1. An Industry Oriented Mini Project Presentation
On
Various Machine Learning Methods for Diabetes Prediction
By
K.SAI VARUN REDDY(20WJ1A6624)
M.NITHIN(20WJ1A6632)
Under the Esteemed Guidance of
Reddi. Durga Sree
Guru Nanak Institutions Technical Campus(Autonomous)
2. • Abstract
• Objective
• Introduction
• Existing System
• Proposed System
• System Design &Analysis
• System Requirements
• Implementation
• Conclusion
• Thank You
OVERVIEW of the Presentation
3. • Diabetes Mellitus (DM) is the group of diseases where the patient suffers from
higher levels of sugar in blood over a prolonged time. Machine learning classifier
helps to predict the disease based on the condition of the symptom suffered by the
patient. The aim of this project is to compare the performance of the machine learning
tree classifiers in predicting Diabetes Mellitus. Machine learning tree classifiers such
as Random Forest, C4.5, Random Tree, and Logistic Model Tree (LMT) were
analyzed based on their accuracy and True Positive Rate(TPR). In this analysis of
predicting DM, Logistic Model Tree classifier achieved higher accuracy than other
classifiers under study.
ABSTRACT
4. • The objective of this mini project is to systematically evaluate and
compare the performance of different machine learning tree classifiers,
in predicting Diabetes Mellitus (DM). The focus is on assessing the
accuracy and True Positive Rate (TPR) of each classifier. By
conducting a thorough analysis, the project aims to identify the most
effective machine learning tree classifier for accurate and efficient
prediction of Diabetes Mellitus.
OBJECTIVE
5. • Diabetes Mellitus (DM) stands as a prevalent metabolic disorder
characterized by elevated blood sugar levels, posing substantial health risks
if left unmanaged. Given its global impact, timely prediction and
intervention become paramount. Our project takes strides in addressing this
challenge by employing various machine learning methods for predictive
analysis. By using various machine learning algorithms , we aim to
enhance the accuracy of diabetes prediction. This endeavor helps identify
individuals at risk and initiate timely interventions, ultimately contributing
to improved management and outcomes in Diabetes Mellitus.
INTRODUCTION
6. • There are several possibilities of diabetes mellitus such as obesity, irregular food
intake timings etc, Primary prevention of diabetes mellitus avoids higher risks at
later stages, such as regular tests of blood sugar content after and before fasting.
• People with diabetes are also subjected to chronic kidney disease, the diabetes
mellitus are also called as silent killers, where the symptoms are undetectable, the
symptoms includes being thirsty often, abnormal loss of weight, feeling very
hungry, and have a blurry eyesight.
EXISTING SYSTEM
7. EXISTING TECHNIQUE
• Traditional Data Classification Algorithms.
TECHNIQUE DEFINITION
• Data classification is the process of organizing data into categories for its most effective and
efficient use. A well-planned data classification system makes essential data easy to find and
retrieve.
• This can be of particular importance for risk management, legal discovery, and compliance.
EXISTING SYSTEM
8. DRAWBACKS
• Memory or computational power limitations.
• Static nature hampers adaptability : lack of adaptability to dynamic changes in patient data,
potentially impacting prediction accuracy.
• Very high rate.
EXISTING SYSTEM
9. • This section briefly introduces the five desired classifiers for predicting diabetes
mellitus under comparison analysis. Random forest are the ensemble learning
approaches which constructs many decision trees at the time of training and
predicts the output by analyzing individual trees. It is more efficient than
decision tree.
• C4.5 uses the extension of the ID3 algorithm for generating decision trees for
both nominal and numerical values. It is a statistical classifier which classifies as
classes used in the diagnosis of certain patterns. A Random tree is a graphical
representation of decisions with their respective consequences.
PROPOSED SYSTEM
10. PROPOSED TECHNIQUE
• Logistic Model Trees
TECHNIQUE DEFINITION
• Logistic model tree (LMT) is a classification model with an associated supervised training
algorithm that combines logistic regression (LR) and decision tree learning.
• A decision tree that has linear regression models at its leaves to provide a piecewise linear
regression model.
ADVANTAGES
• Performs well with large dataset.
PROPOSED SYSTEM
12. • Hardware Requirements
• PROCESSOR : DUAL CORE 2 DUO
• RAM : 2GB DD RAM
• HARD DISK : 250 GB
• Software Requirements
• FRONT END : PYTHON
• OPERATING SYSTEM : WINDOWS 7(or above)
• IDE : SPYDER3,JUPYTER
SYSTEM REQUIREMENTS
24. • Machine learning tree classifiers were analyzed for predicting diabetes mellitus in this
study. The accuracy of these classifies were calculated based on the True Positive, True
Negative, False Positive, and False Negative of the classifiers. The classifier with the
highest accuracy and true positive rate is identified.
CONCLUSION