AI

Efficient Prediction of At-Risk University Students Using Reduced Training Vector-Based SVM (RTV-SVM)

3 Mins read

Predicting At-Risk University Students Using a Machine Learning Algorithm:
University education plays a critical role in societal progress, making it essential for students to succeed in their courses and graduate on time. However, many students face academic challenges that lead to course failure, depression, or withdrawal, increasing the faculty workload and the financial strain on institutions. The study presents an RTV-SVM designed to predict at-risk and marginal students while reducing computational costs in response to this issue. The RTV-SVM eliminates redundant training vectors, thus decreasing training time without compromising the accuracy of the support vectors. A study involving 32,593 students across seven courses showed the RTV-SVM reduced training vectors by 59.7% while maintaining classification accuracy. The model achieved an accuracy rate of 92.2-93.8% in identifying at-risk students and 91.3-93.5% for marginal students.

The RTV-SVM leverages support vector machine (SVM) techniques, commonly applied in fields like imaging, bioinformatics, and energy management, to enhance prediction in the education sector. Traditional SVM models often struggle with large datasets, so this paper focuses on improving efficiency while maintaining accuracy. The RTV-SVM addresses computational challenges and enhances the prediction process by reducing the number of training vectors used in SVM without affecting the support vectors. This method offers a promising solution to help universities identify at-risk students early, mitigating academic failure and its associated social and economic consequences.

Challenges and Approaches in Learning Analytics for At-Risk Students:
Learning analytics uses data processing, predictive models, and educational data mining to support meaningful interventions, such as identifying at-risk students. Custom models tailored to specific learner needs can enhance educational outcomes. Traditional methods like replacing exams with attendance are ineffective, leading to lowered academic standards. Predictive models, including random forest, SVM, and decision trees, address this issue by forecasting student failure and dropout risks. Key challenges in learning analytics involve handling big data, collecting sufficient and relevant data, ensuring privacy and security, and choosing the most effective machine learning algorithms.

RTV-SVM Methodology for Optimized SVM Classification:
The RTV-SVM methodology consists of four steps: defining inputs, tier-1 elimination using multivariate normal distribution, tier-2 elimination via vector transformation, and building an SVM classifier using SMO. The process begins by defining training feature vectors and eliminating redundant vectors based on their probability distribution. Tier 2 reduces vectors further by projecting them onto class centers. The remaining vectors are used to build the SVM classifier through SMO, optimizing the decision boundary. This approach aims to enhance classification efficiency by minimizing the number of training vectors while preserving accuracy.

Predicting At-Risk University Students with RTV-SVM:
The RTV-SVM methodology was applied to the Open University Learning Analytics (OULA) dataset to predict at-risk students. The study evaluated four scenarios: no reduction, tier-1 elimination, tier-2 elimination, and both tiers combined. The classifier was assessed using metrics such as training vector reduction, training and testing time, sensitivity, specificity, and overall accuracy. Results showed that tier-1 and tier-2 eliminations significantly reduced training vectors without sacrificing accuracy. In multi-class classification (Pass, Marginal, Fail), the RTV-SVM maintained good performance, with accuracy exceeding 91% across all scenarios.

Performance Comparison Between RTV-SVM and Related Methods:
The RTV-SVM model demonstrated superior performance in predicting at-risk students, particularly those likely to fail. It achieved higher accuracy than other methods, with the benefit of identifying students who may achieve marginal results. This ability to detect marginal students is significant, as they are more prone to failure. Additionally, the RTV-SVM model could outperform models designed for more complex predictions, such as student dropouts or graduation delays. Its accuracy and efficiency make it a strong tool for predicting student outcomes.


Check out the Report. All credit for this research goes to the researchers of this project. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. If you like our work, you will love our newsletter..

Don’t Forget to join our 50k+ ML SubReddit

⏩ ⏩ FREE AI WEBINAR: ‘SAM 2 for Video: How to Fine-tune On Your Data’ (Wed, Sep 25, 4:00 AM – 4:45 AM EST)


Sana Hassan, a consulting intern at Marktechpost and dual-degree student at IIT Madras, is passionate about applying technology and AI to address real-world challenges. With a keen interest in solving practical problems, he brings a fresh perspective to the intersection of AI and real-life solutions.



Source link

Related posts
AI

Salesforce AI Research Proposes a Novel Threat Model: Building Secure LLM Applications Against Prompt Leakage Attacks

6 Mins read
Large Language Models (LLMs) have gained significant attention in recent years, but they face a critical security challenge known as prompt leakage….
AI

Microsoft’s Dynamic Few-Shot Prompting Redefines NLP Efficiency: A Comprehensive Look into Azure OpenAI’s Advanced Model Optimization Techniques

3 Mins read
Microsoft’s approach to leveraging dynamic few-shot prompts with Azure OpenAI offers an innovative technique that optimizes the application of few-shot learning by…
AI

Codeium vs. Tabnine: Comparison of Key Features and Benefits

2 Mins read
When comparing Codeium and Tabnine, both tools offer powerful AI-based code assistance, but they differ significantly in features, security, pricing, and target…

 

 

Leave a Reply

Your email address will not be published. Required fields are marked *