Health Prediction System
Using Machine Learning
& Python
Presented By
Team -5
CONTENTS
INTRODUCTION
WORK FLOW
WORKING
RESULTS
CONCLUSION
2
INTRODUCTION
Identifying diabetes, heart disease, and Parkinson’s is tough due to unclear
symptoms.
Variability in data and human subjectivity leads to misclassification and
delayed diagnosis.
ML boosts diagnostic accuracy by spotting patterns in patient information.
It predicts diseases early, aiding in targeted care for patients.
Using ML Early detection becomes simpler, facilitating prompt treatment.
ML provides objective, data-driven diagnostic insights, supporting better
medical decision-making.
3
WORK FLOW
Data Collection ---> Data Preprocessing ---> Model Building ---> Model Evaluation
---> Save Model ---> Deploying the model ---> Displaying the results
4
WORKING
DATA COLLECTION
We collect the datasets from Kaggle
The sizes of heart disease dataset is 768*9
The sizes of Parkinson's disease dataset is 195*24
The sizes of diabetes disease dataset is 303*14
DATA PREPROCESSING
We check for null values in dataset and eliminate them using appropriate techniques
We use one hot encoding for converting categorical data to numerical data(0 healthy ,1
unhealthy)
We select important features by correlation using heatmap
5
Model Building
We consider Logistic Regression for building heart diseases predictive system
We consider SVM for building Parkinson's disease predictive system
We consider SVM for building diabetes disease predictive system
Performance Measure
We consider Accuracy as performance measure
The results on train and test set by each model is as shown in the following graph
6
88
87.17 87.17
86
85.12
84
81.96
82
80
78.33
78
77.27
76
74
72
Diabetes Parkinsons Heart
Train Accuracy Test Accuracy
7
DEPLOYING THE MODEL
Streamlit streamlines web app development, minimizing code for
accessible deployment at all skill levels.
Seamless integration with machine learning models in Streamlit
allows for quick deployment and interaction.
Streamlit offers easy-to-use widgets, simplifying user input for
dynamic and user-friendly interfaces.
So we used streamlit for our model deployment and results of stremlit
interface is show below
8
9
10
11
RESULTS
12
13
14
CONCLUSION
The Multiple Disease Prediction System employs Streamlit and machine learning
models to offer a user-friendly interface for predicting diabetes, heart disease, and
Parkinson's disease. With efficient deployment and adaptable design, it provides
accurate health predictions, making it a valuable tool for quick and accessible
disease diagnosis.
15
PROJECT GITHUB LINK
https://github.com/19PA1A0465/Disease-predictive-system
16
Thank
17