X (FORMALLY TWITTER) PRODUCT CLASSIFICATION USING NAÏVE BASE
Keywords:
Polarities, Textual, Sentiment, Vectorizing, UniqueAbstract
Finding polarities in a textual data is very important in sentiment analysis. Navie Bayes is one of the most effective machine learning classifier techniques and a probabilistic classier that applies Bayes theorem and assuming feature independence for the classification of data. The objective of this research is to implement the Naïve Bayes algorithm for the classification of sentiment in the context of X (Formally Twitter) product classification. Our aim is developing a model that can conveniently classify product-related text into positive and negative sentiment categories. The process begins with the collection of customer product reviews from X (Formally Twitter) and vectorizing each reviews making a long array of unique words, whose attributes include the independent vectors while assigned values are the number of times each independent vector appear in the product review. We had a total of one hundred and thirty-three (133) unique words in the training set for both the positive and Negative statements of which ten (10) documents were with positive (+) outcomes and ten (10) documents were with negative (-) outcomes. From the product reviews we derived the probability of each positive outcomes of independent word, probability of each negative outcomes of independent words and then the probability of each word in the product review. Our results shows that the naïve bayes classifier is a good classification technique and will be effective for both large and small businesses in making decisions related to their product development, marketing campaigns and customer support.
Published
How to Cite
Issue
Section
FUDMA Journal of Sciences