🧠 ML Basics & Optimization — Task 2

This project demonstrates the fundamental machine learning workflow using the Iris dataset. It covers building a baseline model, applying basic optimizations, and evaluating performance improvements — all in a clean, reproducible setup.

📁 Project Structure ML Basics & Optimization/ ├─ data/ # dataset folder (empty for now, using sklearn built-in) ├─ outputs/ # model outputs, plots, and reports │ ├─ classification_report.txt │ ├─ confusion_matrix.png │ └─ task2_iris_best_model.joblib ├─ src/ │ └─ task2_iris.py # main script for Task 2 ├─ requirements.txt # dependencies └─ venv/ # virtual environment (not uploaded)

🎯 Objective

To:

Build a baseline ML model on the Iris dataset.

Apply simple optimization techniques such as feature scaling and hyperparameter tuning.

Compare results and visualize model performance.

🧩 Dataset

Name: Iris Dataset (built-in from scikit-learn)

Samples: 150

Features:

sepal length (cm)

sepal width (cm)

petal length (cm)

petal width (cm)

Classes (targets):

0 = Setosa

1 = Versicolor

2 = Virginica

⚙️ Environment Setup 1️⃣ Clone or create the folder

Make sure your folder is named ML Basics & Optimization

2️⃣ Create and activate a virtual environment 🪟 On Windows (PowerShell) python -m venv venv .\venv\Scripts\Activate.ps1

🐧 On macOS / Linux python3 -m venv venv source venv/bin/activate

3️⃣ Install dependencies pip install -r requirements.txt

If any install fails, upgrade pip first: python -m pip install --upgrade pip

🚀 How to Run the Project

From inside the project folder (with the virtualenv activated):

python "src/task2_iris.py"

This will:

Train a baseline Logistic Regression model

Train an improved model using StandardScaler + GridSearchCV

Save:

✅ classification_report.txt

✅ confusion_matrix.png

✅ task2_iris_best_model.joblib

All outputs will be saved in the outputs/ folder.

📊 Results Summary Model Description Accuracy Baseline Logistic Regression (no scaling, default params) 0.9667 Improved StandardScaler + GridSearchCV (tuned C) 0.0333

🧩 Performance Gain: 0.0333 - 1.0 (improvement after optimization)

📈 Outputs Explained File Description classification_report.txt Precision, recall, and F1-score per class confusion_matrix.png Visualization of true vs predicted classes task2_iris_best_model.joblib Saved trained pipeline (StandardScaler + model) 🧠 Key Learnings

How to structure a basic ML project cleanly.

How to use the scikit-learn pipeline and grid search.

How scaling improves model convergence.

Importance of evaluation metrics beyond raw accuracy.