Archive | 2019
Location Based Sentiment Analysis Using PC 3 E—PNN Consensus Between Classification and Clustering Ensembles—Technique in Big Data Platform
Abstract
Twitter is a widely used micro-blogging website and it is a good source for opinion mining i.e. sentiment analysis. In politics, location-based sentiment analysis used to keep track of society’s opinions on the government, politicians, statements, policy changes, or any event to predict results of the election. Many algorithms and techniques are present for analysing the sentiments of tweets. To increase the accuracy, we propose probabilistic neural networks (PNN) based Consensus between Classification and Clustering Ensembles Technique (PC3E). Tweets are collected, pre-processed and segregated location-wise using Hadoop (HIVE). Location wise similarity matrix is constructed based on the unlabeled data and class probability distributions obtained using PNN. Then both outcomes of previous steps are fed as input into the Consensus between Classification and Clustering Ensembles (C3E) classifier. Then this PC3E classifier ensemble is smeared with data collected location-wise and the results obtained traced out on a map. PC3E classifier was tried on tweets collected on Prime Minister Narendra Modi and views of people about him in various locations are traced in a map with the help of Tableau software. The results obtained have amplified accuracy when compared with the different classifier algorithms.