according to the sentiments it gives about any product, person or any political party. It also has many applications in different field like in political science, social and stock market research etc. In various previous studies, it was seen that the studies deal with classification of a tweet using a suboptimal approach, it is because of their goal was not to estimate the label of the class of tweets of people, but estimating the relative frequency of various classes in the dataset. In this paper, we proposed Deep LSTM-RNN which uses the Factor approach of sentences and make the semantic relationship between words. After semantic relation makesthe non-linear mapping of activation function and get abstract features. These features are used for learning in Soft-max approach and make the classifier model. In experiments, the tweets are classified according to veg and non-veg and also classified in positive and negative sentiments. In the analysis, the proposed approach is compared to machine learning and deep learning approach and proposed approach shows significant improvement in Precision, Recall,and Accuracy.
Volume 11 | 02-Special Issue
Pages: 1432-1438