Team 04 Final Review

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 17

PROJECT PRESENTATION

Team No: 04

Review No: Final Review

Project Title: Drug Recommendation System using ML and NLP

Date: 26-04-2024

1
KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India
DRUG RECOMMENDATION SYSTEM

Arjun Balaji - 20AD005


Jaisri V - 20AD018
Mithra S - 20AD026

Project Guide :

Mrs.Punidha

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


2
PROBLEM STATEMENT

• With huge amount of medical and research data available, it can be


challenging for Healthcare professionals to stay up-to-date with the latest
treatments and medications.

• AI-powered drug recommendation systems can provide valuable assistance to


healthcare professionals in making more informed decisions regarding the best
course of treatment for their patients.

• Drug Reviews dataset available on UCI ML repository.

• Many medical conditions being available, our current scope is to recommend drugs
only 4 conditions (Acne, Anxiety, Depression, Birth Control
KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India
3
BACKGROUND STUDY

• This AI system leverages the power of ML to analyze the Patient data including Drug
reviews and other features like Health Condition, Drug-Taken, Support Count to
identify potential issues faced and recommend alternative drug.

• It utilizes NLP techniques to extract meaningful insights from feature(Review),


ultimately enabling more informed decision-making for healthcare professionals.

• This system helps in reducing the risk of medication errors.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


4
LITERATURE SURVEY
[1]The Drug Recommender System uses machine learning for sentiment analysis,
helping patients and medical professionals make informed decisions on drug
prescriptions. By analyzing user sentiments in text data, the system provides insights on
medication effectiveness and patient satisfaction. Combining sentiment analysis with
drug recommendation algorithms, it aims to optimize medicine selection, improve
outcomes, and reduce adverse reactions. The approach has the potential to
revolutionize healthcare by offering tools for informed decisions. Achieves an accuracy
of 94%.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 5
LITERATURE SURVEY
[2]It discusses the challenges faced in the healthcare system due to the COVID-19
pandemic, including shortages of medical professionals, equipment, and medicines. This
has led to individuals self-medicating without proper consultation, worsening their health
conditions. The paper proposes a solution in the form of a drug recommender system
using machine learning. The system utilizes patient reviews and various vectorization
processes to predict sentiments and recommend suitable drugs for specific diseases. The
evaluation of predicted sentiments shows that the LinearSVC classifier with TF-IDF
vectorization performs the best with 93% accuracy, aiming to alleviate the burden on
specialists in the medical field.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 6
LITERATURE SURVEY

[3]The Drug Recommender System uses machine learning for sentiment analysis, helping
patients and medical professionals make informed decisions on drug prescriptions. By
analyzing user sentiments in text data, the system provides insights on medication
effectiveness and patient satisfaction. Combining sentiment analysis with drug
recommendation algorithms, it aims to optimize medicine selection, improve outcomes,
and reduce adverse reactions. The approach has the potential to revolutionize healthcare
by offering tools for informed decisions.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 7
LITERATURE SURVEY
[4] TF-IDF vectorization is employed in drug recommendation systems to convert textual
data, such as patient records, into numerical representations. This involves calculating
TF-IDF scores for terms in each document, where TF measures term frequency, and IDF
measures term importance across all documents. The resulting vectors represent
documents, and drug recommendations are made based on the similarity between
these vectors. While TF-IDF is valuable, advanced techniques like machine learning and
deep learning are often used for improved accuracy. Integrating domain expertise and
addressing ethical considerations are crucial in developing effective drug
recommendation systems.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 8
LITERATURE SURVEY

[5] The article uses LDA topic modeling to analyze 136,630 online pharmacy reviews from
JD Pharmacy and J1.COM. Key findings reveal logistics as the most significant satisfaction
factor (38.5% of reviews), followed by drug prices (25.5%). Positive opinions exist for both
pharmacies, but concerns about logistics and drug prices are noted. The primary
recommendation is for online pharmacies to prioritize improving logistics, possibly
through self-built systems, and enhance consumer engagement through effective
marketing strategies. Improving employee service attitudes is also emphasized for better
customer service.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 9
ARCHITECTURE DESIGN

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 10
PROPOSED SOLUTION
The two phases involved in Drug Recommendation System

• Data Transformation using NLP


Involves preprocessing of patient data ,vectorizing it using Tf-idf Vectorizer
and making it model ready.

• Drug Recommendation using ML


Try out with different Classifier models and chosen the best performance
Classifier model based on Evaluation metrics

Tech Stack Used


• Python
• Pandas
• NumPy
• sklearn Library KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India
11
Model Performance
Based on the performance results, we chose Passive Aggressive Classifier for Drug
recommendation with 95% train accuracy .

Algorithms Accuracy

Passive Aggressive Classifier 0.957546

Multinomial Naive Bayes 0.666066

Random Forest 0.518740

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 12
EXPECTED OUTCOMES

• AI based Drug Recommendation system being trained on Huge Patient data


and accurately identify the Top 3 alternative drugs for the review provided as an
input.

• It ranks medications based on positive ratings and useful count ensuring


better user happiness and efficacy.

• The system also considers negative reviews to prevent prescribing


medicines with unfavorable ratings, preventing unfavorable consequences or
patient dissatisfaction.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


13
Project User Interface

Review Entered

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 14
AI based Drug Recommender

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 15
REFERENCES

[1] Wisdom ML (2023 , March 30 Drug Recommendation System Using NLP and
Machine Learning Approach in Python [online]
https://wisdomml.in/drug-recommendation-system-using-nlp-and-machine-learning-appro
ach-in-python/
[2] Solan (April 2022) Drug Recommendation System based on Sentiment Analysis
of Drug Reviews using Machine Learning https://arxiv.org/pdf/2104.01113v2.pdf
[3] IEEE Sinchana N M (2023) Model-based Filtering Techniques for Recommendation
Systems in Healthcare Domain https://ieeexplore.ieee.org/document/10290568
[4] IEEE (Jan 2020)-Realizing an Efficient IoMT-Assisted Patient Diet Recommendation
System Through Machine Learning Model .
[5] Springer(17 Aug 2020) - Consumers’ satisfaction factors mining and sentiment analysis
of B2C online pharmacy reviews.
KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India
16
Thank you.

KPR Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India


* 17

You might also like