Scribd
Upload
11 views12 pages

Diabetes Prediction Presentation

The document presents a project on diabetes prediction using machine learning, specifically employing a Support Vector Machine (SVM) classifier. It details the methodology, including data preprocessing, model development, and evaluation metrics, achieving a training accuracy of 78.5% and testing accuracy of 76.4%. The study concludes that the SVM model shows potential for diabetes prediction, while highlighting limitations such as the dataset's lack of certain health features and the need for broader applicability testing.

Uploaded by

Atul suryavanshi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
11 views12 pages

Diabetes Prediction Presentation

The document presents a project on diabetes prediction using machine learning, specifically employing a Support Vector Machine (SVM) classifier. It details the methodology, including data preprocessing, model development, and evaluation metrics, achieving a training accuracy of 78.5% and testing accuracy of 76.4%. The study concludes that the SVM model shows potential for diabetes prediction, while highlighting limitations such as the dataset's lack of certain health features and the need for broader applicability testing.

Uploaded by

Atul suryavanshi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 12

GOVERNMENT

ENGINEERING
COLLEGE RAIPUR
(Affiliated to CSVTU Bhilai/Approved by AICTE , New Delhi )
Raipur-492015

“A Machine Learning Approach to


Diabetes Prediction Using Support
Vector Machines
”
GUIDED BY
PRESENTED
MANJEET SONWANI
SIR BY
ATUL
SURYVANSH
I
Agenda
 Introduction
 Dataset Description
 Methodology
 Result
 Web Applications
 Limitations.
Conclusion.
References.
Introduction

Di ab ete s is a c hro nic h ea lth c o nd itio n a ffe c ting m illi on s w orldw id e.


Di ab ete s is a c hro nic h ea lth c o nd itio n a ffe c ting m illi on s w orldw id e.

Ea rly d ete c ti on is c ri tic al for ma na gi ng the d is e as e e ffe c ti v e ly .


Ea rly d ete c ti on is c ri tic al for ma na gi ng the d is e as e e ffe c ti v e ly .

Trad itio na l d iag no s ti c m e tho ds c a n be inv as iv e a nd in ac c es s i ble .


Trad itio na l d iag no s ti c m e tho ds c a n be inv as iv e a nd in ac c es s i ble .

Th is proje c t us es ma c h in e l earnin g (M L ) w ith a n S VM c las s i fie r to pred ic t dia be te s .


Th is proje c t us es ma c h in e l earnin g (M L ) w ith a n S VM c las s i fie r to pred ic t dia be te s .


Dataset Description
Methodology
D ata Prepr oces si ng:
D ata Prepr oces si ng:

- M i ss i ng val ues handl ed us i ng m edi an repl acement.


- M i ss i ng val ues handl ed us i ng m edi an repl acement.

- Features standardi zed w i t h Standar d Sc al er .


- Features standardi zed w i t h Standar d Sc al er .

M odel D evel opment :


M odel D evel opment :

- A l gori thm: Support V ect or M achi ne ( SV M ) w i th a l i near kernel .


- A l gori thm: Support V ect or M achi ne ( SV M ) w i th a l i near kernel .

- Trai n- Tes t Spl i t: 80% trai ni ng, 20% t esti ng.


- Trai n- Tes t Spl i t: 80% trai ni ng, 20% t esti ng.

- M etr i cs: A c curac y, Preci si on, R ecal l , F1- Scor e.


- M etr i cs: A c curac y, Preci si on, R ecal l , F1- Scor e.


Methodology

Data from Kaggle


Methodology
D ata Prepr oces si ng:
D ata Prepr oces si ng:

- M i ss i ng val ues handl ed us i ng m edi an repl acement.


- M i ss i ng val ues handl ed us i ng m edi an repl acement.

- Features standardi zed w i t h Standar d Sc al er .


- Features standardi zed w i t h Standar d Sc al er .

M odel D evel opment :


M odel D evel opment :

- A l gori thm: Support V ect or M achi ne ( SV M ) w i th a l i near kernel .


- A l gori thm: Support V ect or M achi ne ( SV M ) w i th a l i near kernel .

- Trai n- Tes t Spl i t: 80% trai ni ng, 20% t esti ng.


- Trai n- Tes t Spl i t: 80% trai ni ng, 20% t esti ng.

- M etr i cs: A c curac y, Preci si on, R ecal l , F1- Scor e.


- M etr i cs: A c curac y, Preci si on, R ecal l , F1- Scor e.


Results

M o d e lP e rfo rm a n c e :

M o d e lP e rfo rm a n c e :

-T ra in in g A c c u ra c y :7 8 .5 %

-T ra in in g A c c u ra c y :7 8 .5 %

-T e s tin g A c c u ra c y :7 6 .4 %

-T e s tin g A c c u ra c y :7 6 .4 %

E v a lu a tio n M e tric s :

E v a lu a tio n M e tric s :

-P re c is io n :0 .7 7

-P re c is io n :0 .7 7

-R e c a l:0 .7 5

-R e c a l:0 .7 5

-F 1 -S c o re :0 .7 6

-F 1 -S c o re :0 .7 6
Web Application
An interactive web app was developed using Flask:
An interactive web app was developed using Flask:

Users can input health parameters to receive predictions.


Users can input health parameters to receive predictions.

Provides real-time feedback on diabetes likelihood.


Provides real-time feedback on diabetes likelihood.

Designed for accessibility and user-friendliness.


Designed for accessibility and user-friendliness.


Limitations

- Dataset lacks features like lifestyle habits and genetic history.


- Dataset lacks features like lifestyle habits and genetic history.

- Model trained on a single dataset; broader applicability needs testing.


- Model trained on a single dataset; broader applicability needs testing.


Conclusion

Th is s tud y d em on s trate s the po te nti al o f m a c h ine le arn ing in h ea lthc are:


Th is s tud y d em on s trate s the po te nti al o f m a c h ine le arn ing in h ea lthc are:

- S VM a c h iev ed re lia ble a c c ura c y fo r di ab ete s pre dic tio n.


- S VM a c h iev ed re lia ble a c c ura c y fo r di ab ete s pre dic tio n.

- T he w eb ap pl ic a tio n m a ke s th e s y s te m ac c e s s i ble fo r re al -w orl d u s e .


- T he w eb ap pl ic a tio n m a ke s th e s y s te m ac c e s s i ble fo r re al -w orl d u s e .

Fu tu re w o rk w i ll foc u s on im pro v i ng ac c u rac y , inc orpo rati ng di v e rs e da ta , an d s c ali ng the s y s tem .


Fu tu re w o rk w i ll foc u s on im pro v i ng ac c u rac y , inc orpo rati ng di v e rs e da ta , an d s c ali ng the s y s tem .


References
1. Kaggle. PIMA Indians Diabetes Dataset. https://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database .
1. Kaggle. PIMA Indians Diabetes Dataset. https://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database .

2. Cortes, C., & Vapnik, V. (1995). Support-Vector Networks. Machine Learning, 20(3), 273-297.
2. Cortes, C., & Vapnik, V. (1995). Support-Vector Networks. Machine Learning, 20(3), 273-297.

3. Kuhn, M., & Johnson, K. (2013). Applied Predictive Modeling. Springer.


3. Kuhn, M., & Johnson, K. (2013). Applied Predictive Modeling. Springer.

4. Kingma, D. P., & Ba, J. (2015). Adam: A Method for Stochastic Optimization. arXiv preprint arXiv:1412.6980.
4. Kingma, D. P., & Ba, J. (2015). Adam: A Method for Stochastic Optimization. arXiv preprint arXiv:1412.6980.

5. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). " Why Should I Trust You?": Explaining the Predictions of Any Classifier. Proceedings of the 22nd ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.
5. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). " Why Should I Trust You?": Explaining the Predictions of Any Classifier. Proceedings of the 22nd ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy