Multilingual Contemplation for Social Media data by Hybrid Classifier
Challapalli Manoj, Talluru Tejaswi, Devi Sohana Nekkanti, Sriram Sai Valiveti, Vithya Ganesan, Viswanathan Ramasamy · 2022 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES) · 2022
Sentiment analysis procure substantial contemplation for abandon growth of the web and social media. The goal is to identify opinions, sensation, or conviction of multilingual tweet. A large set of manually labeled tweets in different languages, use them as training data, and construct sentiment analysis model. A machine-learning algorithm is used to identify and extract characteristics such as adjectives, adverbs, and abstract nouns from each cleaned tweet. If the input text is not in English, the Google Machine Translator is utilized to convert it into English. A hybrid model called voting-based ensemble model which integrates Naive Bayes, Multinomial Naive Bayes, Bernoulli’s Naive Bayes, Logistic Regression, Stochastic Gradient Descent, Support Vector Machines, and Maximum Entropy classifiers to resolve the ambiguous language. If, same word appears in both English and non-English it utilizes a hybrid model retrieved features to translate a tweet /string. The classifiers identify the sentiment of translated multilingual tweets and extended with prediction.