Research On The Influencing Factors Of Online Recruitment Salary Based On Machine Learning

With the optimization and adjustment of China's economic structure and the advent of the information age,online recruitment is becoming more and more popular,and online recruitment information is more and more,and the salary situation is one of the most concerned things for every job seeker,so the research on the factors affecting the salary can provide relevant reference for job seeker,so that job seeker can obtain the employment demand from the massive recruitment information Information.In this paper,we use Python language to write a crawler program to obtain the job data of data analysis,machine learning,data mining and deep learning in Zhilian recruitment,build a prediction model of salary level,and analyze the factors that affect the salary level in these job data.This paper mainly studies the salary forecast model from the two aspects of xgboost model and gbdt model.The main work is as follows:(1)Two crawler strategies,depth first and width first,are used to crawl the position data in recruitment information through multiple processes.Through mastering the algorithmic process of xgboost model and gbdt model,the theoretical basis for the research of the model is established.(2)There are a lot of structured and unstructured text data in the recruitment data obtained in this paper.Exploratory analysis and visualization technology are used to analyze the influence relationship between each variable and salary.Text processing technology and visualization technology are used for the text data of unstructured variables.These two technologies are used to visualize the information in the text data and to extract the skill information in the job recruitment data by using the word cloud diagram in the visualization.Through the analysis of structured data and the extraction of skill information of unstructured data,the discrete classification features are transformed into binary vector representation by using the unique heat coding technology,and these data encoded by unique heat are combined with numerical salary data to build the data features required by the model.It paves the way for the prediction of the model and the analysis of the influencing factors of salary.(3)Xgboost model and gbdt model are used to analyze the influencing factors of salary.According to the exploratory analysis,we know the variable characteristicsthat affect the salary factors.The gbdt and xgboost algorithm models in machine learning algorithm are used to optimize the parameters and get the optimal parameters.The parameters selected through the optimization are used to predict the model of salary,score the importance of variable characteristics and rank the factors affecting salary.The main influencing factors are obtained through sorting.Finally,using the training set accuracy,test set accuracy and RMSE evaluation index,the prediction effect of the model is compared and analyzed.The results show that xgboost model is better than gbdt model,so xgboost model is more suitable for the analysis of salary factors.