preliminary screening of the work types, and screen out the industry or enterprise object groups aiming at the labor skills of the users;
the method divides the work types of the agricultural and civil workers into 6 categories, and respectively corresponds the streamline image, the building site image, the stall image, the express package and express vehicle image, the kitchen ware and other images to the manufacturing industry, the building industry (including carpentry, water conservancy projects, electricians, tile workers and the like), the wholesale and retail industry, the transportation and postal storage industry, the lodging catering industry, the residential service repair and other service industries through the image classification method based on deep learning. And when the images uploaded by the rural worker users do not belong to the first five types, the images are classified into resident service repair and other service industries. The classified images are the most common daily images, and the farmer worker user can conveniently take pictures and select the pictures. For users who have not mastered labor skills, skipping this step;
uploading image categories Preliminary screening of work category
Pipeline image Manufacturing industry
Construction site image Construction industry (including carpenters, water works, electricians, masons, etc.)
Booth image Wholesale and retail trade
Express package, express vehicle image Transportation and storage postal industry
Kitchen tool Housing catering industry
Other images Residential service repair and other service industries
Secondly, performing secondary screening according to information of the hometown location in the registration information in the enterprise objects primarily screened in the first step to determine a sub-group with the most similar employment backgrounds of people in the same country and in nearby areas of the user;
setting a grading conversion table based on the intelligent logic;
in the classical set theory, the uncertainty of elements in a set cannot be evaluated and described, so the uncertainty in application can be processed by a fuzzy set, which proposes the concept of membership degree and indirectly expresses the uncertainty of elements by quantifying the membership degree of different sets. Whereas in the mesopic set, the uncertainty of an element is explicitly quantified, the element is divided into three components: the real component T, the uncertain component I and the false component F, and the relationship among the real component T, the uncertain component I and the false component F is not strongly restricted.
The intelligent set is a tool for describing the uncertain information most completely, can describe fuzzy information such as inaccuracy and incompleteness more accurately, and is a further continuation of the fuzzy set and the intuitive fuzzy set. However, the concept of the wisdom set is derived from philosophy, and its definition includes nonstandard subintervals, which cannot be conveniently applied to the actual scientific and engineering fields, so that the single-valued wisdom set arises from the operation, the single-valued wisdom set is represented by that its true value membership, uncertain membership and unreal membership are all defined in the standard unit subintervals, and the membership function is a specific real number, which can be more conveniently applied to the actual scientific and engineering fields.
Therefore, the grading conversion table based on the intelligent logic is set based on the theory, the rigid grading in the traditional recommendation method is avoided, the uncertainty of a farmer user in the grading process is better understood through the intelligent logic grading conversion with uncertainty processing capacity, and the grade number is non-even and is set as n for convenience of evaluation; the step size between each level is equal interval, and in order to ensure that the median value is about 0.5, the interval size δ can be set as:
Figure BDA0002931563400000071
then for single value median intelligent score
Figure BDA0002931563400000072
c is 1,2, …, n is grade number, and its true value membership degree
Figure BDA0002931563400000073
Uncertain degree of membership
Figure BDA0002931563400000074
And degree of unreal membership
Figure BDA0002931563400000075
Can be determined by the following conversion equation:
Figure BDA0002931563400000076
wherein the content of the first and second substances,
Figure BDA0002931563400000077
for a scoring table with 9 ranks, i.e., n-9, the conversion to single-valued noon-score is shown in the following table:
scale description (numerical score) Single value median intelligent score
Extremely good/full compliance (9) <1.00,0.00,0.00>
Very good/very good fit (8) <0.875,0.125,0.125>
Good/fit (7) <0.75,0.25,0.25>
Slightly better/slightly in line (6) <0.625,0.375,0.375>
Medium/average level (5) <0.50,0.50,0.50>
Slightly poor/not conform (4) <0.375,0.625,0.625>
Poor/not conform (3) <0.25,0.75,0.75>
Very poor/very bad (2) <0.125,0.875,0.875>
Extremely poor/completely incongruous (1) <0.00,1.00,1.00>
Fourthly, according to a grading conversion table, converting the historical grading information of the farmer users in the enterprise objects obtained by secondary screening into a farmer user and post grading matrix based on the Chinese wisdom number;
the agricultural and civil users can automatically estimate values and fill in intention scores (or not) of all posts according to the past employment situation and upload the intention scores to the system to obtain a historical score information table of the agricultural and civil users for all posts; the historical scoring information comprises signing post and expected conformity scoring information of interested posts and the historical scoring information; the more explanation is that the signing post is a post which has been made or has been signed and is doing, the interested post is a post which has not been made but is interested, the score information of the degree of conformity with the self expectation is that the satisfaction degree of the user on the welfare treatment, the working environment, the development space and the like of the signing post and the interested post is in accordance with the expected conformity degree when the user finds the work, the better the conformity degree is, the higher the score is made for the post, so that the grade score of 1-9 is formed, and the subsequent conversion into the single-value noose score is facilitated. For example:
Figure BDA0002931563400000078
Figure BDA0002931563400000081
after conversion, the following steps are changed:
Figure BDA0002931563400000082
wherein? And (c) indicates an unscored position, i.e., a position to be predicted which is recommended to a farmer user.
The fifth step is to calculate the intensified truth degree membership
Figure BDA0002931563400000083
The weighted similarity of the specific gravity is used as the final similarity;
the traditional similarity calculation is as follows:
(1) cosine similarity:
Figure BDA0002931563400000084
the cosine similarity is calculated by measuring the size of an included angle between a user vector u and a user vector u', and when the hollow value of a user-scoring matrix is large, namely the user does not score most posts, the cosine similarity method cannot accurately calculate the similarity of the user, so that the accuracy of a recommendation result is reduced, and the problem of fuzzy scoring of the user cannot be solved.
(2) Pearson correlation coefficient:
Figure BDA0002931563400000085
wherein I represents a set of positions, ru,iRepresents the user u's score, r, for position iu',iIndicating the rating of position i by user u',
Figure BDA0002931563400000086
representing the average rating of user u for all positions,
Figure BDA0002931563400000087
representing the average rating of user u' for all positions.
The Pearson correlation coefficient method depends heavily on user scores with a common score set when calculating user similarity, and when common score intersections exist between users in the system, errors of the Pearson correlation coefficient similarity calculation are large, and meanwhile the problem of user fuzzy score cannot be solved.
(3) Jaccard coefficient:
Figure BDA0002931563400000091
wherein, IuRepresenting a set of items scored by user u, Iu'Representing the set of items scored by user u'.
The Jaccard coefficient method ignores the influence of the difference of the score of the user on the recommendation effect when calculating the similarity of the user, and as the scale of the user is continuously increased, the result of the similarity calculation of the Jaccard coefficient is more fuzzy, and the problem of fuzzy score of the user cannot be solved.
Therefore, the invention adopts the weighted similarity of the proportion of the intensified truth value membership degree as the final similarity, and the final similarity sim (u, u ') of the two agricultural and civil users u and u' can be represented by the following formula:
Figure BDA0002931563400000092
wherein, wj...wpThe weight values of the p attributes are,
Figure BDA0002931563400000093
Figure BDA0002931563400000094
and
Figure BDA0002931563400000095
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u; mu, lambda and gamma are weight values of true value membership, uncertain membership and unreal membership respectively, and satisfy the condition that mu + lambda + gamma is 1, and in order to strengthen the proportion of true value membership, mu is 0.4, lambda is 0.3 and gamma is 0.3.
The method adopts the weighted similarity of the proportion of the reinforced true value membership as the final similarity, can more accurately express the employment preference of the agricultural and civil users, and can effectively improve the accuracy of the recommendation result and effectively solve the problem of fuzzy scoring of the users under the conditions that the users do not score most posts, the common scoring intersection between the users is less and the user scale is larger.
Sixthly, solving a user most similar to the current user as a neighbor user according to the similarity value;
and the seventh step is to generate a final recommendation list and recommend the most consistent information to the user.
The following scoring prediction value calculation formula can be adopted to predict the predicted scoring of the user on a certain unscored post, and TOP-N posts are generated according to the predicted scoring value and the descending order of the scores and recommended to the farmer worker user. The scoring prediction value calculation formula is as follows:
Figure BDA0002931563400000101
wherein the content of the first and second substances,
Figure BDA0002931563400000102
represents the prediction score of the farmer user u for position i,
Figure BDA0002931563400000103
representing the average value of the scores of the rural worker users for the interested posts, Neigh representing the number of the nearest neighbors of the rural worker users, namely the number of the most similar rural worker users calculated by the similarity, sim (u, u ') representing the similarity of two rural worker users u and u', ru′,iRepresents the rating of the farmer user u' for position i,
Figure BDA0002931563400000104
and represents the average value of the scores of the rural worker user u' on the interested posts. Therefore, TOP-N positions with the highest prediction scores in the positions which are not scored by the farmer worker user are obtained and recommended to the user.
Detailed description of the invention
With reference to fig. 2, the invention provides a farmer personalized employment recommendation system based on the intelligent logic collaborative filtering, which comprises:
the information acquisition module 10 is used for acquiring pictures uploaded by a user and user information; wherein, the picture comprises the working tools or working environment for the work of the peasant civil workers; the user information comprises the information of the place where the user is located and the historical scoring information of each working post by the user;
the primary screening module 20 is used for judging the type of the belonging work category according to the picture and preliminarily screening out an enterprise object group corresponding to the type of the belonging work category; wherein, the industry category comprises six categories of manufacturing industry, construction industry, wholesale and retail industry, transportation and postal storage industry, lodging and catering industry, resident service repair and other service industry; judging the class of the work by an image classification method based on deep learning;
the secondary screening module 30 is used for performing secondary screening in the enterprise object group according to the information of the user home location, and screening out the enterprise object group associated with the information of the user home location; the enterprise object group associated with the information of the place where the user hometown belongs comprises an enterprise object group to which employment personnel in the same hometown or a region near the hometown of the user belong;
the intelligent conversion module 40 is used for converting historical scoring information in the enterprise object group obtained after secondary screening into rural workers user and post scoring information based on the intelligent number according to a preset intelligent logic scoring conversion formula with uncertain processing capacity and historical scoring information of each work post by the user; the historical scoring information comprises signing post and expected conformity scoring information of interested posts and the historical scoring information;
the similarity calculation module 50 is used for calculating and obtaining the similarity between the user and other users according to the rural labor user and the post scoring information based on the Chinese wisdom number, arranging the similarity calculation scores in a descending order, and obtaining a plurality of other users with the similarity calculation scores arranged in the front as nearest neighbor users; the similarity between the user and other users is obtained by calculating the weighted similarity of the membership proportion of the reinforced truth value;
and the prediction recommendation module 60 is used for calculating and acquiring a predicted user employment recommendation list according to the acquired rural labor user and post scoring information based on the noon data and the nearest neighbor user information.
Further, the conversion formula of the zhongzhi logic score in the zhongzhi conversion module 40 is as follows:
Figure BDA0002931563400000111
wherein the content of the first and second substances,
Figure BDA0002931563400000112
c is 1,2, …, n is grade number;
Figure BDA0002931563400000113
representing true value membership;
Figure BDA0002931563400000114
representing an uncertain membership;
Figure BDA0002931563400000115
representing a non-true degree of membership; delta represents the step size between the levels,
Figure BDA0002931563400000116
m-0.5 represents the median value; n is the number of levels.
Further, the similarity calculation formula in the similarity calculation module 50 is as follows:
Figure BDA0002931563400000117
wherein, wjRepresents the attribute weight, p represents the number of attributes, and
Figure BDA0002931563400000118
Figure BDA0002931563400000119
and
Figure BDA00029315634000001110
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u;
Figure BDA00029315634000001111
and
Figure BDA00029315634000001112
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u'; mu, lambda and gamma are weights of true value membership, uncertain membership and unreal membership respectively, and satisfy that mu + lambda + gamma is 1.
Further, the predicted employment recommendation list of the user obtaining the prediction is calculated in the predicted recommendation module 60 according to the following score prediction value calculation formula:
Figure BDA00029315634000001113
wherein the content of the first and second substances,
Figure BDA00029315634000001114
representing the predicted score of the user u for the position i;
Figure BDA00029315634000001115
the average value of the scores of the user u for the interesting positions is represented; neigh represents the nearest neighbor number of the user u; sim (u, u ') represents the similarity of users u and u'; r isu′,iRepresenting the rating of the user u' for the position i;
Figure BDA00029315634000001116
represents the average value of the scores of the user u' for the interesting positions.
The functions of the individual employment recommendation system for the agricultural workers based on the intelligent logic collaborative filtering in this embodiment can be described by the individual employment recommendation method for the agricultural workers based on the intelligent logic collaborative filtering, so that the detailed parts in this embodiment can be referred to the above method embodiments, and are not described herein again.
The invention fully considers the actual situation of the user group of the rural workers and carries out closer to actual analysis on the employment demand and employment tendency problems of the rural workers. Aiming at the characteristics that the population is large and the average age is large over 50 years old of the rural civilian workers, the information acquisition process is simplified, the employment information collection barrier of the rural civilian workers is reduced, and the historical information of the population with similar background is screened out by uploading the working tool or working environment photos and the information of the same country; meanwhile, rigid evaluation is avoided in the scoring process, subjective uncertainty of people is not considered, incomplete inconsistent evaluation cannot be well measured, accuracy of a recommendation result is affected, intelligent logic is introduced into a collaborative filtering algorithm, three membership conversion scoring means of true value membership, uncertain membership and unreal membership are used for reflecting employment preference of agricultural workers more truly, a traditional similarity formula is improved, weighting intelligent similarity of reinforced true value logic is used for measurement, similar users are further found out on the basis of background similar groups, and recommendation accuracy is improved.
While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of this description, will appreciate that other embodiments can be devised which do not depart from the scope of the invention as described herein. Furthermore, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes, and may not have been selected to delineate or circumscribe the inventive subject matter. The present invention has been disclosed in an illustrative rather than a restrictive sense, and the scope of the present invention is defined by the appended claims.

Claims (10)

1. A rural folk worker personalized employment recommendation method based on intelligent logic collaborative filtering is characterized by comprising the following steps:
the method comprises the steps of firstly, obtaining pictures and user information uploaded by a user; wherein, the picture comprises a working tool or working environment for the work of the agricultural workers; the user information comprises the information of the place where the user is located and the historical scoring information of each working post by the user;
step two, judging the type of the belonging work category according to the picture, and preliminarily screening out an enterprise object group corresponding to the type of the belonging work category;
thirdly, secondary screening is carried out in the enterprise object group according to the information of the user home location, and the enterprise object group associated with the information of the user home location is screened out;
step four, according to a preset intelligent logic scoring conversion formula with uncertainty processing capability and historical scoring information of the user on each working post, converting the historical scoring information in the enterprise object group obtained after secondary screening into rural worker user and post scoring information based on the intelligent number;
calculating according to the rural and civil worker users based on the Chinese wisdom number and the post scoring information to obtain the similarity between the users and other users, arranging the similarity calculation scores in a descending order, and obtaining a plurality of other users with the similarity calculation scores arranged in the front as nearest neighbor users;
and step six, calculating and acquiring a predicted user employment recommendation list according to the rural worker user and post scoring information based on the noon data acquired in the step four and the nearest neighbor user information acquired in the step five.
2. The individual employment recommendation method for the agricultural workers based on the intelligent logic collaborative filtering is characterized in that the category of the industry in the step two comprises six categories of manufacturing industry, construction industry, wholesale and retail industry, transportation and postal warehousing industry, lodging and catering industry, residential service repair and other service industry.
3. The individual employment recommendation method for the agricultural workers based on the intelligent logic collaborative filtering as claimed in claim 2, wherein in the step two, the category of the work is judged by an image classification method based on deep learning.
4. The individual employment recommendation method for farmers based on intelligent logic collaborative filtering as claimed in claim 1, wherein the enterprise object group associated with the information of the location of the user's hometown in the third step comprises an enterprise object group to which the employment personnel belonging to the same hometown or the area nearby the hometown of the user belong.
5. The individual employment recommendation method for farmers based on intelligent logic collaborative filtering as claimed in claim 1, wherein the conversion formula of the intelligent logic score in step four is as follows:
Figure FDA0002931563390000011
wherein the content of the first and second substances,
Figure FDA0002931563390000021
c is 1,2, …, n is grade number;
Figure FDA0002931563390000022
representing true value membership;
Figure FDA0002931563390000023
representing an uncertain membership;
Figure FDA0002931563390000024
representing a non-true degree of membership; delta represents the step size between the levels,
Figure FDA0002931563390000025
m-0.5 represents the median value, and n represents the number of ranks.
6. The rural folk worker personalized employment recommendation method based on intelligent logic collaborative filtering as claimed in claim 5, wherein the historical scoring information in step four comprises the scoring information of the degree of agreement between the sign-up post and the interested post with the self expectation.
7. The individual farmer employment recommendation method based on the collaborative filtering of the intelligent logic as claimed in claim 1, wherein the similarity between the user and other users is obtained by calculating the weighted similarity of the reinforced truth value membership proportion, wherein the similarity calculation formula is as follows:
Figure FDA0002931563390000026
wherein, wjRepresents the attribute weight, p represents the number of attributes, and
Figure FDA0002931563390000027
Figure FDA0002931563390000028
and
Figure FDA0002931563390000029
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u;
Figure FDA00029315633900000210
and
Figure FDA00029315633900000211
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u'; mu, lambda and gamma are weights of true value membership, uncertain membership and unreal membership respectively, and satisfy that mu + lambda + gamma is 1.
8. The individual employment recommendation method for the agricultural workers based on the intelligent logic collaborative filtering as claimed in claim 1, wherein in the sixth step, the predicted employment recommendation list of the user is obtained by calculation according to the following calculation formula of the score prediction value, wherein the calculation formula of the score prediction value is as follows:
Figure FDA00029315633900000212
wherein the content of the first and second substances,
Figure FDA00029315633900000213
representing the predicted score of the user u for the position i;
Figure FDA00029315633900000214
the average value of the scores of the user u for the interesting positions is represented; neigh represents the nearest neighbor number of the user u; sim (u, u ') represents the similarity of users u and u'; r isu′,iRepresenting the rating of the user u' for the position i;
Figure FDA00029315633900000215
represents the average value of the scores of the user u' for the interesting positions.
9. The utility model provides a peasant civilian industry individualized employment recommendation system based on well-being logic collaborative filtering which characterized in that includes:
the information acquisition module is used for acquiring the picture uploaded by the user and the user information; wherein, the picture comprises a working tool or working environment for the work of the agricultural workers; the user information comprises the information of the place where the user is located and the historical scoring information of each working post by the user;
the primary screening module is used for judging the type of the work and the seed according to the picture and preliminarily screening out an enterprise object group corresponding to the type of the work and the seed; wherein the work category comprises six categories of manufacturing industry, construction industry, wholesale and retail industry, transportation and postal storage industry, lodging and catering industry, and residential service repair and other service industry; judging the class of the work by an image classification method based on deep learning;
the secondary screening module is used for carrying out secondary screening in the enterprise object group according to the information of the user home location and screening out the enterprise object group associated with the information of the user home location; the enterprise object group associated with the information of the place where the user hometown belongs comprises an enterprise object group to which employment personnel in the same hometown or a region near the hometown of the user belong;
the intelligent conversion module is used for converting historical scoring information in the enterprise object group obtained after secondary screening into rural workers user and post scoring information based on the intelligent number according to a preset intelligent logic scoring conversion formula with uncertain processing capacity and the historical scoring information of the user on each work post; the historical scoring information comprises signing post and expected conformity scoring information of interested posts and the signing post; the conversion formula of the intelligent logic scoring is as follows:
Figure FDA0002931563390000031
wherein the content of the first and second substances,
Figure FDA0002931563390000032
c is 1,2, …, n is grade number;
Figure FDA0002931563390000033
representing true value membership;
Figure FDA0002931563390000034
representing an uncertain membership;
Figure FDA0002931563390000035
representing a non-true degree of membership; delta represents the step size between the levels,
Figure FDA0002931563390000036
m-0.5 represents the median value; n is the number of grades;
the similarity calculation module is used for calculating and obtaining the similarity between the user and other users according to the rural labor user and the post scoring information based on the Chinese wisdom number, arranging the similarity calculation scores in a descending order and obtaining a plurality of other users with the similarity calculation scores arranged in the front as nearest neighbor users; the similarity between the user and other users is obtained by calculating the weighted similarity of the membership proportion of the reinforced truth value;
the prediction recommendation module is used for calculating and acquiring a predicted user employment recommendation list according to the acquired rural and civil users and post scoring information based on the Chinese wisdom number and the nearest neighbor user information; the user employment recommendation list for obtaining the prediction is calculated according to the following scoring prediction value calculation formula:
Figure FDA0002931563390000037
wherein the content of the first and second substances,
Figure FDA0002931563390000038
representing the predicted score of the user u for the position i;
Figure FDA0002931563390000039
the average value of the scores of the user u for the interesting positions is represented; neigh represents the nearest neighbor number of the user u; sim (u, u ') represents the similarity of users u and u'; r isu′,iRepresenting the rating of the user u' for the position i;
Figure FDA0002931563390000041
represents the average value of the scores of the user u' for the interesting positions.
10. The individual farmer employment recommendation system based on intelligent logic collaborative filtering as claimed in claim 9, wherein the similarity calculation formula in the similarity calculation module is as follows:
Figure FDA0002931563390000042
wherein, wjRepresents the attribute weight, p represents the number of attributes, and
Figure FDA0002931563390000043
Figure FDA0002931563390000044
and
Figure FDA0002931563390000045
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u;
Figure FDA0002931563390000046
and
Figure FDA0002931563390000047
respectively is true value membership, uncertain membership and unreal membership of the zhongzhi score of the user u'; mu, lambda and gamma are weights of true value membership, uncertain membership and unreal membership respectively, and satisfy that mu + lambda + gamma is 1.
CN202110148315.0A 2021-02-03 2021-02-03 Chinese-intelligence-logic-collaborative-filtering-based individual employment recommendation method and system for rural workers Active CN112818254B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110148315.0A CN112818254B (en) 2021-02-03 2021-02-03 Chinese-intelligence-logic-collaborative-filtering-based individual employment recommendation method and system for rural workers

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110148315.0A CN112818254B (en) 2021-02-03 2021-02-03 Chinese-intelligence-logic-collaborative-filtering-based individual employment recommendation method and system for rural workers

Publications (2)

Publication Number Publication Date
CN112818254A true CN112818254A (en) 2021-05-18
CN112818254B CN112818254B (en) 2022-06-14

Family

ID=75860757

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110148315.0A Active CN112818254B (en) 2021-02-03 2021-02-03 Chinese-intelligence-logic-collaborative-filtering-based individual employment recommendation method and system for rural workers

Country Status (1)

Country Link
CN (1) CN112818254B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115577185A (en) * 2022-11-15 2023-01-06 湖南师范大学 Muting course recommendation method and device based on mixed reasoning and mesopic group decision

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103530304A (en) * 2013-05-10 2014-01-22 Tcl集团股份有限公司 On-line recommendation method, system and mobile terminal based on self-adaption distributed computation
CN104899246A (en) * 2015-04-12 2015-09-09 西安电子科技大学 Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism
CN110427565A (en) * 2019-06-10 2019-11-08 深圳大学 A kind of item recommendation method based on collaborative filtering, intelligent terminal and storage medium
AU2020101885A4 (en) * 2020-08-19 2020-09-24 Xinjiang University A Novel Tensor Factorization Using Trust and Rating for Recommendation, system and method thereof

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103530304A (en) * 2013-05-10 2014-01-22 Tcl集团股份有限公司 On-line recommendation method, system and mobile terminal based on self-adaption distributed computation
CN104899246A (en) * 2015-04-12 2015-09-09 西安电子科技大学 Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism
CN110427565A (en) * 2019-06-10 2019-11-08 深圳大学 A kind of item recommendation method based on collaborative filtering, intelligent terminal and storage medium
AU2020101885A4 (en) * 2020-08-19 2020-09-24 Xinjiang University A Novel Tensor Factorization Using Trust and Rating for Recommendation, system and method thereof

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
SHAN YE,JING FU AND JUN YE: "Medical diagnosis using distance-based similarity measures of single valued neutrosophic multisets", 《NEUTROSOPHIC SETS AND SYSTEMS》 *
SHAN YE,JING FU AND JUN YE: "Medical diagnosis using distance-based similarity measures of single valued neutrosophic multisets", 《NEUTROSOPHIC SETS AND SYSTEMS》, 31 December 2015 (2015-12-31) *
毛雨欣: "基于用户行为的个性化推荐***的研究与实现", 《中国优秀硕士论文电子期刊网》 *
毛雨欣: "基于用户行为的个性化推荐***的研究与实现", 《中国优秀硕士论文电子期刊网》, 15 February 2019 (2019-02-15) *
钟前建等: "基于Mahout框架的协同过滤视频推荐算法", 《广东技术师范学院学报》, no. 08, 15 August 2016 (2016-08-15) *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115577185A (en) * 2022-11-15 2023-01-06 湖南师范大学 Muting course recommendation method and device based on mixed reasoning and mesopic group decision
CN115577185B (en) * 2022-11-15 2023-04-07 湖南师范大学 Muting course recommendation method and device based on mixed reasoning and mesopic group decision

Also Published As

Publication number Publication date
CN112818254B (en) 2022-06-14

Similar Documents

Publication Publication Date Title
EP3916668A1 (en) Urban land automatic identification system integrating industrial big data and building forms
Batty et al. Scaling and allometry in the building geometries of Greater London
Weinand et al. Developing a municipality typology for modelling decentralised energy systems
CN110096634B (en) House property data vector alignment method based on particle swarm optimization
CN111561734B (en) Heating household valve adjusting method, system and equipment based on artificial intelligence
CN107784394A (en) Consider that the highway route plan of prospect theory does not know more attribute method for optimizing
CN114331031B (en) Building traditional feature recognition and evaluation method and system
Haining Designing a health needs GIS with spatial analysis capability
CN102096633A (en) Application field oriented software quality standard evaluating method
CN112818254B (en) Chinese-intelligence-logic-collaborative-filtering-based individual employment recommendation method and system for rural workers
CN114398951A (en) Land use change driving factor mining method based on random forest and crowd-sourced geographic information
CN114529038A (en) Intelligent matching business recruitment strategy system and method based on enterprise demands
Woodcock Uncertainty in remote sensing
Nizeyimana et al. A study on the classfication of households in Rwanda based on factor scores
CN114707785A (en) Rural residential point multi-scale spatial feature analysis method based on deep learning
CN102609721B (en) Remote sensing image clustering method
CN113901348A (en) Oncomelania snail distribution influence factor identification and prediction method based on mathematical model
CN111401683B (en) Method and device for measuring tradition of ancient villages
Yang et al. A zoning method for the extreme wind pressure coefficients of buildings based on weighted K-means clustering
Zhang et al. Evaluating water resource carrying capacity in Pearl River-West River economic Belt based on portfolio weights and GRA-TOPSIS-CCDM
CN108133296B (en) Event attendance prediction method combining environmental data under social network based on events
Song et al. Developing urban building energy models for shanghai city with multi-source open data
Liu et al. Application of artificial neural network and SARIMA in portland cement supply chain to forecast demand
Gao et al. Higher education evaluation system based on factor Analysis
Tian et al. Efficiency of high-standard farmland development based on matter-element extension and TOPSIS modeling.

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant