Volume V | Issue 4 | 2026 | Paper ID: IJATEM-V05I04P3

Student Academic Performance Prediction Using Hybrid Mamba-Transformer Network with Adaptive Attention Learning

Author(s)

, Ramya. R

Cite This

", International Journal of Advanced Trends in Engineering and Management, vol. 05, no. 04, pp. 24-30, 2026.

Abstract

The prediction of student performance has become an essential component for online educational platforms, as it improves student engagement in learning, supports adaptive learning strategies and enables timely academic interventions. The diverse data sources provide valuable insights into student behaviour and academic progress, making virtual learning environments a rich and promising field for educational research and predictive analytics. Each learning activity record consists of two categories of features: student behavioural features and exercise-related features. In this analysis, it is acknowledged that each factor has a distinct effect on student performance and combination of activity with behaviour among students features. It is crucial for enhancing intelligence prediction reliability among students for providing better options in their employments. In this study, raw educational data is transformed to structured and deep machine learning format through pre-processing unit that helps in achieving robustness and better accuracy, while data visualization paves way in selecting patterns and variables relationships before training the model. Then the processed data is splitted for training, validation and testing which in turn selection of data takes place and fed into classifier model for performance prediction. Mamba transformer network is employed to improve quality of image data, using multi-layer processing recognition accuracy is improved. Use of deep learning, which is classified under artificial intelligence (AI), is crucial to the field of performance prediction of intelligent students in universities. Therefore, this study develops a Mamba transformer model along with adaptive attention mechanism to find out performance of students based on activity-related data and student behaviours, which performs with better accuracy of 97.3%, precision of 97% and its F1 score value is 96% than previous integrative models.

Keywords

Artificial intelligence (AI), Hybrid Mamba-Transformer Network, Adaptive Attention Learning, Student performance prediction.

References

[1] Mfowabo Maphosa; Wesley Doorsamy; Babu S. Paul, Year: 2023, “Student Performance Pa tterns in Engineering at the University of Johannesburg: An Exploratory Data Analysis”, in IEEE Access, Vol: 11, pp. 48977-48987.

CrossRef Google Scholar Publisher Link
[2] Hongkun Qiu; Haojie Zhu; Xiaoyong Hu; Yajie Wang; Yue Xu; Wenlong Li, Year: 2026, “Predicting Student Performance Using a Cost-Sensitive Deep Learning Network With Sparse Attention Mechanism”, in IEEE Transactions on Education, Vol: 69, No: 1, pp. 10-18.

CrossRef Google Scholar Publisher Link
[3] Zhonghua Zhan; Tongping Shen; Jiaming Wang, Year: 2025, “Predicting Student Performance Based on Knowledge Characteristics and Learning Ability”, in IEEE Access, Vol: 13, pp. 26105-26115.

CrossRef Google Scholar Publisher Link
[4] Essa Alhazmi; Abdullah Sheneamer, Year: 2023, “Early Predicting of Students Performance in Higher Education”, in IEEE Access, Vol: 11, pp. 27579-27589.

CrossRef Google Scholar Publisher Link
[5] Ziling Chen; Gang Cen; Ying Wei; Zifei Li, Year: 2023, “Student Performance Prediction Approach Based on Educational Data Mining”, in IEEE Access, Vol: 11, pp. 131260-131272.

CrossRef Google Scholar Publisher Link
[6] Jin Li; Shu Li; Yuan Zhao; Longjiang Guo; Fei Hao; Meirui Ren; Keqin Li, Year: 2023, “Predicting dropouts before enrollments in MOOCs: An explainable and selfsupervised model”, IEEE Transactions on Services Computing, Vol: 16, No: 6, pp. 4154–4167.

CrossRef Google Scholar Publisher Link
[7] Nesren Farhah; Adnan Muhammad; Ahmed Abdullah Alqarni; M. Irfan Uddin; Theyazn HH Aldhyani, Year: 2025, “AI-Driven Innovation Using Multimodal and Personalized Adaptive Education for Students With Special Needs”, in IEEE Access, Vol: 13, pp. 69790-69820.

CrossRef Google Scholar Publisher Link
[8] Wei Zhang; Zhongwei Gong; Peihua Luo; Zhixin Li, Year: 2024, “DKVMN-KAPS: Dynamic Key-Value Memory Networks Knowledge Tracing With Students’ Knowledge-Absorption Ability and Problem-Solving Ability”, in IEEE Access, Vol: 12, pp. 55146-55156.

CrossRef Google Scholar Publisher Link
[9] Ike Fibriani; Eko Mulyanto Yuniarno; Ronny Mardiyanto; Mauridhi Hery Purnomo, Year: 2024, “ViTMa: A Novel Hybrid Vision Transformer and Mamba for Kinship Recognition in Indonesian Facial Micro-Expressions”, in IEEE Access, Vol: 12, pp. 164002-164017.

CrossRef Google Scholar Publisher Link
[10] Yu Feng; Minghui Li; Mingjie Chen; Benning Song; Dongjun Yu, Year: 2026, “An Efficient Ensemble Mamba Architecture for IMU Data Processing in the FAST Feed Cabin”, in IEEE Sensors Journal, Vol: 26, No: 7, pp. 10589-10598.

CrossRef Google Scholar Publisher Link
[11] Sukrit Leelaluk; Cheng Tang; Tsubasa Minematsu; Yuta Taniguchi; Fumiya Okubo; Takayoshi Yamashita; Atsushi Shimada, Year: 2024, “Attention-Based Artificial Neural Network for Student Performance Prediction Based on Learning Activities”, in IEEE Access, Vol: 12, pp. 100659-100675.

CrossRef Google Scholar Publisher Link
[12] Yupei Zhang; Rui An; Shuhui Liu; Jiaqi Cui; Xuequn Shang, Year: 2023, “Predicting and Understanding Student Learning Performance Using Multi-Source Sparse Attention Convolutional Neural Networks”, in IEEE Transactions on Big Data, Vol: 9, No: 1, pp. 118-132.

CrossRef Google Scholar Publisher Link
[13] Daozong Sun; Rongxin Luo; Qi Guo; Jiaxing Xie; Hongshan Liu; Shilei Lyu; Xiuyun Xue; Zhen Li; Shuran Song, Year: 2023, “A University Student Performance Prediction Model and Experiment Based on Multi-Feature Fusion and Attention Mechanism”, in IEEE Access, Vol: 11, pp. 112307-112319.

CrossRef Google Scholar Publisher Link
[14] Han Wan; Shiyang Yue; Mengying Li; Xin Luo; Yaofeng Hu; Baoliang Che; Jingyuan Wang, Year: 2026, “Integrating Blended Learning Behaviors via Multimodal Fusion for Student Performance Prediction”, in IEEE Transactions on Learning Technologies, Vol: 19, pp. 87-104.

CrossRef Google Scholar Publisher Link
[15] Yu Liu; Yanchuan Hui; Dongxu Hou; Xiao Liu, Year: 2023, “A Novel Student Achievement Prediction Method Based on Deep Learning and Attention Mechanism”, in IEEE Access, Vol: 11, pp. 87245-87255.

CrossRef Google Scholar Publisher Link

Full Length Article PDF

Article Details