KLASIFIKASI TIPE HARGA PONSEL MENGGUNAKAN RECURSIVE FEATURE ELIMINATION DAN SUPPORT VECTOR MACHINE

The transition of the Industry 4.0 era to Society 5.0 is happening, making cellphone business players strive to increase their product sales. Through innovation that focuses on developing the production process with consideration of observing customer and market needs. Researchers want to initiate a study that raises the utilization of artificial intelligence with the aim of being an alternative
solution in the process of developing cellphone production. In this case, researchers will make a classification of cellphone price types using datasets from kaggle which will be modeled in Recursive Feature Elimination (RFE) and Support Vector Machine (SVM). Not only that, although the recommended use of SVM kernels in RFE is linear, but researchers want to make a comparison with non-linear kernels (rbf and sigmoid). The accuracy results given, the first rank was occupied by linear SVM from 96% to 97%, then SVM rbf from 90.8% to 93.5%, and SVM sigmoid from 88.5% to 83%. The highest increase in accuracy was obtained in SVM rbf by +2.7% and a decrease in accuracy was obtained by SVM sigmoid by -5.5%.