Paper Title
Study of factors affecting program ratings in Workpoint Entertainment stations via a Machine Learning algorithm

Abstract - Television influences the mindset of Thais. Thailand has comparable television viewing rates. The success of a television program depends on its ratings. Television stations' longevity and financial viability depended on advertising rates, which were affected by viewer competition. Therefore, channel content is essential. This study gathered data from Workpoint Entertainment Public Company Limited and inserted them into Machine Learning (ML) algorithms utilizing the RFR method. A total of 91 datasets. A schematic representation of the RFR algorithm was the ML algorithm used in this study. The accuracy of the RFR model was about 0.9310 and 0.9569 of R2 for prediction and regression, respectively. From the prediction results, Programs were the most critical variable affecting the program's rating, 48%, followed by time and period, 15% and 10%, respectively. According to the Partial dependence report, the broadcast time of Workpoint Entertainment television stations with higher ratings should be broadcast in the range of 90 to 120 minutes, it can raise the rating, and from predictions, Two-dimensional partial, the broadcast date and the highest ratings of the program were Friday and Sunday at 7:00 p.m. Time slots from 10:00 a.m. to 4:00 p.m. do not affect the rating increase. In Random Forest Regression (RFR) in ML for the model, prediction helps forecast future new programs being produced on Workpoint Entertainment television stations. It can also be used for improvement and planning within the television station to be more successful in the programs. Keywords - Artificial intelligence, Random Forest Regression (RFR), Rating, Television station management