Loading...
Please wait, while we are loading the content...
Similar Documents
Analisis Sentimen Twitter dengan Klasifikasi Naïve Bayes menggunakan Seleksi Fitur Mutual Information dan Inverse Document Frequency
| Content Provider | Semantic Scholar |
|---|---|
| Author | Putra, Riky Sutriadi |
| Copyright Year | 2017 |
| Abstract | RIKY SUTRIADI PUTRA. Twitter Sentiment Analysis using Naïve Bayes Classification with Mutual Information and Inverse Document Frequency for Feature Selection. Supervised by JULIO ADISANTOSO. The objective of this research is to classify Twitter tweet data into three sentiments: positive, negative, and neutral. The data used in this research were 3195 tweets consisting of 1065 positive, 1065 negative, and 1065 neutral. The data were divided into two subsets which consist of 80% for training and 20% for testing. The training data were used in feature selection step with 5 experiments and the test data were used for testing in an existing classification system made using Multinomial Naıve Bayes method. This research yielded the average accuracy of 75.27% for Mutual Information and 74.33% for Inverse Document Frequency feature selection. The average accuracy rate for each sentiment for Mutual Information and Inverse Document Frequency models are respectively 65.07% and 50.33% for the positive sentiment. For the negative sentiment, the numbers are 69.86% and 50.52%, and 80.84% and 80.75% for the neutral sentiment. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://repository.ipb.ac.id/bitstream/handle/123456789/87323/G17rsp.pdf?isAllowed=y&sequence=1 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
