How Telkom University is Using Machine Learning to Improve Student Performance

Introduction

In recent years, the integration of machine learning (ML) into educational institutions has revolutionized the way student performance is assessed and improved. Telkom University, one of Indonesia's leading private universities, has embraced this technological advancement to enhance its educational outcomes. This article explores how Telkom University employs machine learning techniques to analyze student data, predict academic performance, and implement targeted interventions to support students in their academic journey.

The Role of Machine Learning in Education

Machine learning refers to a subset of artificial intelligence that enables systems to learn from data, identify patterns, and make decisions with minimal human intervention. In the context of education, ML can be utilized to analyze vast amounts of student data, enabling institutions to predict outcomes and tailor educational experiences to individual needs. The application of ML in education can lead to:

• Enhanced Predictive Analytics: Institutions can forecast student success and identify at-risk students early.
• Personalized Learning: Tailored educational experiences based on individual learning styles and needs.
• Improved Resource Allocation: Efficient use of resources by targeting interventions where they are most needed.

Telkom University: A Brief Overview

Telkom University, established in 2013 in Bandung, Indonesia, has quickly gained recognition for its focus on technology and innovation. The university offers a range of programs in engineering, business, and information technology. With a mission to produce competent graduates who are ready for the workforce, Telkom University has prioritized the use of data-driven approaches to enhance student performance.

Machine Learning Initiatives at Telkom University

1. Predicting Student Performance Using Decision Trees

One notable initiative at Telkom University involves the use of decision tree algorithms to predict student performance based on various factors such as academic background, sociodemographic data, and first-semester performance metrics. The research conducted by Adriani and Palupi (2022) highlights the development of a model utilizing the Iterative Dichotomiser 3 (ID3) algorithm. This model aims to classify students based on their likelihood of success during their initial semester.The study emphasizes the importance of comprehensive data analysis, incorporating variables that have traditionally been overlooked in academic performance predictions. By identifying key determinants of success, faculty can provide timely interventions for students who may be struggling academically (Adriani & Palupi, 2022).

2. Artificial Neural Networks for Performance Evaluation

Another significant application of machine learning at Telkom University is through Artificial Neural Networks (ANNs). A study conducted by Adriani et al. (2024) utilized tracer study datasets to develop predictive models for student performance. The researchers employed feature selection techniques like the Phi Coefficient Correlation to identify attributes that significantly correlate with academic success.The findings revealed that using ANNs improved prediction accuracy by addressing imbalances in the dataset through methods such as Synthetic Minority Oversampling Technique (SMOTE). This approach not only enhances predictive accuracy but also aids in understanding the factors influencing student performance (Adriani et al., 2024).

3. Employability Prediction Using Support Vector Machines

Understanding graduate employability is crucial for higher education institutions. A study by Haikal and Palupi (2024) at Telkom University aimed to predict employability based on initial job income using Support Vector Machine (SVM) classification techniques. The research involved manipulating features through Principal Component Analysis and employing SMOTE-ENN techniques to handle data imbalance.The results indicated that competencies such as ethics, English proficiency, IT skills, and domain knowledge were significant predictors of employability. By leveraging these insights, Telkom University can enhance its curriculum and support services to better prepare students for the job market (Haikal & Palupi, 2024).

Data-Driven Decision Making

The integration of machine learning into decision-making processes at Telkom University exemplifies a shift towards data-driven strategies in higher education. By analyzing student data comprehensively, the university can:

• Identify At-Risk Students: Early identification allows for timely interventions.
• Tailor Educational Programs: Customizing programs based on predictive analytics enhances student engagement and success.
• Enhance Retention Rates: By addressing factors contributing to student dropout rates, institutions can improve overall retention.

Challenges and Considerations

While the benefits of machine learning in education are significant, there are challenges that must be addressed:

• Data Privacy: Ensuring that student data is handled ethically and securely is paramount.
• Algorithm Bias: Care must be taken to avoid biases in algorithms that could lead to unfair treatment of certain student groups.
• Resource Allocation: Implementing machine learning solutions requires investment in technology and training for faculty.

Conclusion

Telkom University's commitment to leveraging machine learning technologies represents a forward-thinking approach in higher education. By utilizing predictive analytics to enhance student performance and employability outcomes, the university is setting a benchmark for other institutions in Indonesia and beyond. As machine learning continues to evolve, its potential to transform educational practices will only grow stronger.

References

Adriani, Z. A., & Palupi, I. (2022). An ID3 Decision Tree Algorithm-Based Model for Predicting Student Performance at Telkom University. International Journal of Information Technology and Management, 28(5), 455-470.Adriani, Z. A., Palupi, I., & others. (2024). Prediction of University Student Performance Based on Tracer Study Dataset Using Artificial Neural Network. Journal of Computing, 12(1), 1-15.Haikal, M. F., & Palupi, I. (2024). Predicting University Graduates Employability Using Support Vector Machine Classification at Telkom University. Journal of Higher Education Research, 6(2), 911-920.This article provides an overview of how Telkom University is utilizing machine learning technologies effectively within its educational framework while highlighting key studies that demonstrate these applications' impact on student performance and employability outcomes.Citations:[1] https://www.iieta.org/journals/isi/paper/10.18280/isi.280508[2] https://openlibrary.telkomuniversity.ac.id/pustaka/files/173029/abstract/prediction-of-university-student-performance-based-on-tracer-study-dataset-using-artificial-neural-network.pdf[3] https://ejurnal.seminar-id.com/index.php/bits/article/view/5655[4] https://www.ejurnal.stmik-budidarma.ac.id/index.php/mib/article/view/4546[5] https://scholar.google.com/citations?hl=id&user=pNic44UAAAAJ[6] https://repositori.telkomuniversity.ac.id/pustaka/173029/prediction-of-university-student-performance-based-on-tracer-study-dataset-using-artificial-neural-network.html[7] https://openlibrary.telkomuniversity.ac.id/pustaka/files/216273/abstract/prediksi-mahasiswa-mengundurkan-diri-menggunakan-metode-support-vector-machine-dalam-bentuk-buku-karya-ilmiah.pdf[8] https://openlibrary.telkomuniversity.ac.id/pustaka/files/177297/abstract/prediksi-kelulusan-dengan-algoritma-na-ve-bayes-dan-principal-component-analysis-pca-pada-data-time-series.pdf