Abstract: Nowadays, participating moments on communicative networks have enhanced something comprehensive. Sharing impressions, opinions, and good representations to formulate our sentiments within the text without using a lot of information. Twitter, for occurrence, is a rich reservoir of data that is a destination for directions for which they can use to investigate people’s judgments, thoughts and emotions. Sentiment analysis gives a more informed overview of the characteristics of an author opinion typically. In large Social Media examination, nearly all projects have focused on analyzing the expressions as positive, negative or neutral. In this work, we design to characterize the terms based on sentiment classes called happiness, anger, fear, and sadness. Several procedures have been brought in the area of dynamic textual sentiment identification in the case of additional communications, but only an insufficient number were based on deep training. This work represents the expansion of a novel deep learning-based scheme that discusses the various emotion distribution difficulties on in-formative data. We recommend an innovative method to reconstruct it into a binary distribution as well as traditional machine learning classification problem and utilize an indepth knowledge strategy to determine the reconstructed problem. Our hybrid approach will provide better classification accuracy over classical machine learning algorithms.Keywords: machine learning, RNN, NLP, text sentiment analysis, social data analytics