Social media changes the way of communicating between people, where people could share their opinions which are used for predicting although those changes improve sentiment analysis techniques to predict how people think, and how they react. Using people’s opinions is a powerful strategy to enhance a Machine Learning model to predict election results. Regarding the big changes in the communications between people collecting their information across social is the most powerful platform to gain valuable insights.
The accuracy of algorithms like Naive Bayes and SVM which perform sentiment analysis depends on the quantity as well as the quality of training data, and most of our models suffer from a lack of training data for text classification which affect accuracy, so we must go sentiment analysis to determine the polarity and inclination of vast population towards a specific topic, item or entity. We recommend creating an active learning model that will be able to identify what data should be labeled to minimize efforts for labelling.
Political field benefits, helping us to promote the regions where the candidate’s opinions are approved or not approved. Obtaining real opinions from people through social media like Twitter by creating a dataset using tweets API applied the Naive Bayes Classifier to obtain public and predict election results. we determine public preference before, during, and after elections and compared them with actual election results. Increase awareness of political candidates’ public image and the sentiment on their social media
Various algorithms for natural language processing especially sentiment analysis is used nowadays as Multinomial Naïve Bases Algorithm and SVM which fall under the category of supervised learning algorithms that depend on training datasets to perform Sentiment analysis on retrieved tweets. In “Election Result Prediction using Twitter sentiment analysis” in 2018. The authors compare the performance of two popular sentiment analysis algorithms, namely Naive Bayes and SVM, which perform multistage classification and identify whether the sentiment of a tweet is positive or negative.
For more Info about the used techniques Please Check the provided presentation.