“…To comprehensively evaluate the model's performance, the researchers implemented eight different scenarios for splitting the data, encompassing ratios of [0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95]. The grid search was conducted over a range of parameters, including the degree parameter [2,3,4,5,6], C values [0.1, 1, 10, 100, 1000], and coef0 values [0.0, 0.1, 0.5, 1.0].The evaluation of the model's performance was conducted through various metrics, including confusion matrix metrics such as precision (Pre), recall (Rec), F1score (F1-scr), and accuracy (Acc). These metrics are based on the outcomes of TP, TN, FP, and FN.In summary, the study encompasses a comprehensive analysis of sentiment classification in Indonesian online marketplace reviews, considering different data split scenarios and employing SVM with linear and non-linear kernels.…”