Optimalisasi Model Klasifikasi Naive Bayes dan Support Vector Machine Dengan Fast Text dan Chi Square Pada Analisis Sentimen Penyelenggaraan Pembelajaran Pemrograman di Fasilkom Universitas Mercu Buana

Fajriah, Riri (Universitas Budi Luhur)
Kurniawan, Denni (Universitas Budi Luhur)

Faktor Exacta
Vol 17, No 4 (2024)

The implementation of effective programming learning at the Faculty of Computer Science, Universitas Mercu Buana is one of important strategy. This expectation is constrained because the results of the evaluation of the competency achievements of many graduates have not mastered programming skills well. Therefore, the research conducted is related to analyzing the sentiments of all stakeholders who have been involved with the implementation of programming learning. The data source based on the results of an online questionnaire. The sentiment data analysis process uses the Cross Industry Standard Process for Data Mining method with the Naive Bayes and Support Vector Machine classification models. The result of the research is an increase in the accuracy of sentiment analysis data processing which previously only used the Naive Bayes Algorithm only achieving an accuracy of 65.56% and by optimizing with Feature Extraction Fast Text, the accuracy achievement increased to 90.49%. While optimizing the algorithm using Feature Selection Chi Square can make the Support Vector Machine classification model optimized to achieve an accuracy value of 99.58% from the previous accuracy achievement was 90.72%. This research can prove that optimizing the application classification model algorithms can use using Fast Text and Chi Square techniques.