Abstract
In this paper we propose a method to analyze the sentiments of a user on a particular topic on the basis of their tweets. This method breaks up the texts into words, remove all the unnecessary stop words and then using a dictionary determines the sentiments by grouping certain words together. Using Tokenization it first splits those tweets into smaller units and using lemmatization it reduces the infected words to meaningfulwords before comparing those with the dictionary elements. Our purpose is to establish perceptive links between tweets and public sentiment for a better understanding of the public opinion. In this paper we present two applications of this method: recognizing depression in tweets and recognizing mass opinion. In the first application we use certain keywords to search for tweets that tend towards depression and our program analyzes the exact sentiments of the tweet to confirm the existence of depression. In the second application, tweets regarding a certain topic are scraped and they are analyzed together to give a result that represents the public opinion.
Recommended Citation
Singh, Sakshi; Nandy, Alekhya; Das, Panchajanya; Bhowmik, Sudeshna; Biswas, Subham; Gupta, Rajdeep Sen; and Mandal, Dr.Sandip
(2024)
"A Novel Approach for Data Scraping and Sentiment Analysis from Twitter using Machine Learning Techniques,"
American Journal of Electronics & Communication (AJEC): Vol. 2:
Iss.
3, Article 1.
Available at:
https://research.smartsociety.org/ajec/vol2/iss3/1