🌸 Iris Species Classification with Machine Learning

This repository contains an end-to-end Machine Learning classification pipeline built on the classic Iris Flower Dataset. The project explores feature relationships, applies data preprocessing techniques, and benchmarks three popular classification algorithms: Gaussian Naive Bayes, Logistic Regression, and Support Vector Classifier (SVC).

---

📌 Project Overview

• Problem Type: Multi-Class Supervised Classification
• Dataset: Iris Flower Dataset (Sepal Length, Sepal Width, Petal Length, Petal Width)
• Target Classes: Iris-setosa, Iris-versicolor, Iris-virginica
• Models Evaluated: Gaussian Naive Bayes, Logistic Regression, Support Vector Classifier (SVC)
• Evaluation Metrics: Accuracy Score, Confusion Matrix, Classification Report (Precision, Recall, F1-Score)

---

🛠️ Tech Stack & Libraries

• Language: Python 3.14
• Data Handling & Manipulation: pandas, numpy
• Data Visualization: matplotlib, seaborn
• Machine Learning Framework: scikit-learn
• Environment: Jupyter Notebook / PyCharm

---

🔄 End-to-End Workflow & Methodology

1. Data Cleaning & Feature Dropping

• Unnecessary Column Removal: Dropped non-informative identifier columns (e.g., Id) to prevent feature redundancy and noise.

2. Categorical Encoding

• Target Encoding: Converted categorical target labels (Setosa, Versicolor, Virginica) into numerical representations using LabelEncoder.

3. Exploratory Data Analysis (EDA) & Visualization

• Conducted distributions and pairwise feature relationship analyses to identify class separability among species.
• Generated pair plots and feature correlation visualizations to observe non-linear boundaries (particularly petal vs. sepal features).

4. Data Splitting & Feature Scaling

• Train-Test Split: Separated features ($X$) and target ($y$), splitting the data into training and validation sets.
• Feature Scaling: Standardized feature distributions via StandardScaler to optimize distance-based estimators (especially SVC) and linear solvers.

---

🤖 Model Evaluation & Performance Metrics

Each model was trained on the standardized training data and evaluated on the hold-out test set using:

1. Accuracy Score: Overall classification correctness.
2. Confusion Matrix: Heatmap visualization of true vs. predicted species counts.
3. Classification Report: Detailed class-level Precision, Recall, and F1-Scores.

Models Benchmarked

Algorithm	Key Characteristics	Strengths on Iris Dataset
Gaussian Naive Bayes	Probabilistic classifier assuming feature independence	Fast training with strong baseline probability estimation
Logistic Regression	Linear multi-class model using Softmax/OVR	Highly interpretable decision boundaries
Support Vector Classifier (SVC)	Kernel-based decision boundary optimization	Excellent margin separation between flower species

---

📁 Repository Structure

.
├── Iris_Species_Classification.ipynb   # Main notebook containing EDA, preprocessing & modeling
└── README.md                           # Project documentation