An Automated Users Profile Classification in Instagram to Improve the Precision Using an Integrated Neural Network Model and Decision Tree Algorithm

The aim of the proposed work is to improve the precision of automated fake profile detection and to reduce security threats on Instagram using a Novel Integrated Neural Networks (NINN) and comparing it to a Decision Tree (DT) algorithm. The research used two sample groups, each with 20 samples. The calculations were performed using the clinic calc tool with an alpha value of 0.05, G-power rate of 0.80, last beta rate of 0.2, and a confidential interval (CI) value of 95%. NINN are trained using a dataset and a cost function, and their performance is measured by the precision of their output. They consist of the standard layers of Neural Network that applies the NINN algorithm, and an output layer that displays the precision of the algorithm. After conducting a thorough analysis, the Novel Integrated Neural Networks Algorithm was found to have a precision of 90.91%, while the Naive Bayes Algorithm had a precision of 85.45%. An Independent samples T - Test was then performed which has p=0.000 (p<0.05) and significant in results. In this research article on Automated fake profile classification on Instagram, the Novel Integrated Neural Networks algorithm was found to have a higher prediction precision percentage (90.91%) than the DT (85.45
^%
). The proposed use of the NINN shows some improvement in the precision of fake profile detection compared to the DT algorithm, providing users with a more accurate way to identify fake profiles. This research provide a more effective solution for detecting fake profiles on Instagram and achieved the proposed aim of this article.