Disclosure of Invention
The invention overcomes the defects of the prior art and provides a high-end talent recommending method and system based on data analysis.
In order to achieve the above purpose, the invention adopts the following technical scheme:
The first aspect of the invention provides a high-end talent recommendation method based on data analysis, which comprises the following steps:
the method comprises the steps of obtaining high-end talent recruitment requirements, dividing the matching priority of recruitment indexes according to the high-end talent recruitment requirements to obtain a priority matching area and a secondary matching area, and training the priority matching area and the secondary matching area to obtain a recruitment requirement screening framework;
Acquiring the application information of each high-end talent, introducing an LDA algorithm to extract keywords in the application information of each high-end talent, constructing a database according to the extracted keywords to obtain an application information database, and introducing the recruitment requirement screening frame into the application information database to carry out reverse index;
Introducing a fuzzy clustering algorithm to calculate membership clustering relations between each recruitment index and each keyword in the application information database when the inverted index is adopted to obtain a priority aggregation set, and analyzing whether the priority aggregation set accords with a screening standard of a priority matching area or not to obtain a first high-end talent application information set;
And acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, screening high-end talents conforming to the secondary matching area by combining the keywords of the residual high-end talent application information and the recruitment indexes in the secondary matching area to obtain a second high-end talent application information set, combining the first high-end talent application information set and the second high-end talent application information set, and generating a high-end talent recommendation scheme.
Further, in a preferred embodiment of the present invention, the acquiring a high-end talent recruitment requirement, dividing the matching priority of the recruitment index according to the high-end talent recruitment requirement to obtain a priority matching area and a secondary matching area, and training the priority matching area and the secondary matching area to obtain a recruitment requirement screening framework, which specifically includes the following steps:
Acquiring browsing information of a user on a recruitment platform, and carrying out demand analysis on the browsing information to obtain one or more high-end talent recruitment demands; wherein the high-end talent recruitment requirements include professions, academics, graduation institutions, and skill attributes;
extracting a plurality of recruitment indexes through one or more high-end talent recruitment requirements, introducing an entropy weight algorithm to carry out weight assignment on the plurality of recruitment indexes, calculating the information entropy of each recruitment index to obtain a plurality of information entropy values, and dividing the information entropy value corresponding to each recruitment index by the sum of the information entropies of all indexes to obtain a plurality of recruitment index weight values;
presetting a weight grading threshold, constructing an ascending list, and importing the weight values of the plurality of recruitment indexes into the ascending list for sorting to obtain a weight ascending list;
After the sorting is finished, the matching priority of the high-end talent indexes is divided in a weight ascending sorting list based on the weight grading threshold, all the recruitment index weight values higher than the weight grading threshold are defined as priority matching areas, and all the recruitment index weight values lower than the weight grading threshold are defined as secondary matching areas;
And constructing a screening framework based on a deep learning algorithm, and guiding the priority matching area and the secondary matching area into the screening framework for training to obtain a recruitment requirement screening framework.
Further, in a preferred embodiment of the present invention, the acquiring the application information of each high-end talent, introducing an LDA algorithm to extract keywords in the application information of each high-end talent, constructing a database according to the extracted keywords, obtaining an application information database, and introducing the recruitment requirement screening frame into the application information database to perform reverse indexing, which specifically includes the following steps:
Acquiring all the talent information of a recruitment platform, comprehensively analyzing all the talent information, generating a high-end talent list, acquiring the talent information of each high-end talent based on the high-end talent list, and extracting a plurality of pieces of application text data from each piece of the talent information;
Introducing an LDA algorithm to extract keywords from the plurality of application text data, distributing a theme to each application text data set, distributing a word to the theme, iterating theme distribution of each word until the preset iteration times are met, so as to obtain a plurality of theme distribution probabilities and word distribution probabilities in the plurality of application text data, and extracting keywords according to the plurality of theme distribution probabilities and word distribution probabilities in the plurality of application text data to obtain a plurality of keywords;
Acquiring the application information corresponding to each keyword, mapping and arranging the keywords and the application information corresponding to each keyword through an inverted index technology, recording an application information list corresponding to each keyword, and outputting the list to obtain an application information database;
and importing the recruitment demand screening framework into the recruitment information database to carry out reverse index, and preferentially screening all high-end talent recruitment information meeting the priority matching area to carry out merging and pushing.
Further, in a preferred embodiment of the present invention, the acquiring all the talent information of the recruitment platform, performing comprehensive analysis on all the talent information, and generating a high-end talent list specifically includes the following steps:
Acquiring all the recruitment talents of a recruitment platform, and extracting a plurality of assessment score information and performance score information based on the all the recruitment talents;
introducing a hash algorithm to calculate hash values among the plurality of assessment achievement information and the plurality of performance evaluation information to obtain a plurality of hash values;
Constructing a Manhattan distance matrix, calculating Manhattan distances between a plurality of hash values and a preset hash value in the Manhattan distance matrix to obtain a plurality of Manhattan distances, and extracting hash values corresponding to the Manhattan distances larger than the preset Manhattan distances;
And determining the applied talent information corresponding to each hash value according to the extracted hash values, defining the applied talents as high-end talents, and drawing a list based on the applied talent information corresponding to each hash value to obtain a high-end talent list.
Further, in a preferred embodiment of the present invention, a fuzzy clustering algorithm is introduced during reverse indexing to calculate membership clustering relations between each recruitment index and each keyword in the application information database to obtain a priority aggregation class set, and whether the priority aggregation class set meets a screening standard of a priority matching area is analyzed to obtain a first high-end talent application information set, which specifically includes the following steps:
Acquiring all recruitment indexes in a priority matching area, introducing a fuzzy clustering algorithm to calculate membership clustering relations between all recruitment indexes and all keywords in the application information database, presetting a membership value range and a fuzzy coefficient, distributing initial membership between each recruitment index and each keyword based on the membership value range and the fuzzy coefficient, and determining a plurality of initialization membership matrixes;
Calculating Euclidean distances between each recruitment index and each keyword according to a plurality of initialization membership matrixes, and carrying out weighted normalization on the Euclidean distances based on a weighted average algorithm to obtain a plurality of clustering centers;
Acquiring an updating rule in a fuzzy clustering algorithm, iteratively updating each initialized membership matrix in the fuzzy clustering algorithm based on the updating rule and a plurality of clustering centers, extracting membership degrees, and generating a plurality of updated initial membership degrees;
If the updated initial membership is greater than a membership threshold, representing that the keywords corresponding to the updated initial membership meet the recruitment index in the priority matching area, and extracting and combining all the corresponding keywords greater than the membership threshold to obtain a priority aggregation set;
And determining the name number of the recruitment index in the priority matching area based on the high-end talent recruitment requirement, judging whether the initial membership number corresponding to each recruitment index in the priority clustering set is larger than the name number, if so, directly outputting the priority clustering set, and if so, continuing to iterate and extract keywords in a fuzzy clustering algorithm until the number is larger than the name number, and generating a first high-end talent recruitment information set.
Further, in a preferred embodiment of the present invention, the method obtains keywords of remaining high-end talent application information based on the first high-end talent application information set, combines the keywords of the remaining high-end talent application information with recruitment indexes in the secondary matching area to screen high-end talents conforming to the secondary matching area to obtain a second high-end talent application information set, combines the first high-end talent application information set and the second high-end talent application information set, and generates a high-end talent recommendation scheme, which specifically includes the following steps:
after the high-end talents meeting all the recruitment indexes in the priority matching area are matched, carrying out secondary analysis matching on the residual high-end talents screened by the first high-end talents application information set according to all the recruitment indexes in the secondary matching area;
Acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, acquiring all recruitment indexes in a secondary matching area, and calculating membership degrees between the keywords of the residual high-end talent application information and all the recruitment indexes based on a Gaussian membership function method to obtain a plurality of Gaussian membership degrees;
a gray correlation analysis method is introduced to calculate the correlation among the plurality of Gaussian membership degrees, a reference sequence is preset in a gray correlation analysis algorithm, and the correlation coefficient between the sequences of the plurality of Gaussian membership degrees and the reference sequence is calculated to obtain a plurality of correlation coefficients;
Extracting a standard reaching threshold according to the high-end talent recruitment requirement and all recruitment indexes in a secondary matching area, if the association coefficient is larger than the preset association coefficient, extracting the matching keyword of the association coefficient larger than the preset association coefficient, obtaining the extraction quantity of the matching keyword, and judging whether the extraction quantity of the matching keyword is larger than the standard reaching threshold of the keyword extraction;
If the keyword extraction amount is larger than the keyword extraction standard threshold value, outputting high-end talent application information corresponding to the extracted keywords, if the keyword extraction amount is smaller than the keyword extraction standard threshold value, calculating a difference value between the extracted keyword extraction amount and the keyword extraction standard threshold value to obtain a deviation value, and rescreening and matching the keywords based on the deviation value until the deviation value is equal to 0, so that a second high-end talent application information set is generated;
and combining the first high-end talent application information set and the second high-end talent application information set and pushing the first high-end talent application information set and the second high-end talent application information set on a recruitment platform to obtain a high-end talent recommendation scheme.
The second aspect of the present invention provides a high-end talent recommendation system based on data analysis, the high-end talent recommendation system based on data analysis includes a memory and a processor, the memory stores a high-end talent recommendation method program based on data analysis, and when the high-end talent recommendation method program based on data analysis is executed by the processor, the following steps are implemented:
the method comprises the steps of obtaining high-end talent recruitment requirements, dividing the matching priority of recruitment indexes according to the high-end talent recruitment requirements to obtain a priority matching area and a secondary matching area, and training the priority matching area and the secondary matching area to obtain a recruitment requirement screening framework;
Acquiring the application information of each high-end talent, introducing an LDA algorithm to extract keywords in the application information of each high-end talent, constructing a database according to the extracted keywords to obtain an application information database, and introducing the recruitment requirement screening frame into the application information database to carry out reverse index;
Introducing a fuzzy clustering algorithm to calculate membership clustering relations between each recruitment index and each keyword in the application information database when the inverted index is adopted to obtain a priority aggregation set, and analyzing whether the priority aggregation set accords with a screening standard of a priority matching area or not to obtain a first high-end talent application information set;
And acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, screening the high-end talents conforming to the secondary matching area by combining the keywords of the residual high-end talent application information and the recruitment indexes in the secondary matching area, and finally generating a high-end talent application recommendation scheme.
Further, in a preferred embodiment of the present invention, the method obtains keywords of remaining high-end talent application information based on the first high-end talent application information set, combines the keywords of the remaining high-end talent application information with recruitment indexes in the secondary matching area to screen high-end talents conforming to the secondary matching area to obtain a second high-end talent application information set, combines the first high-end talent application information set and the second high-end talent application information set, and generates a high-end talent recommendation scheme, which specifically includes the following steps:
after the high-end talents meeting all the recruitment indexes in the priority matching area are matched, carrying out secondary analysis matching on the residual high-end talents screened by the first high-end talents application information set according to all the recruitment indexes in the secondary matching area;
Acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, acquiring all recruitment indexes in a secondary matching area, and calculating membership degrees between the keywords of the residual high-end talent application information and all the recruitment indexes based on a Gaussian membership function method to obtain a plurality of Gaussian membership degrees;
a gray correlation analysis method is introduced to calculate the correlation among the plurality of Gaussian membership degrees, a reference sequence is preset in a gray correlation analysis algorithm, and the correlation coefficient between the sequences of the plurality of Gaussian membership degrees and the reference sequence is calculated to obtain a plurality of correlation coefficients;
Extracting a standard reaching threshold according to the high-end talent recruitment requirement and all recruitment indexes in a secondary matching area, if the association coefficient is larger than the preset association coefficient, extracting the matching keyword of the association coefficient larger than the preset association coefficient, obtaining the extraction quantity of the matching keyword, and judging whether the extraction quantity of the matching keyword is larger than the standard reaching threshold of the keyword extraction;
If the keyword extraction amount is larger than the keyword extraction standard threshold value, outputting high-end talent application information corresponding to the extracted keywords, if the keyword extraction amount is smaller than the keyword extraction standard threshold value, calculating a difference value between the extracted keyword extraction amount and the keyword extraction standard threshold value to obtain a deviation value, and rescreening and matching the keywords based on the deviation value until the deviation value is equal to 0, so that a second high-end talent application information set is generated;
and combining the first high-end talent application information set and the second high-end talent application information set and pushing the first high-end talent application information set and the second high-end talent application information set on a recruitment platform to obtain a high-end talent recommendation scheme.
The invention solves the technical defects existing in the background technology, and has the beneficial technical effects that:
The method comprises the steps of obtaining high-end talent recruitment requirements, dividing and training matching priorities of recruitment indexes according to the high-end talent recruitment requirements to obtain a recruitment requirement screening framework, obtaining the recruitment information of each high-end talent, constructing a database according to the recruitment information of each high-end talent to obtain an recruitment information database, importing the recruitment requirement screening framework into the recruitment information database to conduct inverted indexing, introducing a fuzzy clustering algorithm to calculate membership clustering relations between each recruitment index and each keyword in the recruitment information database during inverted indexing, obtaining a first high-end talent recruitment information set, obtaining keywords of residual high-end talent recruitment information based on the first high-end talent recruitment information set, and screening the high-end talents conforming to a secondary matching area by combining the keywords of the residual high-end talent recruitment information and the recruitment indexes in the secondary matching area to finally generate a high-end talent recommendation scheme. The invention can realize the recommendation of the personalized high-end talent scheme, and improve the recruitment quality and efficiency of the high-end talents, thereby meeting the recruitment requirement of users.
Detailed Description
In order that the above-recited objects, features and advantages of the present application will be more clearly understood, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present application and features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
The first aspect of the present invention provides a high-end talent recommendation method and system based on data analysis, as shown in fig. 1, comprising the following steps:
S102: the method comprises the steps of obtaining high-end talent recruitment requirements, dividing the matching priority of recruitment indexes according to the high-end talent recruitment requirements to obtain a priority matching area and a secondary matching area, and training the priority matching area and the secondary matching area to obtain a recruitment requirement screening framework;
S104: acquiring the application information of each high-end talent, introducing an LDA algorithm to extract keywords in the application information of each high-end talent, constructing a database according to the extracted keywords to obtain an application information database, and introducing the recruitment requirement screening frame into the application information database to carry out reverse index;
s106: introducing a fuzzy clustering algorithm to calculate membership clustering relations between each recruitment index and each keyword in the application information database when the inverted index is adopted to obtain a priority aggregation set, and analyzing whether the priority aggregation set accords with a screening standard of a priority matching area or not to obtain a first high-end talent application information set;
S108: and acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, screening high-end talents conforming to the secondary matching area by combining the keywords of the residual high-end talent application information and the recruitment indexes in the secondary matching area to obtain a second high-end talent application information set, combining the first high-end talent application information set and the second high-end talent application information set, and generating a high-end talent recommendation scheme.
Further, in a preferred embodiment of the present invention, the acquiring a high-end talent recruitment requirement divides the matching priority of the recruitment index according to the high-end talent recruitment requirement to obtain a priority matching area and a secondary matching area, and trains the priority matching area and the secondary matching area to obtain a recruitment requirement screening framework, as shown in fig. 2, specifically including the following steps:
S202: acquiring browsing information of a user on a recruitment platform, and carrying out demand analysis on the browsing information to obtain one or more high-end talent recruitment demands; wherein the high-end talent recruitment requirements include professions, academics, graduation institutions, and skill attributes;
S204: extracting a plurality of recruitment indexes through one or more high-end talent recruitment requirements, introducing an entropy weight algorithm to carry out weight assignment on the plurality of recruitment indexes, calculating the information entropy of each recruitment index to obtain a plurality of information entropy values, and dividing the information entropy value corresponding to each recruitment index by the sum of the information entropies of all indexes to obtain a plurality of recruitment index weight values;
S206: presetting a weight grading threshold, constructing an ascending list, and importing the weight values of the plurality of recruitment indexes into the ascending list for sorting to obtain a weight ascending list;
S208: after the sorting is finished, the matching priority of the high-end talent indexes is divided in a weight ascending sorting list based on the weight grading threshold, all the recruitment index weight values higher than the weight grading threshold are defined as priority matching areas, and all the recruitment index weight values lower than the weight grading threshold are defined as secondary matching areas;
S210: and constructing a screening framework based on a deep learning algorithm, and guiding the priority matching area and the secondary matching area into the screening framework for training to obtain a recruitment requirement screening framework.
It should be noted that, the analysis of the high-end talent recruitment requirement of the user is a precondition of accurate recommendation, and the personalized customization recommendation scheme is provided for the user by analyzing the high-end talent requirement of the user, so that the recommendation scheme meets the user expectation, and the user requirement customization modules of most recruitment platforms at present have single functions, limited indexes capable of being screened and cannot realize high-quality personalized customization; the high-end talent recruitment requirement of the user comprises a plurality of recruitment indexes, wherein the recruitment indexes are obtained by carrying out weight calculation on each recruitment index through an entropy weight algorithm, the weight value can intuitively embody the importance degree of the index, and favorable data support is provided for primary and secondary grading; the weight values obtained through calculation are arranged in an ascending order, and the weight values after the sorting are divided into primary and secondary grades through a preset weight grading threshold value, so that a primary recruitment index matching area which is required to be met is formed, the primary recruitment index matching area is defined as a priority matching area, a secondary recruitment index matching area is formed, and the secondary recruitment index matching area is defined as a secondary matching area; because the priority matching area and the secondary matching area contain all recruitment requirements of high-end talents required by the user, the priority matching area and the secondary matching area can be used as training bases of a screening framework, and finally the screening framework with the recruitment requirements personalized to the requirements of the user can be obtained through training by a deep learning algorithm. According to the invention, the personalized customization screening framework can be constructed according to the high-end talent recruitment requirement of the user, so that the recruitment platform can accurately and efficiently recommend the high-end talents according to the user requirement, and the high-end talent recruitment expectation of the user is met.
Further, in a preferred embodiment of the present invention, the acquiring the application information of each high-end talent, introducing an LDA algorithm to extract keywords in the application information of each high-end talent, constructing a database according to the extracted keywords, obtaining an application information database, and introducing the recruitment requirement screening frame into the application information database to perform reverse indexing, as shown in fig. 3, specifically including the following steps:
s302: acquiring all the talent information of a recruitment platform, comprehensively analyzing all the talent information, generating a high-end talent list, acquiring the talent information of each high-end talent based on the high-end talent list, and extracting a plurality of pieces of application text data from each piece of the talent information;
S304: introducing an LDA algorithm to extract keywords from the plurality of application text data, distributing a theme to each application text data set, distributing a word to the theme, iterating theme distribution of each word until the preset iteration times are met, so as to obtain a plurality of theme distribution probabilities and word distribution probabilities in the plurality of application text data, and extracting keywords according to the plurality of theme distribution probabilities and word distribution probabilities in the plurality of application text data to obtain a plurality of keywords;
S306: acquiring the application information corresponding to each keyword, mapping and arranging the keywords and the application information corresponding to each keyword through an inverted index technology, recording an application information list corresponding to each keyword, and outputting the list to obtain an application information database;
S308: and importing the recruitment demand screening framework into the recruitment information database to carry out reverse index, and preferentially screening all high-end talent recruitment information meeting the priority matching area to carry out merging and pushing.
After the personalized screening framework is built, if the screening framework is led into a talent screening system of a traditional recruitment platform, the system can be caused to be unadapted, and most importantly, the matching result can have the phenomena of screening omission, screening errors and the like, so that the accuracy of matching screening is reduced, and therefore, a database of high-end talent application information with high matching precision is required to be used for matching the index matching of the screening framework; analyzing all the recruitment talents information of the recruitment platform, screening a list conforming to high-end talents, extracting text data according to the high-end talent recruitment information on the list, introducing an LDA algorithm, and extracting keywords in the text data to construct a database, wherein the keywords comprise the recruitment information which is matched with the recruitment index in height, so that the accuracy of screening and matching can be improved by a method for constructing the database by the keywords, and meanwhile, the suitability of the screening frame can be greatly improved by adopting a mode of matching the keywords with the recruitment index, so that the recommendation accuracy of the high-end talents is ensured; the LDA algorithm can extract information keywords from the text data center without omission, so that the stability and the integrity of the database are further ensured; and finally, constructing a database by an inverted index technology, wherein the inverted index is a data structure for quickly searching the documents, and can correlate the keywords with the documents containing the keywords, so that the keywords required by the recruitment indexes in the screening frame can be quickly searched and matched. The invention can establish the high-end talent search database which has high adaptation degree and can be searched efficiently and rapidly for the recruitment demand screening framework, thereby ensuring that the recommended high-end talents are accurate and correct and meet the demands of users.
Further, in a preferred embodiment of the present invention, the acquiring all the talent information of the recruitment platform, performing comprehensive analysis on all the talent information, and generating a high-end talent list specifically includes the following steps:
Acquiring all the recruitment talents of a recruitment platform, and extracting a plurality of assessment score information and performance score information based on the all the recruitment talents;
introducing a hash algorithm to calculate hash values among the plurality of assessment achievement information and the plurality of performance evaluation information to obtain a plurality of hash values;
Constructing a Manhattan distance matrix, calculating Manhattan distances between a plurality of hash values and a preset hash value in the Manhattan distance matrix to obtain a plurality of Manhattan distances, and extracting hash values corresponding to the Manhattan distances larger than the preset Manhattan distances;
And determining the applied talent information corresponding to each hash value according to the extracted hash values, defining the applied talents as high-end talents, and drawing a list based on the applied talent information corresponding to each hash value to obtain a high-end talent list.
It should be noted that, since there are numerous recruiters on a recruitment platform, including high-end talents and ordinary talents, high-end talents meeting the requirements need to be extracted precisely from numerous recruiters; the assessment score information and the performance score information are selected as the standard for distinguishing the high-end talents from the common talents, so that the difference between the high-end talents and the common talents can be distinguished more directly, and the extraction rate is improved; and a hash algorithm is introduced to comprehensively analyze the assessment achievement information and the performance achievement information, the length of the hash value is obtained by calculating the Manhattan distance, and the applied talents meeting the requirements can be clearly and rapidly screened according to the length, so that the extraction accuracy is improved, and finally a high-end talent list is generated. The invention can extract the high-end talents in the recruitment platform and provide basis for the establishment of the follow-up high-end talent recommendation scheme.
Further, in a preferred embodiment of the present invention, a fuzzy clustering algorithm is introduced during reverse indexing to calculate membership clustering relations between each recruitment index and each keyword in the application information database to obtain a priority aggregation class set, and whether the priority aggregation class set meets a screening standard of a priority matching area is analyzed to obtain a first high-end talent application information set, which specifically includes the following steps:
Acquiring all recruitment indexes in a priority matching area, introducing a fuzzy clustering algorithm to calculate membership clustering relations between all recruitment indexes and all keywords in the application information database, presetting a membership value range and a fuzzy coefficient, distributing initial membership between each recruitment index and each keyword based on the membership value range and the fuzzy coefficient, and determining a plurality of initialization membership matrixes;
Calculating Euclidean distances between each recruitment index and each keyword according to a plurality of initialization membership matrixes, and carrying out weighted normalization on the Euclidean distances based on a weighted average algorithm to obtain a plurality of clustering centers;
Acquiring an updating rule in a fuzzy clustering algorithm, iteratively updating each initialized membership matrix in the fuzzy clustering algorithm based on the updating rule and a plurality of clustering centers, extracting membership degrees, and generating a plurality of updated initial membership degrees;
If the updated initial membership is greater than a membership threshold, representing that the keywords corresponding to the updated initial membership meet the recruitment index in the priority matching area, and extracting and combining all the corresponding keywords greater than the membership threshold to obtain a priority aggregation set;
And determining the name number of the recruitment index in the priority matching area based on the high-end talent recruitment requirement, judging whether the initial membership number corresponding to each recruitment index in the priority clustering set is larger than the name number, if so, directly outputting the priority clustering set, and if so, continuing to iterate and extract keywords in a fuzzy clustering algorithm until the number is larger than the name number, and generating a first high-end talent recruitment information set.
It should be noted that, the recruitment requirement screening frame is imported into the recruitment information database to perform reverse index, and because the priority matching area in the screening frame contains the recruitment index which must be met by the high-end talents, the high-end talents' recruitment information required in the priority matching area needs to be subjected to priority matching screening; calculating and analyzing membership cluster relations between recruitment indexes in a priority matching area and all keywords in an recruitment information database by introducing a fuzzy clustering algorithm, and initializing an initial membership matrix in the fuzzy clustering algorithm by distributing initial membership, so that the possibility that the keywords belong to each recruitment index can be reflected; then calculating and initializing a clustering center of keywords and recruitment indexes in a membership matrix so as to play a role in data decision on the update of membership, and carrying out weighted normalization on Euclidean distance through a weighted average algorithm so as to ensure the accuracy of the clustering center; then, carrying out iterative updating on the membership according to the updating rule and the clustering center, and reflecting the membership between the keywords and the recruitment indexes through continuous updating, thereby improving the screening accuracy of high-end talents; because the names of the recruitment indexes of the priority matching areas are limited, the hard demands of users on the high-end talents can be basically met by screening the high-end talents according to the names of the recruitment indexes in the priority matching areas. The invention can calculate the membership of the recruitment index in the priority matching area in the application information database, match the high-end talents according to the membership, improve screening efficiency and screening matching accuracy, and avoid the occurrence of the phenomenon of error of the matching result, thereby realizing the effect of recommending the hard index demand of the high-end talents according to the priority of the user.
Further, in a preferred embodiment of the present invention, the method obtains keywords of remaining high-end talent application information based on the first high-end talent application information set, combines the keywords of the remaining high-end talent application information with recruitment indexes in the secondary matching area to screen high-end talents conforming to the secondary matching area to obtain a second high-end talent application information set, combines the first high-end talent application information set and the second high-end talent application information set, and generates a high-end talent recommendation scheme, which specifically includes the following steps:
after the high-end talents meeting all the recruitment indexes in the priority matching area are matched, carrying out secondary analysis matching on the residual high-end talents screened by the first high-end talents application information set according to all the recruitment indexes in the secondary matching area;
Acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, acquiring all recruitment indexes in a secondary matching area, and calculating membership degrees between the keywords of the residual high-end talent application information and all the recruitment indexes based on a Gaussian membership function method to obtain a plurality of Gaussian membership degrees;
a gray correlation analysis method is introduced to calculate the correlation among the plurality of Gaussian membership degrees, a reference sequence is preset in a gray correlation analysis algorithm, and the correlation coefficient between the sequences of the plurality of Gaussian membership degrees and the reference sequence is calculated to obtain a plurality of correlation coefficients;
Extracting a standard reaching threshold according to the high-end talent recruitment requirement and all recruitment indexes in a secondary matching area, if the association coefficient is larger than the preset association coefficient, extracting the matching keyword of the association coefficient larger than the preset association coefficient, obtaining the extraction quantity of the matching keyword, and judging whether the extraction quantity of the matching keyword is larger than the standard reaching threshold of the keyword extraction;
If the keyword extraction amount is larger than the keyword extraction standard threshold value, outputting high-end talent application information corresponding to the extracted keywords, if the keyword extraction amount is smaller than the keyword extraction standard threshold value, calculating a difference value between the extracted keyword extraction amount and the keyword extraction standard threshold value to obtain a deviation value, and rescreening and matching the keywords based on the deviation value until the deviation value is equal to 0, so that a second high-end talent application information set is generated;
and combining the first high-end talent application information set and the second high-end talent application information set and pushing the first high-end talent application information set and the second high-end talent application information set on a recruitment platform to obtain a high-end talent recommendation scheme.
After the high-end talents required by the priority matching area are screened and matched, most recruitment indexes contained in the priority matching area are hard indexes, so that the basic requirements of users are basically met after the priority matching area is screened and matched, but the recommendation of the high-end talents is more comprehensive and more referential, so that the high-end talents screened by the set secondary matching area can provide references for the users and provide various selection possibilities for the users; the screening of the secondary matching area is carried out according to the principle of optimal selection, so that the screening and matching of each index are ensured to be as many as possible, the requirements of users can be met, and more reference options can be provided; the priority matching area screens out a part of high-end talents, so that only the residual high-end talents can be screened out during secondary screening, the membership degree between the keywords of the residual high-end talents application information and all the recruitment indexes is calculated through Gaussian membership functions, the relevance between the membership degrees is calculated by introducing a gray relevance analysis method, the application information with higher membership degree with the recruitment indexes is rapidly determined and extracted according to the relevance performance, and complex calculation time and steps are saved; and judging whether the high-end talents screened and matched by each index are as much as possible according to the extraction times, and finally obtaining the high-end talents application recommended scheme. The invention can supplement the additional screening scheme according to the secondary recruitment index in the user demand, thereby providing more high-end talent recommendation options for the user, improving the selection diversity, breaking the problems of selection limitation, non-abundant recommendation content and the like, and ensuring that the high-end talent recruitment demand of the user is met.
In addition, the high-end talent recommendation method based on data analysis further comprises the following steps:
acquiring high-end talent screening records of a user, and analyzing the high-end talent screening records of the user to obtain a plurality of high-end talent indexes required by the user;
Constructing a blank hash table with preset capacity, importing a plurality of high-end talent indexes into the blank hash table with preset capacity for recording to obtain an initial hash table, and importing a hash function to perform traversal calculation on the initial hash table to obtain a high-frequency screening index;
Acquiring high-end talents corresponding to each index based on the high-frequency screening index, acquiring the job achievement degree corresponding to the high-end talents, comparing the job achievement degree with a preset job achievement degree, extracting high-end talents corresponding to the preset job achievement degree, and generating a primary screening result;
Acquiring a recommendation index of each high-end talent on a recruitment platform in the primary screening result, if the recommendation index is lower than a preset recommendation index, eliminating the high-end talents corresponding to the recommendation index lower than the preset recommendation index, and integrating the remaining high-end talents to generate a secondary screening result;
and constructing a ranking table, importing a secondary screening result into the ranking table based on a recommendation index to rank from high to low to obtain a high-quality high-end talent recommendation table, and pushing the high-quality high-end talent recommendation table to a user.
It should be noted that, at present, talent application information on most recruitment platforms is editable, so that conditions such as application information, resume work and the like generally occur, so that the recommended content of high-end talents becomes false, the result of recruitment talents is unreliable, and most of high-end talents recommended to users have the condition that actual working capacity is inconsistent with expected working efficiency, and the recommended quality of the high-end talents is reduced; however, some indexes cannot be edited by self in the recruitment platform, most of the indexes are uploaded to a specially managed system, such as job achievement evaluation, performance evaluation, assessment results and the like, and the indexes intuitively reflect the professional ability and the working benefit of talents, so that high-end talents can be accurately recommended through some indexes, and the phenomenon of false recommended content is avoided. The invention can carry out high-quality screening on the high-end talents by combining the job achievement degree and the recommendation index, so that the recommended high-end talents can meet the requirements of users, and the situations of high-end talent information falsification and the like are avoided in the recommendation scheme.
The second aspect of the present invention provides a high-end talent recommendation system based on data analysis, where the high-end talent recommendation system based on data analysis includes a memory 41 and a processor 42, where the memory 41 stores a high-end talent recommendation method program based on data analysis, and when the high-end talent recommendation method program based on data analysis is executed by the processor 42, as shown in fig. 4, the following steps are implemented:
the method comprises the steps of obtaining high-end talent recruitment requirements, dividing the matching priority of recruitment indexes according to the high-end talent recruitment requirements to obtain a priority matching area and a secondary matching area, and training the priority matching area and the secondary matching area to obtain a recruitment requirement screening framework;
Acquiring the application information of each high-end talent, introducing an LDA algorithm to extract keywords in the application information of each high-end talent, constructing a database according to the extracted keywords to obtain an application information database, and introducing the recruitment requirement screening frame into the application information database to carry out reverse index;
Introducing a fuzzy clustering algorithm to calculate membership clustering relations between each recruitment index and each keyword in the application information database when the inverted index is adopted to obtain a priority aggregation set, and analyzing whether the priority aggregation set accords with a screening standard of a priority matching area or not to obtain a first high-end talent application information set;
And acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, screening the high-end talents conforming to the secondary matching area by combining the keywords of the residual high-end talent application information and the recruitment indexes in the secondary matching area, and finally generating a high-end talent application recommendation scheme.
Further, in a preferred embodiment of the present invention, the method obtains keywords of remaining high-end talent application information based on the first high-end talent application information set, combines the keywords of the remaining high-end talent application information with recruitment indexes in the secondary matching area to screen high-end talents conforming to the secondary matching area to obtain a second high-end talent application information set, combines the first high-end talent application information set and the second high-end talent application information set, and generates a high-end talent recommendation scheme, which specifically includes the following steps:
after the high-end talents meeting all the recruitment indexes in the priority matching area are matched, carrying out secondary analysis matching on the residual high-end talents screened by the first high-end talents application information set according to all the recruitment indexes in the secondary matching area;
Acquiring keywords of residual high-end talent application information based on the first high-end talent application information set, acquiring all recruitment indexes in a secondary matching area, and calculating membership degrees between the keywords of the residual high-end talent application information and all the recruitment indexes based on a Gaussian membership function method to obtain a plurality of Gaussian membership degrees;
a gray correlation analysis method is introduced to calculate the correlation among the plurality of Gaussian membership degrees, a reference sequence is preset in a gray correlation analysis algorithm, and the correlation coefficient between the sequences of the plurality of Gaussian membership degrees and the reference sequence is calculated to obtain a plurality of correlation coefficients;
Extracting a standard reaching threshold according to the high-end talent recruitment requirement and all recruitment indexes in a secondary matching area, if the association coefficient is larger than the preset association coefficient, extracting the matching keyword of the association coefficient larger than the preset association coefficient, obtaining the extraction quantity of the matching keyword, and judging whether the extraction quantity of the matching keyword is larger than the standard reaching threshold of the keyword extraction;
If the keyword extraction amount is larger than the keyword extraction standard threshold value, outputting high-end talent application information corresponding to the extracted keywords, if the keyword extraction amount is smaller than the keyword extraction standard threshold value, calculating a difference value between the extracted keyword extraction amount and the keyword extraction standard threshold value to obtain a deviation value, and rescreening and matching the keywords based on the deviation value until the deviation value is equal to 0, so that a second high-end talent application information set is generated;
and combining the first high-end talent application information set and the second high-end talent application information set and pushing the first high-end talent application information set and the second high-end talent application information set on a recruitment platform to obtain a high-end talent recommendation scheme.
The foregoing is merely illustrative embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think about variations or substitutions within the technical scope of the present invention, and the invention should be covered. Therefore, the protection scope of the invention is subject to the protection scope of the claims.