PERANCANGAN APLIKASI ANALISIS SENTIMEN TERHADAP OPINI PENGHAPUSAN SKRIPSI PADA TWITTER MENGGUNAKAN METODE NAVIE BAYES

Abstract

Sentimen analysis is a text data mining model to obtain data about positive, neutral, or negative sentimen. Sentimen analysis is provided for people on social media in the delivery of reviews related to developing issues. Twitter is a social media where anyone is free to express their opinions. Twitter has almost 600 million users and generates more than 250 million tweets per day. Therefore, Twitter is considered a rich source of information. One of the topics or issues that have been developing lately in the world of education is related to the statement about the thesis being abolished. With the application of text mining techniques and classification methods we can determine whether the sentimen given by these people is positive, neutral, or negative. Algorithms that are often used in sentimen analysis with the Naïve Bayes classifier method. The challenge of this research is that with unstructured data and large amounts of data, it will be difficult to process and classify the data. To make it easier to process and classify the data, preprocessing stages are carried out before the twitter data is analyzed using the Naïve Bayes algorithm. In the initial stage, the input string will be cut from a sentence and then the process of removing words that are ambiguous or not needed. Then the words are first converted into text based on the basic word so that when weighting the value of the word between several sentences with different affixes is the same value and changing capital letters to lowercase letters. The results of the literature study and testing with this method resulted in the Naïve Bayes algorithm getting an accuracy value of 87.40% from a total of 54 tweets that were used as sample data.