Tren Marketplace Berdasarkan Klasifikasi Ulasan Pelanggan Menggunakan Perbandingan Kernel Support Vector Machine

Currently, many Indonesian people like to conduct online trading transactions. However, a number of business people find it difficult to choose a marketplace to market their products. One of the reasons is because they rarely pay attention to the marketplace trends that consumers are discussing. Therefore, analyzing trends on social media such as Twitter, it becomes very important for business people to understand the pattern of consumer tendencies towards their services or products. So the purpose of this study is to create a model that can analyze marketplace trends based on the classification of customer reviews on Twitter using the SVM algorithm. The kernels used are linear, RBF, sigmoid, and polynomial with parameter optimization using grid search. The methodology used is KDD. The results of the evaluation of the best classification model are the sigmoid kernel with 92% accuracy, 92% precision, 92% recall, and 92% F1 score and parameters C=100, =0.01, and r=1. Market trend results based on the highest percentage of positive reviews are Tokopedia, Shopee, and lastly Bukalapak