Premium
Minkowski Sommon Feature Map‐based Densely Connected Deep Convolution Network with LSTM for academic performance prediction
Author(s) -
Ramanathan Kaviyarasi,
Thangavel Balasubramanian
Publication year - 2021
Publication title -
concurrency and computation: practice and experience
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.309
H-Index - 67
eISSN - 1532-0634
pISSN - 1532-0626
DOI - 10.1002/cpe.6244
Subject(s) - computer science , artificial intelligence , classifier (uml) , convolution (computer science) , deep learning , feature (linguistics) , precision and recall , pattern recognition (psychology) , minkowski space , layer (electronics) , machine learning , data mining , artificial neural network , mathematics , chemistry , geometry , organic chemistry , philosophy , linguistics
Summary Student academic performance prediction plays a major role in the current educational systems to improve the quality of education. The conventional single classifier‐based predictive analysis is not efficient to provide accurate results. In this paper, a novel technique called Minkowski Sommon Feature Map Densely connected Deep Convolution Network with LSTM (MSFMDDCN‐LSTM) is introduced to predict the academic performance of students with higher accuracy and lesser time consumption. The MSFMDDCN‐LSTM technique uses a densely connected deep convolution network to learn the given input for accurate prediction. The student activities are collected and stored in the organization dataset. The MSFMDDCN‐LSTM technique starts with the data collection followed by performing the attributes selection and classification. The collected data are given to input layer to predict the students' academic achievement at the end of study program. Secondly, the importance of numerous dissimilar attributes or “features” is considered for student performance prediction using steepest descent Minkowski sommon mapping. After that, the classification is performed using LSTM to classify the input instances for accurate prediction. Finally, the classification results are observed in the output layer. The quantitative outcomes inferred that MSFMDDCN‐LSTM technique performs well in terms of achieving higher precision, recall, f‐measure, and lesser time consumption than the state‐of‐the‐art methods.