Human Resource Recommendation Algorithm Based On Hybrid Genetic Ensemble Learning

In recent years,with the major human resources service companies actively deploying the online recruitment market,the industry's homogenization competition has been explosive.In order to enhance the company's core competitiveness,the online recruitment platform needs human resources recommendation algorithm to provide personalized recommendation services.Therefore,this paper constructs the pre-processing strategy and the storage strategy for collecting human resources data and combines the attribute information and behavior information of job seekers to study the human resources recommendation algorithm based on hybrid genetic ensemble learning.The main work of this article includes the following points:Firstly,Data Acquisition and Preprocessing: This paper designs a data preprocessing method and process.It has preprocessing operations such as cleaning,integration,protocol,and transformation which ensures the accuracy and completeness of data in the data warehouse.Secondly,Data Processing Architecture and Data Warehousing: This paper proposes a parallel human resource data processing architecture based on distributed columnar storage.It designs and implements a human resources data warehouse on distributed columnar storage,and parallelizes the algorithm implementation on Spark.Thirdly,The Human Resources Recommendation Algorithm: Based on the ideas of matrix decomposition,selective ensemble learning and mixed recommendation,this paper designs a human resources recommendation algorithm based on hybrid genetic ensemble learning.In offline training,the algorithm obtains the classification model through selective ensemble learning.In real-time recommendation,the algorithm constructs a matrix decomposition collaborative filtering algorithm based on the implicit feedback of human resources,and obtains rating information and filters out highly rated posts.Then,the characteristics of the user and the positions are used as the input of the classification model to obtain the classification information,and finally the rating information and the classification information are integrated to obtain a higher confidence recommendation result.In addition,the algorithm also improves the selective integration genetic algorithm,introduces the simulated annealing operation,adaptive crossover and mutation operation,realizes the mixed inheritance,and accelerates the speed of the algorithm to search the global optimal solution.In addition,different recommendation strategies are adopted for the "cold start" problem,which improves the ability of the algorithm to handle the "cold start" problem.This paper also verifies the key parameters and effectiveness of the algorithm.The experimental results show that the parameters of this algorithm are reasonable,and compared with the current mainstream recommendation algorithm,the recall rate and F value are better.At the same time,the hybrid genetic algorithm has a stronger global search ability than the traditional genetic algorithm,and can effectively improve the speed of selection integration,thereby improving the overall performance of the algorithm.