SENTIMENT ANALYSIS TERKAIT DAMPAK RUU OMNIBUS LAW CIPTA KERJA MENGGUNAKAN ALGORITMA NAIVE BAYES CLASSIFIER
Total View This Week7
Total View This Week7
Institusion
STMIK Bumi Gora
Author
Adjie, Firman Anshari
Subject
TN Mining engineering. Metallurgy
Datestamp
2021-10-06 07:22:21
Abstract :
In this period of time, many social media make it easy for us to know events and information around the world, some social media including Facebook, Twitter, Instagram, social media, etc. Twitter is one of the social media that has many users at this time. Twitter, apart from being a medium of information with various kinds of tweets, Twitter is also often used to socialize between users and express their opinions to a discussion or issues that are currently happening. The public can get information and various comments from Twitter because the number is very large, so it can be used for sentiment analysis. With so much data, the purpose of this study is to create a public sentiment analysis system in the form of positive sentiment, negative, and neutral towards the tweet of the Omnibus Law on Job Creation. Sentiment Analysis is one of the areas found in Machine Learning. Sentiment Analysis can be done by using a classification method. One of the methods used in sentiment classification is the Naïve Bayes Classifier using TF – IDF (Term Frequency – Inverse Document Frequency) weighting. This research begins by taking twitter data of 5000 tweets, using two methods, namely RapidMiner and the Twitter API (Application Programming Interface) that has been provided. The data format used in this study is CSV (Comma Separated Values). This sentiment will later be used as training data and testing data. Then the sentiment will be done Preprocessing process in which there are several stages. The output results obtained are in the form of a system that is able to classify new data by providing positive sentiment, neutral, and negative on tweets. Sentiment classification experiment using the Naïve Bayes Classifier method using the Split Train test and K-fold Cross Validation produces a Confusion Matrix with an accuracy of 77.22%. and with K-fold Cross Validation of 79.33% , 79% and , 78.7%. Then the percentage of sentiment results obtained in a row is 47.3% for neutral sentiment, 35.4% for positive sentiment, and 17.3% for negative sentimen
In this period of time, many social media make it easy for us to know events and information around the world, some social media including Facebook, Twitter, Instagram, social media, etc. Twitter is one of the social media that has many users at this time. Twitter, apart from being a medium of information with various kinds of tweets, Twitter is also often used to socialize between users and express their opinions to a discussion or issues that are currently happening. The public can get information and various comments from Twitter because the number is very large, so it can be used for sentiment analysis. With so much data, the purpose of this study is to create a public sentiment analysis system in the form of positive sentiment, negative, and neutral towards the tweet of the Omnibus Law on Job Creation. Sentiment Analysis is one of the areas found in Machine Learning. Sentiment Analysis can be done by using a classification method. One of the methods used in sentiment classification is the Naïve Bayes Classifier using TF – IDF (Term Frequency – Inverse Document Frequency) weighting. This research begins by taking twitter data of 5000 tweets, using two methods, namely RapidMiner and the Twitter API (Application Programming Interface) that has been provided. The data format used in this study is CSV (Comma Separated Values). This sentiment will later be used as training data and testing data. Then the sentiment will be done Preprocessing process in which there are several stages. The output results obtained are in the form of a system that is able to classify new data by providing positive sentiment, neutral, and negative on tweets. Sentiment classification experiment using the Naïve Bayes Classifier method using the Split Train test and K-fold Cross Validation produces a Confusion Matrix with an accuracy of 77.22%. and with K-fold Cross Validation of 79.33% , 79% and , 78.7%. Then the percentage of sentiment results obtained in a row is 47.3% for neutral sentiment, 35.4% for positive sentiment, and 17.3% for negative sentimen
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
