The Research And Prediction Of Vertical Industry Job Salary Based On Deep Learning

To take advantage of the immense job data available on the internet,job seekers and employers are increasingly utilizing the internet to job searches and recruitment respectively,which is associated with a large amount of job data.However,while companies gain access to potential employees' information,companies often fail to identify salary.In order to make this process equitable for job seekers,it is critical to establish the fair and competitive salaries for positions they might be applying.This paper mainly analyze and predict job salary through the Internet job data,we put forward two new salary analysis and predicting models,bring more transparency to the job market.The current study seeks to achieve this by first studying the traditional machine learning model of salary prediction.This requires explaining the deficiencies in this method and analyzing the feasibility of the job salary analysis and prediction for the deep learning algorithm.Now that online recruitment is popular,there is a surge in salary data on the Internet,which also leads us to analyze the salary of the position using the deep learning algorithm model.Therefore,this paper analyzes the existing salary forecasting methods and in depth study on deep learning algorithms,and applies the deep learning algorithm to the job salary analysis and prediction research of vertical industry to improve the accuracy of the salary forecast.This paper aims to study the following four aspects:Firstly,this paper studies the traditional machine learning model of salary prediction,and explains the deficiencies in the method,and analyzes the feasibility of the job salary analysis and prediction for the deep learning algorithm.Secondly,to address this process's shortcomings,we put forward two salary prediction models proposed based on the deep learning algorithm model.The structure of both models validated through a comparison experiment.The mean absolute error value of the salary prediction models in the test set reduce 5.7%and 7.2%respectively compared with the traditional models,and the prediction effect of the model was significantly improved.Thirdly,to enhance this,we use deep convolutional neural network prediction model trained by salary data of vertical industry to calculate the salary of UK jobs in Adzuna.The fine adjustment of network parameters then carried out,and we verify the generalization ability of the model on different data sets through comparative experimentsFinally,we optimize two models of salary prediction from two different angles,and we also introduce the attention mechanism and the contextual feature in the model,then the prediction effect of the salary prediction model is further improved.