How Machine Learning is Innovating Today's World

5Mood Detection Using Tokenization: A Review

Kuldeep Vayadande^1*, Preeti A. Bailke¹, Lokesh Sheshrao Khedekar¹, R. Kumar² and Varsha R. Dange¹

¹Vishwakarma Institute of Technology, Pune, Maharashtra, India

²VIT-AP University, Inavolu, Beside AP Secretariat, Amaravati AP, India

Abstract

Tokenization and convolutional neural networks (CNNs) are being used for the task of mood detection, emoji generation, and classification. Tokenization is helpful to break down text into individual sentences, words, and phrases, which is an important step in natural language processing (NLP). It allows the model to focus on individual words and phrases. The CNNs are then trained on labeled text datasets for mood detection, text–emoji pair datasets for emoji generation, and emoji–label pair datasets for emoji classification. In this survey paper, we have summarized various methods for tokenization and CNNs that can be effectively used to understand the sentiment or emotion expressed in text data, to acquire high accuracy in classifying text as having a negative and positive or neutral sentiment, generating emojis that match the sentiment or emotion expressed in the text and classifying emojis as expressing a certain sentiment or emotion. For sentiment analysis, a CNN can be trained on a dataset of labeled text where the labels indicate the sentiment or emotion expressed. Once trained, the CNN can be used to classify new text as expressing a positive, negative, or neutral sentiment. As far as ...