2 pages excluding citations

Reseasrch praposal) based on this need to right the literature review/draft 40 pages Emotion can be expressed in multiple views that can be considered, such as expressions from the face, gestures, speech, and composing text. Detecting emotions from text documents is a content-based classification problem that integrates Natural Language Processing (NLP) and Machine Learning (ML). Detection, classification, and quantification of emotions from the microblogging websites data addressed in this work. Social media, for instance, Twitter, Facebook is loaded with emotions, feelings, and opinions about various products from people throughout the world. For example, plain text in English gathered from these blogs can give data having expressions in different ways, particularly sentiment mining. Be that as it may, analyzing and classifying text based on emotions is a significant challenge and can be considered a challenging/difficult type of Sentiment Analysis. This work focused on categorizing the text into multi-class emotion categories. As mentioned, the work integrates two distinct methodologies and combines them to extract the text emotions adequately. The first approach depends on NLP utilizes textual features, for instance, emotions, word count, parts of speech, etc. The subsequent methodology depends on ML classification algorithms. The proposed strategy will create an enormous bag of emotional words and intensities of emotions. Furthermore, the framework could be set up for automatic refreshing of bag-of-words based on new tweets and data that is to be analyzed. The study's problem statement is to focus on finding the best approaches for improving sentiment analysis in social media data by setting respondents to check the results by creating a hybrid model by combining NLP and Machine Learning models. The current approach of analyzing text through Machine Learning or NLP gives us a maximum accuracy of 60% to 72%. This research proposal aims to improve current efficiency by creating a hybrid model by combining NLP and Machine Learning models. In the initial phase for testing the application efficiency, I will work with benchmark datasets available on standard data source websites like Kaggle. Then, the research application will get trained using the predictive analysis behavior available through benchmark datasets because these datasets had proven to test the efficiency. Furthermore, I will use a real-time survey-based approach to analyze sample user comments confined to a geographical area from public websites. The survey would cover collecting feedback from the user to test the application user emotion quality. The evaluation methods include all the methods that are used while designing and implementing an emotional classifier. They are divided into two essential categories: a lexicon-based method and a Machine Learning (ML) method. The methods in ML use its related algorithms for system training and map a function to classify emotions in the future based on linguistic characteristics that we opted to train the system. Mining the emotions from text documents or directly from micro-blogging websites. However, classification and scoring the emotions from the text poses many challenges. The work is focused on the multi-class classification to identify the emotions from the text data. Manual annotations of the text are a complex process to identify different categories of emotion. Hence, designing a system that generates reliable training set for the classifier and managing the algorithms to label the bag-of-words and algorithms to detect and classify the emotions on multiple categories. In addition to the existing research, the Hybrid model of NLP and machine learning algorithms will give a better emotion score from the given tweet. References Cates, K., Pengcheng Xiao, Zeyu Zhang, & Dailey, C. (2018). Can Emoticons Be Used to Predict Sentiment? Journal of Data Science, 16(2), 355–375. https://doi.org/10.6339/JDS.201804_16(2).0007 Bouazizi, M., & Ohtsuki, T. (2016). Sentiment analysis: From binary to multi-class classification: A pattern-based approach for multi-class sentiment analysis in Twitter. 2016 IEEE International Conference on Communications (ICC). https://doi.org/10.1109/icc.2016.7511392