GDL

Sentiment mining is the extraction of textual data automatically to see the trend of sentiment. The rapid growth of social media, opinions expressed in social media plays an influential role in our life. A need for a sentiment mining model arises, so as to enable the retrieval of tweet for how to evaluate community is responses on a community development program. Sentiment was obtained from community tweets about activities in preprocessing and the analysis. The dataset of used is a tweets about activities of community development programs in the area of Bogor. The method preprocessing, reduction feature using PCA (Principal Component Analysis), next to classification to find out whether a tweet consists of positive, negative or neutral sentiment using SVM (Support Vector Machine). The results showed, best response based on the percentage of positive sentiment is the better GenRe (Generasi Berencana) than PIK-Remaja activities. Accuracy rates showed they was not influenced by the ratio of training data and test data that are used but preprocessing tweet and parameter of SVM.

Beri Komentar ?#(0) | Bookmark