A HYBRID AI-BASED COURSE RECOMMENDATION MODEL FOR DOST USING SENTIMENT-AWARE LEARNING

Sep 22, 2026

The work addresses course-selection complexity created by diverse interdisciplinary undergraduate programs and the limitations of centralized admission platforms that typically lack intelligent academic guidance. It focuses on the Degree Online Services Telangana (DOST) ecosystem, where current processes emphasize administrative automation rather than personalized recommendations. The study proposes a hybrid AI model combining collaborative filtering, latent factor learning, and sentiment-aware analysis. It integrates performance indicators, contextual attributes, peer influence patterns, and qualitative feedback representations, using mathematical formulations for similarity, fusion, and optimization. Validation results from survey observations support the need for recommendation support in centralized environments.

LanguageEnglish

CategoryExam

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Index Terms

Recommender Systems

DOST

Sentiment Analysis

Collaborative Filtering

Educational Data Mining

Machine Learning

1. INTRODUCTION

DOST platform and current limitations

Advances in recommender systems for education

Sentiment-aware learning with NLP and BERT

A HYBRID AI-BASED COURSE RECOMMENDATION MODEL FOR DOST USING SENTIMENT-AWARE LEARNING Dr. Mohd Thousif Ahemad Lecturer, Nagarjuna Government College (Autonomous), Nalgonda, Telangana, India Submission Date: 12-05-2026 , Review Date: 21-05-2026 , Acceptance Date: 29-05-2026 , Publication Date: 31-05-2026 ABSTRACT The rapid expansion of interdisciplinary undergraduate programs has increased the complexity of academic decision-making among higher education aspirants. Although centralized admission platforms simplify institutional admission procedures, they often lack intelligent academic guidance systems capable of assisting students in selecting suitable academic programs aligned with their abilities, interests, and career goals. The Degree Online Services Telangana (DOST) platform provides a unified admission environment for undergraduate programs in Telangana State. However, the present admission process mainly focuses on administrative automation and does not support personalized course recommendation. This study presents a hybrid artificial intelligence-based recommendation model for undergraduate course selection within the DOST ecosystem. The proposed model combines collaborative filtering, latent factor learning, and sentiment-aware analysis to generate personalized course recommendations. Academic performance indicators, contextual attributes, peer influence patterns, and qualitative feedback representations are integrated to estimate recommendation suitability scores. Mathematical formulations for similarity computation, latent interaction learning, recommendation fusion, and optimization are incorporated to establish a theoretically grounded recommendation architecture. A preliminary validation study based on survey observations highlights the necessity of intelligent recommendation support mechanisms within centralized admission environments. The proposed model provides a theoretically grounded approach for integrating academic performance, contextual attributes, and sentiment-aware learning into undergraduate course recommendation. The study contributes a domainspecific recommendation architecture for the DOST ecosystem and offers a foundation for subsequent empirical evaluation and large-scale deployment. | IJCRT26A5073 | International Journal of Creative Research Thoughts (IJCRT) [www.ijcrt.org](www.ijcrt.org) | j506 | | --- | --- | --- | Index Terms: Recommender Systems, DOST, Sentiment Analysis, Collaborative Filtering, Educational Data Mining, Machine Learning 1. INTRODUCTION The growing diversity of undergraduate academic programs has significantly increased the complexity of course selection among higher education aspirants. Emerging disciplines such as artificial intelligence, data science, cybersecurity, and business analytics have transformed conventional undergraduate education into a multidimensional academic environment [1] . While these developments expand educational opportunities, they also create uncertainty among students who often lack sufficient awareness regarding academic suitability, employability trends, and long-term career implications. To simplify undergraduate admissions, the Government of Telangana introduced the Degree Online Services Telangana (DOST) platform, which centralizes the admission process across affiliated colleges and universities. The platform successfully reduces administrative complexity and streamlines seat allocation procedures. However, the current admission mechanism primarily focuses on registration and option-entry activities and does not provide intelligent academic guidance during course selection. As a result, students frequently depend on peer influence, incomplete information, or social trends while selecting academic programs, which may lead to inappropriate academic choices and reduced career alignment. Recent advancements in recommender systems have demonstrated their effectiveness in supporting personalized decision-m

FAQs

What problem does the proposed model address in the DOST admission process?

It addresses the lack of intelligent, personalized course recommendation in a platform that mainly focuses on administrative automation during course selection.

What components are combined in the hybrid recommendation model?

The model integrates collaborative filtering, latent factor learning, and sentiment-aware analysis, along with performance indicators, contextual attributes, peer influence patterns, and qualitative feedback representations.

How does sentiment-aware learning contribute to recommendations?

By using natural language processing to extract semantic information from textual data, it incorporates real-world opinions and satisfaction signals into course recommendation generation.

Sep 22, 2026

Uploaded by Lucas Vance