Diabetes Prediciton Model

You might also like

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 23

(An Autonomous Institute Affiliated to Savitribai Phule Pune University)

Department of Artificial Intelligence

Mini Project Title


Diabetes Prediction
Group members

1) Vicky Jagdale
2) Varad Wagh
3) Vaibhav Kale

Name of Guide: Prof. Sharad Adsure and Prof. Ganesh Kadam


Academic year: 2021-2022

1
Index
Class Page No
DOMAIN INTRODUCTION 4

MOTIVATION 5

INTRODUCTION 6

NEED OF THE PROJECT 8

PROBLEM DEFINATION 9

EXISTING SYSTEM 10
METHODOLOGY
PROPOSED SYSTEM FLOW 11

ALGORITHM 12

OUTCOME 13

2
PROJECT SCOPE 14

ADVANTAGES 15

DISADVANTAGES 16

HARDWARE AND SOFTWARE REQUIREMENTS 17

END USERS 18

COMPARISON RESUTS 19

APPLICATIONS 20

CONCLUSION 21

REFERENCES 22

3
DOMAIN INTRODUCTION

 Domain Name: Deep learning

Deep learning is a subset of machine learning in artificial intelligence that has


networks capable of learning unsupervised from data that is unstructured . Also
known as deep neural learning or deep neural network.

Deep learning is an AI function that mimics the workings of the human brain
in processing data for use in detecting objects, recognizing speech, translating
languages, and making decisions.

Examples: Siri, Alexa

4
MOTIVATION

• According to a report of WHO, about 463 million people in the


world were affected by diabetes in 2019 making it one of a most
significant lethal chronic disease. Early diagnosed of diabetes plays a
very important role in the treatment of the disease, however, the long
asymptomatic phase makes it difficult to diagnose the said disease.
• According to International Diabetes Federation (IDF), about 232
million people (50% of the total population) remain undiagnosed of
the disease

5
INTRODUCTION

We design this project for Doctor. In this project first we collect the
big data. Then we clean the data for machine to predication the accurate
output
By using some algorithm we create the diabetes prediction model
After that when the user gives 8 attributes to the model and model
predict that patient have diabetes or not
It is very simple model to predict the patient is diabetes or not

6
NEED OF THE PROJECT

The need of this project is that it help the patient to know patient is
diabetes or not
Todays world not have time to go to doctor and give some blood and wait
for one or two day to result
Now a days technologic is very increase by use this technology we
create model for diabetes prediction.
For todays fasterer word and save the time we create the model patient
give some attributes to model and model predit the patient is diabetes or
not

7
Existing System

Give some attributes like Age, blood pressure, BMI, Glucose,


Skin thickness etc

Model predit patient have diadetes or not

8
PROBLEM DEFINATION

• Build the application that telling the patient is diabetes or not


by giving some attributes.

9
EXISTING SYSTEM METHODOLOGY

10
PROPOSED SYSTEM FLOW

Diabites dataset

11
ALGORITHM

1) Principal Component Analysis


2) Linear Discriminant Analysis
3) Random Forest
4) Logistic Regression
5) Convolutional Neural Network

12
OUTCOME

Give some attributes like Age, blood pressure, BMI,


Glucose,
Skin thickness etc
Get that patient is diabetes or not

13
PROJECT SCOPE

• The scope of the project is to build a system which


recognition the patient is diabetes or not by using machine
learing

14
ADVANTAGES

• 1. By give some input model predict the diabetes.


• 2. By using this mode we save the time as well as money.
• 3 .By using this mode we find the diabetes.

15
DISADVANTAGES

• 1. If user enter worng inputs then model predict the worng outputs
• 2. (normaly age range is 0 to 100) If user enter age = 200 then model we
predict the worng output
• For better output must have enter correct inputs

Department of AI & Data Science 16


HARDWARE AND SOFTWARE
REQUIREMENTS

Hardware Specification:

1) System type: 64-bit or 32-bit Operating System


2) Processor :Intel(R) Core(TM)i3-5005U CPU 2.00GHz
3) Installed Memory(RAM):4.00GB

Software Specification:

1) Operating System: Windows 7 , 10 or 11


2) IDE: Jupyter Notebook
3) Database: MySQL

17
END USERS

The end user of this project is the patient and doctor


By entering some attributes model predicts the patient is diabetes or

not.

18
COMPARISON RESUTS

• The performance of the given machine learning techniques shows us


the prediction result where we recall the algorithms used as Random
Forest, KNearest Neighbor, Decision Tree, Logistic Regression. Here
we see that logistic regression classifier shows the better results than
other classifier to predict diabetes.

• According to the figure logistic regression shows 78% accuracy on


this dataset, which is greater than all other machine learning
techniques. This it shows that logistic regression performs well on
this given medical dataset for predicting diabetes mellitus.

19
APPLICATIONS

1)We can also use this application for the other disease.

2) We can also use this application for multiple labs and hospital.

3) We can also use the same application for library.

20
CONCLUSION

• In this experiment we have analysed the early prediction of diabetes


by taking all the related factors in its tests and implementing using
machine learning techniques by extracting knowledge from our real
health care medical dataset to predict diabetic patients thus we have
done.
• our experiment using some various machine learning algorithms
namely Random Forest, K-Nearest Neighbour, Decision Tree,
Logistic Regression on Indian datasets to predict diabetes.

21
REFERENCES

•J.S. Skyler, G.L. Bakris, E. Bonifacio, T. Darsow, R.H. Eckel, L. Groop, et al.Differentiation of diabetes


by pathophysiology, natural history, and prognosis

•N. Cho, J. Shaw, S. Karuranga, Y. Huang, J. da Rocha Fernandes, A. Ohlrogge, et al.IDF Diabetes Atlas:


global estimates of diabetes prevalence for 2017 and projections for 2045

•Unreliable hemoglobin A1C (HBA1C) in a patient with new onset diabetes after transplant (nodat)
•Endocr Pract, 24 (2018), pp. 43-44

22
23

You might also like