Abstract:

The vast increment in data on social media creates a need for reducing such dataset to get valuable insights. Sentiment analysis addresses such need by determining opinions or emotions and the true meaning on the social media text. Sentiment analysis has aroused the interest of many researchers in recent years, especially on online reviews and it has become a hot research field. This study analyses the true meaning of the retrieved keyword on a Twitter’s tweet and computes the polarity of it. Social media is the domain of study on Twitter’s tweet on the act of pedophiles that leads to child sexual abuse. The research being going on to determine the keywords with the correct meaning using various methods, it is still differed in terms of accuracy of the meaning. Hence, this system will give the output that identifies the polarity of the keyword and the whole sentences by using Machine Learning’s Naïve Bayes technique. This machine learning technique can assist in processing a natural human language on the proposed keyword.Stanford NLP Library is used to run this Sentiment Analysis program in which an existing model or also known as classifier is provided by the Stanford NLP Group and therefore, there is no need to create a new model for this. The model is integrated into Stanford CoreNLP as of version 3.8.0. This includes the model and the source code, as well as the parser and sentence splitter needed to use the sentiment tool . Hence, the default model is used to test it out and conduct several experiments and modification been done upon further enhancing the program to possible meet the objectives. When it comes to modelling the program, a classifier/model is needed and it has been trained with sentences that has already been tokenized and annotated with their respective label which is in the scale of 0-4.This study shows the prove of concept of Naïve Bayes algorithm into supervised data such as Twitter’s tweet on the act of pedophiles on social media. Naïve Bayes is simple model and can be implemented where this classifier returns the polarity with the higher count. If there is a tie, then positive polarity of the majority class is returned

Keywords:

Social Media, Sentiment Analysis, Machine Learning, Naïve Bayes