🌸 Iris Flower Classification

CodeAlpha Data Science Internship — Task 1

A complete machine learning pipeline that classifies Iris flowers into 3 species using 4 classifiers — achieving up to 100% test accuracy with full EDA, evaluation, and visualization.

📓 View Notebook • 📊 Results • 🚀 How to Run • 📁 Project Structure

---

📌 Project Overview

This project is Task 1 of the CodeAlpha Data Science Internship. The goal is to build a supervised machine learning model that classifies Iris flowers into one of three species — Setosa, Versicolor, or Virginica — based on 4 physical measurements.

The project covers the complete ML pipeline from data exploration to model evaluation and comparison, using 4 different classifiers side-by-side.

---

🎯 Objectives

• ✅ Perform thorough Exploratory Data Analysis (EDA) with visualizations
• ✅ Preprocess data with StandardScaler for optimal model performance
• ✅ Train and compare 4 machine learning classifiers
• ✅ Evaluate models using accuracy, cross-validation, and confusion matrices
• ✅ Identify the best model and extract feature importance insights

---

📂 Dataset

Property	Detail
Source	UCI Machine Learning Repository (via sklearn.datasets)
Samples	150 (50 per class)
Features	4 — Sepal Length, Sepal Width, Petal Length, Petal Width
Target	3 classes — Setosa, Versicolor, Virginica
Missing Values	None
Class Balance	Perfectly balanced (50 samples each)

📋 Feature Description

Feature	Unit	Description
sepal_length	cm	Length of the flower's sepal
sepal_width	cm	Width of the flower's sepal
petal_length	cm	Length of the flower's petal
petal_width	cm	Width of the flower's petal
species	—	Target: Setosa / Versicolor / Virginica

---

🛠️ Tech Stack

Tool	Version	Purpose
Python	3.10+	Core language
Pandas	2.0+	Data manipulation
NumPy	1.24+	Numerical operations
Scikit-learn	1.3+	ML models & evaluation
Matplotlib	3.7+	Visualizations
Seaborn	0.12+	Statistical plots
Jupyter Notebook	—	Development environment

---

🤖 Models Used

#	Model	Key Hyperparameters
1	Logistic Regression	max_iter=300
2	Decision Tree	max_depth=4
3	Random Forest	n_estimators=100, max_depth=5
4	K-Nearest Neighbors (KNN)	n_neighbors=5, metric='euclidean'

All models evaluated with:

• 80/20 Train-Test Split (stratified)
• 5-Fold Stratified Cross-Validation
• StandardScaler normalization applied

---

📊 Results

Rank	Model	Test Accuracy	CV Mean	CV Std
🥇	Random Forest	~97–100%	~96%	±2%
🥈	KNN	~96–100%	~95%	±3%
🥉	Logistic Regression	~96–97%	~95%	±3%
4	Decision Tree	~93–97%	~93%	±4%

💡 Exact values depend on the random seed. Run the notebook to see your results.

---

📈 Visualizations

The project generates 7 professional plots, all saved as .png files:

Plot	File	Description
🔷 Feature Distributions	feature_distributions.png	Histograms by species for all 4 features
🔗 Correlation Heatmap	correlation_heatmap.png	Feature correlation matrix
🌐 Pairplot	pairplot.png	All feature pair combinations colored by species
📦 Boxplots	boxplots.png	Feature spread and outliers per species
🏆 Model Comparison	model_comparison.png	Test vs CV accuracy bar chart for all models
🔢 Confusion Matrices	confusion_matrices.png	4 confusion matrices side by side
🌳 Feature Importance	feature_importance.png	Random Forest feature importance
🗺️ Decision Boundary	decision_boundary.png	RF decision boundary (petal features)

---

💡 Key Insights

1. Petal features dominate — petal_length and petal_width together explain ~95%+ of the class separability
2. Setosa is trivially separable — linearly separable from other species in petal space
3. Versicolor–Virginica boundary is the challenge — slight overlap in feature space
4. Random Forest handles noise best — ensemble averaging reduces boundary errors
5. No overfitting detected — consistent train/CV scores across all models

---

📁 Project Structure

CodeAlpha_IrisFlowerClassification/
│
├── 📓 iris_classification.ipynb    ← Main Jupyter Notebook (full pipeline)
├── 📄 README.md                    ← This file
├── 📋 requirements.txt             ← Python dependencies
│
└── 📊 Generated Plots/
    ├── feature_distributions.png
    ├── correlation_heatmap.png
    ├── pairplot.png
    ├── boxplots.png
    ├── model_comparison.png
    ├── confusion_matrices.png
    ├── feature_importance.png
    └── decision_boundary.png

---

🚀 How to Run

Option 1 — Clone & Run Locally

# 1. Clone the repository
git clone https://github.com/MOHAMMED-ABUZAR317/CodeAlpha_IrisFlowerClassification.git
cd CodeAlpha_IrisFlowerClassification

# 2. Install dependencies
pip install -r requirements.txt

# 3. Launch Jupyter Notebook
jupyter notebook iris_classification.ipynb

Option 2 — Run on Google Colab (No setup needed)

Click the badge below to open directly in Google Colab:

---

📦 Requirements

pandas>=2.0.0
numpy>=1.24.0
scikit-learn>=1.3.0
matplotlib>=3.7.0
seaborn>=0.12.0
jupyter>=1.0.0

Install all at once:

pip install -r requirements.txt

---

📚 What I Learned

• How to perform structured EDA with multiple visualization techniques
• Understanding class separability through pairplots and correlation analysis
• Importance of feature scaling (StandardScaler) for KNN and Logistic Regression
• Comparing models objectively using cross-validation vs test accuracy
• How ensemble methods (Random Forest) outperform single models on real datasets

---

🔗 Connect

Platform	Link
💼 LinkedIn	[(https://www.linkedin.com/in/mohammed-abuzar-9061a1375/)
🐙 GitHub	[(https://github.com/MOHAMMED-ABUZAR317)
🏢 Internship	CodeAlpha

---

🌸 Made with ❤️ during the CodeAlpha Data Science Internship

If you found this project helpful, please give it a ⭐ on GitHub!