| dc.description.abstract |
This research aims to improve the performance of the Random Forest algorithm in predictingemployee salaries by improving the hyperparameter tuning method using GridSearchCV. This method is intended to optimize critical parameters in the Random Forest algorithm to increase the accuracy of employee salary predictions. This research uses a dataset that includes various employee-related features, such as education, work experience, and others.Through the experiments carried out, the results show that by using GridSearchCV for hyperparameter tuning, the performance of the Random Forest algorithm in predicting employee salaries can be significantly improved compared to the manual tuning approach. The results of this research: By conducting 125 iteration experiments, the smallest MSE number was obtained in 125 iterations with a combination of model search using a learningrate of 0.01, batch size 100, epoch hidden state 512 and window size 30. There are results from calculating the MSE error rate to get the results 0.8730706456934657. These findingsprovide an important contribution in improving the effectiveness of employee salary prediction models and have practical implications in the context of human resource management.
Key word : Random forest, salary, prediction, employees, hyperparameters
| en_US |
