<![CDATA[Cryptocurrency is a virtual currency that is used as an alternative currency, where the currency is generated and traded through a cryptographic process. One of the Cryptocurrency digital currency applications is Tokocrypto, a legal means of payment using only rupiah. The sentimen classification method in this study uses Support Vector Machine (SVM) and the parameters are tested extensively by applying the K-fold cross validation technique to find the optimal configuration. The initial stages in this sentimen analysis are the data collection stage, pre-processing, which consists of case folding & cleaning, filtering (stopword removal), tokenizing, changing negated words, normalization and stemming. After that, data labeling and scoring are carried out using the Lexicon Based method. The dataset from pre-processing and Lexicon Based is used for the classification process using SVM. The best evaluation results were obtained with an AUC (Area Under the Curve) of 90.50% and an Accuracy of 85.80% with Linear kernel and parameter C = 1 and using K-fold cross validation value of 10. Visualization of the results of the sentimen analysis of the Tokocrypto Application is displayed in a bar chart and word cloud.]]>
Copyrights © 2024
