<![CDATA[The test design of the Public Sentiment Analysis on the Ratification of the Job Creation Bill with the RapidMiner Studio application. The initial stage is to collect data in the form of tweets of Twitter users and then put it into a CSV file, the data obtained will be divided into training data and test data. Furthermore, the training data will be labeled consisting of 2 types of labels, namely Positive and Negative labels, then the data will be cleaned from unneeded words such as Mention or Hastag, then the data will go through several stages in the Preprocessing stage to convert raw data into data that is ready to be processed. Furthermore, each word will be weighted with the TF-IDF method. The final result of the comparison with these two test methods, namely the prediction of Public Sentiment Towards the Issue of Determining the Job Creation Bill based on data obtained from Twitter and implemented by the SVM (Support Vector Machine) method, showed an accuracy value of 96.52%. Of the 605 test data, 492 data were predicted as Negative Sentiment and 112 data as Positive Sentiment and the Naive Bayes Method showed an accuracy value of 49.67%. Of the 605 test data, 492 data were predicted as Negative Sentiment and 112 data as Positive Sentiment.]]>
Copyrights © 2022
