Log Linear Regression Model for the Identification and Blocking of Fraudulent Application
Benitha Christinal J, Johnson A, Hariharan T, Anandha Kumar E · 2024
Our lives now revolve around mobile applications, yet it can be difficult to assess their reliability and safety. To solve this, a method that predicts app safety using characteristics including ratings, reviews, in-app purchases, and adverts has been developed. To assess the efficacy of the system, three ML (Machine Learning) models–DT (Decision Tree), LR (Log linear Regression), and NB (Naïve Bayes)–were compared. The LR model performed the best, with a precision of 0.87, a recall of 0.85, an F1 score of 0.815, and an accuracy of 85%. These findings demonstrate the model's dependability in evaluating app safety. This approach of using ML models automates evaluation and yields more reliable and consistent outcomes. Mobile apps are becoming increasingly crucial in our day-to-day lives, and this technology represents a promising step towards assuring the safety and reliability of mobile applications.