Open AccessExploration of Opinion from Twitter Data
Author(s) -
Nilesh Korde,
Gaurav Kawade,
Sunita Rawat,
Kavita Kalambe,
Abhijeet Ramesh Thakare
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.b3854.118419
Subject(s) - popularity , social media , computer science , sentiment analysis , data science , big data , opinion leadership , reading (process) , world wide web , profit (economics) , unstructured data , information retrieval , artificial intelligence , data mining , psychology , sociology , social psychology , social science , political science , law , economics , microeconomics
To share information nowadays, people use social media sites from all around the world. For example, Twitter is a social media site enable users with facilities like reading, sending post recognized as ‘tweets’ and interrelate with diverse peoples. People often post their sentiments about their day-to-day lives, the whole thing for example places and brands. Businesses makes profit from this vast social media site by gathering data related to sentiments of people. Presenting a model that can accomplish opinion analysis on collected data from Twitter is the aim of this paper. To analyze highly unstructured and unorganized data in Twitter makes it difficult to manage and use. In our proposed model we are combining the work of unsupervised and supervised algorithms. Extracting and classifying each tweet depending on its opinion considered to be a neutral, positive or negative. Zomato and Swiggy are the two subjects about which data were collected to show which online food delivery business has more popularity. We have used diverse machine learning algorithms for testing. Testing metrics like f-score and cross validation were used for testing the result from these models. Our model has shown performance which is considered to be robust on directly mining Twitter texts.