Automated Fake News Detection Using Logistic Regression over Decision Tree with Improved Accuracy
Chakravarthy Balaji, Anitha G. S., V PAVITHRA, Salim Lahmiri · 2024
This paper's primary goal is to use machine learning techniques, specifically Logistic Regression and Decision Trees, to identify bogus news on social media. An innovative logistic model is employed to attain accuracy. To measure accuracy and loss, datasets from the Kaggle library are used. There are twenty samples in all. Decision trees (N=10) and logistic regression (N=10) were the two categories that were taken into account. Logistic regression yields a 93.68% accuracy and a 6.32% loss, which seems to be better than Decision Tree, which yields an 81.71% accuracy and an 18.39% loss, respectively. In conclusion, it seems that Logistic Regression performs substantially better than Decision Trees. The independent sample T-Test value (p= 0.001, 2 tailed in SPSS analysis) for both the Decision Tree (DT) and Logistic Regression (LR) algorithms shows statistical satisfaction with a confidence level of 95%. Logistic regression appears to be a considerably superior method than decision trees for detecting bogus news.