THE SENTIMENTAL ANALYSIS USING DEEP LEARNING MODELS

THE SENTIMENTAL ANALYSIS USING DEEP LEARNING MODELS

M.SRISANKAR1   Dr.K.P.LOCHANAMBAL2

1Research Scholar, Department of Computer Science, Government Arts College, Udumalpet, Bharathiar University, Tamilnadu, India. Shankar276@gmail.com

2Assistant Professor, Department of Computer Science, Government Arts College, Udumalpet, Bharathiar University, Tamilnadu, India. kplmca@rediffmail.com

ABSTRACT:The tweets are brief and come in a steady stream. Emotions have a significant impact on feelings. People can express their ideas about anything and anything on social media. Public perception is divided into three categories: positive, negative, and neutral. In this study, Twitter hotel reviews are gathered and pre-processed before being analyzed using Python's Tweepy package. Re-tweets, tags, URLs, hash tag symbols, and duplicate entries are all eliminated as part of a screening procedure to remove any discrepancies in the data. Using Python's scikit-learn module, tweets are up-sampled and divided. Python turns textual data into vectors using the keras Tokenizer. Bi-sense Emoji Embedding (BSEE) is used to perform a sentimental analysis. Sentiment is categorized using Support Vector Machines (SVM) and Random Forest (RF),and  LSTM (Long Short Term Memory) where compared based on accuracy, recall, F-measure, precision, time duration, and performance. It is clear that the proposed classifier produces better results.

Keywords: Bi-Sense Emoji Embedding (BSEE), Long Short Term Memory (LSTM), Support Vector Machine (SVM), Random Forest (RF),