Downloads: 8 | Views: 237 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Research Paper | Computer Science and Information Technology | India | Volume 13 Issue 6, June 2024 | Popularity: 5.2 / 10
Semantic Search - Based Medicine Recommender with LangChain Google Search Agent
Aakash Walavalkar, Jay Ajmera, Kallind Soni
Abstract: In the field of medicine, timely and accurate medication suggestions are critical to ensuring that patients receive top - quality care. This paper leverages the functionality of LangChain agents along with advanced semantic search algorithms to address the challenge of finding relevant medications for specific illnesses. Here, we introduce an innovative architecture, the "Semantic Search - Based Medicine Recommender with LangChain Google Search Agent. " Our system makes use of a large dataset of approximately 23, 000 drugs. Our pipeline includes the steps of data scraping and structuring, NLP sentence vectorization, and 1024 - dimensional embeddings using the SBERT model. Because these embeddings are stored in a Neo4j graph database, effective retrieval is possible due to cosine similarity calculations. Through the use of a chatbot as the interface, the right medicines can be suggested in response to a question by users, and simple interaction is feasible with the user. Our experimentation showcases remarkable improvement over the traditional keyword - based search approach in both the accuracy of suggestions as well as precision and recall. The provided outcomes point out how the integration of LangChain agents with semantic search algorithms will enhance the efficiency and robustness of medication recommender systems.
Keywords: Medicine recommendation system, Medicine Recommender, LangChain Agents, SBERT embeddings, Contextual medicine recommender, Google search agent, LLM Grounding, Neo4j Database, Neo4j Vector Database
Edition: Volume 13 Issue 6, June 2024
Pages: 643 - 646
DOI: https://www.doi.org/10.21275/SR24609111905
Make Sure to Disable the Pop-Up Blocker of Web Browser
