Performance Analysis of Machine Learning Algorithms using Fake News Detection

Abstract

Abstract: Machine learning has vast algorithms in which each and everything is specialized to predict and compute certain functionalities and tasks. A neural network is a collection of neurons in which every neuron holds a numerical value as the output of other neurons. Furthermore, neural networks are classified more as regular neural networks, convolutional neural networks, feed-forward neural networks, and long-term memory networks. Each is specialized in unique scenarios, such as regular neural networks work better in position-based image detection and convolutional neural networks work better in edgebased detection. Therefore, this paper bases the comparison of different algorithms on a base prediction problem, which is fake news detection. This project displays various performance metrics such as accuracy score and various visualization plots like scatter, pie, and bar. so that users will know the different algorithms and their scalability on non-relatable text patterns. Keywords: Fake news detection, Machine learning, Naïve bayes Gaussian NB, Classification, Regression, Support Vector Classifier, Decision Tree Classifier.