Effective prediction of fake news using two machine learning algorithms

Abstract

The study's main aim is to detect the fake news from the political information. To detecting this types of fake news we are implementing more effective machine learning classifier algorithm. These algorithms will compare their performance by using methods and materials. The materials and methods are belonging to the LRA (Logistic Regression algorithm) and the NBA (Naive Bayes algorithm). These two types of methods and algorithms are implemented to test the data's. This algorithm tested more than 44,000 records. These two algorithms are performed well to detect the fake information from the dataset. The experimental performance rate is N = 10. These two algorithms identified different types of fake news and provide accurate news classification. The eighty percentage result are used by the G-power test. Finally, the experimental accuracy results of the fake political news for these two algorithms as follows 98.7080 using the Logistic Regression algorithm and 94.8490 using the Naive Bayes algorithm and 0.013 is the statistical difference between these two algorithms. We used T-test to find out the accuracy of the independent samples. When we compared these two results the Logistic Regression algorithm provided better performance.