Sentiment Analysis on Twitter Data Using Deep Learning approach
Vishu Tyagi, Ashwini Kumar, Sanjoy Das
Abstract
The recent developments of many social networking websites have created large collections of product reviews and polarity of opinions etc, data for customer around the world. Data collection from these social media can be utilized to solve objectives such as market prediction, product recommendation, and reviewer's sentiment. It is very difficult task to manage of unstructured data available on social media. To handle such type of data, Deep learning algorithms is an appropriate solution for analysis these challenges. In this paper, we propose a CNN- LSTM based deep learning method with pre trained embedding approach learns to extract feature automatically for analysing sentiments and classification of reviews or opinions labelled into two polarity as positive or negative. Our proposed model to implement the result gives better performance on benchmark dataset. The performance of CNN-LSTM based deep learning method has compared with baseline machine learning methods.