Efficient Prediction of Malicious Social Bots with URL Features using DenseNet Compared over ResNet50 with Improved Accuracy

Thestudy's goal is to more accurately identify dangerous social bots by combining URL characteristics and reinforcement learning techniques to detect them on the Twitter network for enhancing accuracy. Recurrent Neural Network algorithm detects dangerous social bots more accurately than Support Vector Machine algorithm by combining URL characteristics with reinforcement learning technique. In this case, the initial test statistical evaluation was done using 80%, and a sample size of 20 persons was chosen, with 10 participants in each of the two groups.. RNN algorithm with 92.11% accuracy was compared with SVM algorithm with 90.28% accuracy to identify dangerous social bots utilizing the URL features in the Twitter network. The independent sample test reveals a statistically significant value in accuracy between two algorithms is 0.008 <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(p &lt; 0.05)$</tex>. From the obtained results, the Novel RNN algorithm performs better than the SVM algorithm in terms of accuracy.