Sentiment analysis is the process of automatically extracting, processing, and understanding unstructured text data to obtain sentiment information contained in an opinion. This study conducted a sentiment analysis of cashless reviews using a web-based knowledge discovery database method. The knowledge discovery database method aims to extract hidden knowledge or information in previously unknown data, through the stages of data selection, preprocessing, transformation, data mining, and evaluation of 300 training data and 250 test data. Based on the original labels, there were 46 positive comments and 204 negative comments, while the prediction results showed 34 positive comments and 216 negative comments. The evaluation process produced an accuracy rate of 83.2% for positive labels and 83.2% for negative labels. The precision and recall values differed for each label, namely: positive labels had a precision of 55.88% and negative labels of 87.50%; positive labels had a recall of 41.30% and negative labels of 92.65%. Based on these scores, negative reviews were more dominant than positive reviews regarding the use of cashless. Positive sentiment focused on the ease of transactions through cashless payments, while negative sentiment focused on technical issues during payments, refunds, and frozen PayLater accounts. Therefore, e-commerce platforms implementing cashless payments need to provide better education and ensure transparency of digital processes to ensure users feel safer and more comfortable with cashless transactions.