Project on salary classification

Abstract

In this project, The results use three different machine learning algorithms to approach salary classification. The analyzed data used many different variables such as education level, age, and work-class to label each person into two categories, one with a salary greater than 50k and the other with a salary less than or equal to 50k. First of all, this work uses a single decision tree model to visualize data because it is more concise and understandable, and then by using the support vector machine method, the result becomes more accurate. After building two different models, The accuracy was found to be about 86.32%, which is relatively high and reliable. However, higher accuracy may be more persuasive. So, this project uses another model which is the random forest model. This algorithm is considered a highly accurate method because of the number of decision trees that participated. This model explained 87.03% of the accuracy of my result. According to my models, if a person desires a wage increase, that person should do his best to improve his education level, and he needs to have a stable marriage situation and be able to start his own business as much as possible between the ages of 20 to 60.