Integrated CNN- and LSTM-DNN-based sentiment analysis over big social data for opinion mining

The interactive and real-time characteristics of gathering public opinion through the process of investigating big social data have gained more popularity and attention from the recent past. Moreover, massive social media data wide opened an immense opportunity for businesses for extracting potential insights. However, big data analytics applications pose a crucial challenge in distinguishing opinions from the factors. In this paper, an Integrated Convolutional Neural Network and Long Short Term Memory (LSTM) Recurrent Neural Network-based Deep Neural Networks-based Sentiment Analysis Methodology (ICNN-LSTM-DNN) was proposed over the big social data for opinion mining. This proposed ICNN-LSTM-DNN-based sentiment analysis approach is an adaptable sentimental analysis mechanism that is capable of investigating social media posts and extracts user's facts and opinion in real-time. This proposed ICNN-LSTM-DNN-based sentiment analysis approach is mainly for facilitating automatic separation of facts from the opinions extracted from twitter messages posted online. This proposed ICNN-LSTM-DNN-based sentiment analysis approach was applied over the tweets associated with the 2019 Indian Election for opinion mining. This proposed ICNN-LSTM-DNN-based sentiment analysis approach outperformed different baseline techniques used for investigation in terms of accuracy, precision, recall, and F-Measure.

Automated summary

The paper introduces a sentiment analysis method based on ICNN-LSTM-DNN, which is capable of extracting user facts and opinions in real-time from large-scale social data, particularly suitable for analyzing opinions and facts in tweets.