2019 International Conference on Innovative Computing (ICIC) | 2019
A Dataset for the Sentiment Analysis of Indo-Pak Music Industry
Abstract
The continuous increase in data creates a need, that data be analysed and useful hidden patterns be found and explored. If the data is readily available, it can easily be analysed. But most of the time it needs to be dug. Substantial increase in the use of social media and online services can be witnessed nowadays. People not only buy and sell things online but also give their remarks on those items/services. Many websites provide such services along with a dedicated section for reviews and comments. These items/services can be ranked and analysed based upon these reviews, which shows the sentiments of the reviewer. These reviews are in millions and promptly in billions. The huge increase in reviews there is a need to analyse them through a proper mechanism. This research is targeting the mining of the sentiments from these reviews. Three songs from YouTube are selected and their reviews are scrapped, pre-processed and analysed using Decision Tree (ID3) and Naïve Bayes. Both presented 75%accuracy on test data. This article presents a dataset to perform Sentiment Analysis on Roman Urdu/Hindi reviews. Dataset is a combination of Indo-Pak song reviews.