Salary Predictor System for Thailand Labour Workforce using Deep Learning

Abstract

The purpose of this research is to build the Salary Predictor System to predict monthly salary of employees in Thailand using the Deep Learning approach, which has rapidly increased distinguish attention in machine learning field. The dataset has gathered from well-known job search website which has more than 1.7 million users. Personal data from the first five months of 2018 is applied to the analysis and construct this model. We compared the performance with related algorithms such as Random Forest and Gradient Boost Trees. The feature selection methods were applied to Deep Learning after comparing. As a result of combining the feature selection with Deep Learning, the optimal result was 0.462 in R-squared and the rapid runtime is 15.37 seconds.