Symptom Significance in Diabetes Diagnosis

authors: Gabriela Serrano and JuanCarlos Jimenez
date: May 2, 2022

Description

This repository contains the final project "Symptom Significance in Diabetes Diagnosis". It includes the R Markdown file, dataset and report.  

A variety of techniques are applied, such as cross-validation, hyperparameter tuning, and supervised learning binary classification algorithms to a UCL Machine Learning Repository diabetes dataset in order to predict whether a patient has or not the condition based on a set of symptoms.  

The project will be submitted in fulfillment of the requirements of CSC 597 (Introduction to Statistical Learning with Applications in R) at the University of Miami.

Table of Contents (R Markdown)

Section I: Loading the data

Section II: Data distribution / visualization

Section III: Data preparation

Section IV: Validation of classification models

Logistic Regression

Decision Tree

Random Forest

Naïve Bayes Classifier

Section V: Analysis of variable importance

 

Table of Contents (Report)

I. Introduction

II. Methods: data, techniques and evaluation

III. Results

IV. Discussion and Conclusions