Position recommendation method

Abstract

The invention discloses a position recommendation method which comprises the following steps: S0, acquiring all position data, user historical behavior data and a position currently browsed by a target user, and acquiring a position browsing sequence of each user according to all the user historical behavior data; S1, generating a new position sequence group by using the position browsing sequence of each user in combination with a random walk algorithm; S2, training a Word2vec model by adopting a Skip-gram framework to generate a feature vector of each position; and S3, calculating feature vector cosine similarity between the position currently browsed by the target user and all positions, and forming a position candidate set. The position candidate set is obtained as a position recommendation set by using the feature vector cosine similarity between the position currently browsed by the target user and all the positions, and the position recommendation method is high in real-time performance and accuracy, and avoids the problems of Matthew effect and cold start.

Description

technical field [0001] The invention belongs to the technical field of recommendation systems, and more specifically, relates to a job recommendation method. Background technique [0002] Most of the job recommendation methods in current human resources communication platforms and recruitment websites are based on traditional recommendation algorithms, including collaborative filtering, content-based recommendation methods, and hybrid recommendation methods. [0003] After searching, the authorized announcement number is CN 105893641 B, which is a Chinese patent titled a job recommendation method. According to the user preference model and job model, the content-based multi-domain rating value is calculated to obtain the first rating value of the job, and the jobs are sorted. set; if a job has a posting record and belongs to the job set, then according to the user preference model and job data, calculate the job’s second rating value based on the similarity of user backgroun...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the technical problems of poor real-time performance and low accuracy of the existing job recommendation algorithm, the present invention provides a job recommendation method. The job candidate set is obtained as a job recommendation set at a high degree, which improves the real-time performance and accuracy of the recommendation method; in addition, relevant jobs with high popularity from multiple dimensions are added to the job candidate set, which further solves the cold start problem and improves the recommendation effect; Finally, the online recommendation model is obtained according to the historical browsing position time of all users and whether they are collected as tags, and the online recommendation model is used to further predict and score the positions in the position candidate set, which further improves the accuracy of the recommendation method; the position of the present invention The recommended method avoids the problems of Matthew effect and cold start, and has high real-time performance and accuracy

Images