discrimant_analysis

Discriminant Analysis

This repository contains R implementations of supervised classification techniques completed as part of a Data Mining course.

Repository Contents

• codermd.Rmd – R Markdown file implementing Logistic Regression, Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), and Naive Bayes classification models.
• README.md – Project documentation.

Methods Implemented

• Logistic Regression
• Linear Discriminant Analysis (LDA)
• Quadratic Discriminant Analysis (QDA)
• Naive Bayes Classification
• Confusion Matrix Evaluation
• Model Performance Comparison

Dataset

The project uses the built-in Iris dataset in R. The Setosa class is removed, and the remaining species are classified using multiple supervised learning algorithms.

Results

The classification models were evaluated using confusion matrices. Logistic Regression achieved the highest accuracy (approximately 98%), while LDA and QDA achieved about 97%, and Naive Bayes achieved approximately 94%.

Tools & Technologies

• R
• R Markdown
• MASS
• e1071
• caret

Skills Demonstrated

• Supervised Machine Learning
• Statistical Classification
• Model Evaluation
• Confusion Matrix Analysis
• Predictive Modeling
• Data Mining