Strange Approach of Movie Rating Prediction Using Logistic Regression Comparing to Gaussian Naive Bayes Algorithm

Abstract

The aim is to find movie ratings using logistic regression and comparing the result with naive bayes based on Accuracy. A total of 6040 samples were collected from movie datasets available in kaggle. Two algorithms are used; one is Logistic Regression and another is naive bayes algorithm. The computation processes were executed and verified for exactness. Sample size N=5 is taken for both algorithms. SPSS was used for predicting significance value of the dataset considering G-Power value as 80%. Logistic Regression achieved mean accuracy of 80.83% when compared to Naive Bayes Algorithm with 82.53%. Results were obtained with a level of significance with 0.003 (p<0.05). Applied strange recommendation model confirms to have higher accuracy than Naive Bayes algorithm.