<![CDATA[According to data on the number of credit card users in Indonesia, it has decreased from late 2019 to 2021 oneThe reason is because of the Covid-19 pandemic that hit. Because of this condition, many people are starting to switch to digital credit because they are considered to minimize transmission of viruses and the process is felt to be more efficient than having to use a credit card. This study aims to compare the level of accuracy between the three algorithms, namely the naïve Bayes classifier, k-nearest neighbor and support vector machine for digital credit applications or often called loan services, namely Kredivo. Akulaku, and Indodana in Indonesia by classifying it into two classes namely positive and negative by using the help of the Python programming language to analyze a sentiment by going through text preprocessing and weighting processes said TF-IDF. The results for the accuracy of the Kredivo application using K-NN get a score of 84%, Naïve Bayes 88%, and SVM get 89%. For the application of the K-NN method, it gets 79%, Naïve Bayes 86%, and SVM 87%. As for the indodana application, the K-NN method gets 81%, Naïve Bayes 88%, and SVM 88%. From the results of this accuracy it can be concluded that the Support Vector Machine method has better accuracy results compared to the K-Nearest Neighbor and Naïve Bayes methods.]]>
Copyrights © 2023
