Implementation of CART (Classification and Regression Trees) Algorithm for Determining Factors Affecting Employee Performance

Abstract

The purpose of this research is to identify the problems that affect the performance of employees in the company. This research is a type of quantitative research involving data with different types ranging from nominal, continuous, categorical, ordinal and using 15 independent variables and 1 dependent variable. The population in this research are all employees of some companies in Sukabumi Indonesia. The sample used was 120 employees and the determination of the sample used a random sampling method. Methods of data collection in this study were questionnaires, interviews, and documentation. Data analysis method uses CART (Classification and Regression Trees) algorithm because it involves many kinds of data. The results showed that there are five variables that can be used for work motivation, corporate social environment, job type, leadership style, and organizational support (company policy). From these variables, the best formula can be generated to improve the performance of the employee is: increase the motivation of work (X1), making the company more conducive social environment (x11), divide the types of work (X14), which became more specific to be done individually, and if an employee is required to work in a team, it should be supported by specific policy organization/company (X3). Of the overall variables, the most influential factor in the performance of employees is the motivation of work, when higher the motivation of work, then the possibility of increased employee performance could be higher.