<![CDATA[With the increasing use of the internet, more and more services are provided online, including in the culinary world. This paper presents the classification results from the sentiment analysis of customer satisfaction in a food delivery service provider. The data used is a dataset from one of the well-known social media, namely Twitter. The data retrieved is 1000 people's tweets about their views on the services provided by service providers. The analysis begins by cleaning up data that has been drawn from some of the relatively unfavorable things in sentiment analysis research such as punctuation and links. In this study, three algorithms will be used, namely Stocastic Gradient Descent, Support Vector Machine, and K-Nearest Neighbor. The data for modeling were given two different assignments, namely without and with TF-IDF. The results obtained from the three models were evaluated with accuracy, F value and ROC curve area. The three models get results above 80% for each evaluation result and get an increase of about 1% after using TF-IDF assistance in selecting each feature used in this study.]]>
Copyrights © 2022
