Trend Analysis and Predictive Modeling Using Machine Learning Models on Indian Election Historical Dataset

Abstract

AbstractIndia has been among the largest democracy around the world, and democratic election process has played an important role in achieving this status. Predicting elections result is an important factor as it influences the market situation at local and global level. This paper had focused on data analytics and machine learning using Python, on historical dataset. Historical datasets of Indian Parliamentary Elections have been taken for a larger time span of 1977–2014 and results of same have been plotted using Python. In this paper, five machine learning models have been used for predicting win or loss for a seat in an election. The models primarily used for this analysis are Gaussian Naive Bayes, extra tree classifier, K-nearest neighbors classifier (KNN), logistic regression and decision tree classifier. In result section, results of five model used for study were evaluated and compared. This study had depicted that decision tree classifier provided comparatively good accuracy score among the chosen five models.KeywordsMachine learning (ML)Logistic regressionDecision treeData analyticsElectionNaive BayesPredictive modelingKNN