Finding suitable algorithm for Text Analysis the basic concept of Big Data

  • Mr.S.Gopalakrishnan, Dr.A.Prema


Data analytics plays a major role in each and every field of text analysis .It is one of the basic methods in data analytics. It helps in feedback analysis, customer behavior analysis, movie and product reviews. The behavior and sentiment is resulted as emotions of a person that is decided by the words used by the individuals. To brief it if a WhatsApp chat if a person is analyzed, they must have used mixed word that depicts both positive and negative words. If there are maximum numbers of positive words it is concluded that the person is happy or else it is concluded that the person is in a problematic situation.

 Sentiment analysis algorithm is dealing with the classification of those positive, negative and neutral words. Text analysis is done by some deep learning algorithms such as Support Vector Machine (SVM), Naive Basie (NB), K Nearest Neighbor (KNN),K* and DT-J48.

This Proposed research paper is based on the idea to find a suitable algorithm for text analysis .Based on the best results, comparison of each results is performed. It can analyze Structured, Unstructured and Semi Structured data. The data set for this analysis is a simple whatsapp chat combination of both text and emojis and results the emotion of an individual by counting Positive, Negative and Neutral words using Sentiment Analysis and this proposed idea is to fix the best algorithm for text analysis.

Keywords – SVM, NB, KNN, K – MEANS, DT-J48, feedback analysis and customer behavior analysis