Linear Regression on Iris Dataset

This project demonstrates the implementation of Linear Regression using the Iris dataset. Although the dataset is typically used for classification, here we adapt it for regression to predict versicolor features such as petal width and petal length.

• Even though it can be done without using gradient descent by closed-form solution but it would be constrained.
• Also I took it as a thing to do with gradient descent only.

---

🧠 BSoC Attribution

This project was completed during Week 1 (June 16–22, 2025) of the Summer of Machine Learning (SoM) track under BSoC.

---

🚀 Enhancements

• 📊 Hosted on GitHub Pages for interactive visualizations
• 📁 Plotted Plotly Express Plots along with Seaborn.

---

📌 Objective

1. To explore relationships between different features in the Iris dataset and build a linear regression model to predict one feature from another using gradient descent.
2. To share EDA report results via a web page.

---

📊 Dataset

The project uses the Iris dataset from scikit-learn. It includes:

• Sepal length
• Sepal width
• Petal length
• Petal width

Only the Versicolor subset is used for training and evaluation.

---

📈 EDA Report

An automated exploratory data analysis (EDA) report is included in HTML format (Report_flower.html). It contains:

• Summary statistics
• Feature distributions
• Correlation heatmaps
• Outlier analysis Open it in your browser to explore the dataset insights before modeling.

Click here for: EDA Report

---

🛠️ Technologies Used

• Python
• Jupyter Notebook
• NumPy
• Pandas
• Matplotlib Pyplot
• Seaborn
• scikit-learn
• Plotly Express

---

📁 Project Structure

Linear_Regression_on_iris_flowers/
├── Flower_cleaned_fully_fixed.ipynb           🔹 EDA + model notebook
├── LICENSE                                    🔹 MIT License
└── README.md                                  🔹 This file!

---

👤 Author

Anuj Kulkarni - aka - steam-bell-92