Open AccessANALISIS SENTIMEN PADA TWITTER TERHADAP UIN RADEN FATAH MENGGUNAKAN SUPPORT VECTOR MACHINE
Author(s) -
Gusmelia Testiana
Publication year - 2022
Publication title -
jatisi (jurnal teknik informatika dan sistem informasi)/jatisi: jurnal teknik informatika dan sistem informasi
Language(s) - English
Resource type - Journals
eISSN - 2503-2933
pISSN - 2407-4322
DOI - 10.35957/jatisi.v9i1.1433
Subject(s) - support vector machine , social media , sentiment analysis , computer science , service (business) , measure (data warehouse) , perception , quality (philosophy) , public opinion , recall , advertising , artificial intelligence , psychology , world wide web , data mining , politics , political science , business , marketing , philosophy , epistemology , neuroscience , law , cognitive psychology
Measuring customer satisfaction is one of the most important aspects of any successful companyto improve the quality of its service,therefore collecting reviews is highly recommended.However, gathering datais not enough, without having efficient and reliable automated systemsthat capable of analysing the data and extracting valuable information for further improvement.Nowadays, social media is getting more and more attention as public and private opinions onvarious subjects are expressed and disseminated continuously through various social media.Twitter offers a fast and effective way to analyse people's perspectiveson important things forsuccessful universities. Developing a program for sentimentanalysis is the approach tocomputationally measure people's perceptions. In this study, public opinion regarding UIN RadenFatah Palembang was analysed using Support VectorMachine (SVM) method in determining thepositive or negative sentiment of a tweet by doing initial processing for unstructured data fromTwitter. The results indicated that the polarity of sentiment towards UIN Raden Fatah Palembangon Twitter as seen from 100 samples of tweets, 89 (89%) had positive sentiments and 11 (11%)had negative sentiments. The level of accuracy in the classification of sentiment using SVM was70% with an average precision of 20.6%, anaverage recall of 70% and an average f-measure of62.7%.