Optimized deep network based students performance analysis for college admissions

The careful analysis of students' data can help to know the behaviour and performance of the students. With the support of technology-enhanced learning platforms, we can use the available educational data to analyze students' academic elegance, solve their problems, improve the learning environment, and allow statistical decision-making. Therefore, we presented a Butterfly-based Deep Neural Framework (BbDNF) to analyze the student data to predict academic performance for college admissions. Initially, the students are collected from the OULA dataset and trained in the system. The collected data are preprocessed to produce noise and error-free data. Furthermore, the preprocessed data is fed into the feature selection phase to select the relevant features utilizing the Butterfly fitness function. Moreover, they executed the predictions phase to analyze the students' performance and categorize them. The proposed system is implemented in the Python tool. The comparison of the suggested system's performance with the other existing system is made in the name of f1-score, precision, accuracy and recall, and the overall efficiency of the designed system is studied.