What question did this study set out to answer?

The aim is to develop a recommender system that provides medicine alternatives based on semantic similarities in cases where prescribed drugs are unavailable or unaffordable.

April 1, 2026Iconic Research and Engineering Journals0 citations

A Semantic Similarity–Based Medicine Alternative Recommender System Using Lightweight Transformer Models

SRS. Jothir Ram KMK. Maniraj

Key Points

The aim is to develop a recommender system that provides medicine alternatives based on semantic similarities in cases where prescribed drugs are unavailable or unaffordable.
Utilized Sentence-BERT embeddings for semantic similarity assessment.
Implemented hybrid filtering techniques for effective recommendations.
Developed a scalable full-stack system for integration into practice.
Achieved significant improvements in recommendations compared to lexical baselines.
Generated clinically relevant alternatives for patients and pharmacists.

Abstract

This paper presents a semantic similarity–based medicine alternative recommender system using Sentence-BERT embeddings, hybrid filtering, and a scalable full-stack imple- mentation. The expanded version integrates additional technical sections such as Problem Statement, Motivation, Research Gap, and Algorithmic Workflow, increasing academic rigor while maintaining IEEE format. The system assists pharmacists and patients by generating clinically relevant alternatives when the prescribed drug is unavailable or unaffordable. Experimental evaluation demonstrates large gains over lexical baselines.

Ask AI

Helpful

Bookmark

View Full Paper