🤖 Supervised Machine Learning – Regression, Classification & Model Evaluation

This repository contains my completed coursework for Machine Learning Assignment 1. The focus is on building and evaluating supervised learning models from scratch using real-world datasets.

From fitting regression curves to tuning classification algorithms, this project demonstrates my ability to handle data preprocessing, apply core ML algorithms, and assess model performance with industry-standard metrics — all implemented in Python using the scikit-learn.

---

What I Learned / What I Can Do

✅ Apply linear and polynomial regression to numerical data
✅ Use GridSearchCV for hyperparameter optimization
✅ Preprocess datasets: handle missing values, normalize, and encode features
✅ Train and evaluate classifiers: Logistic Regression, KNN, Naive Bayes
✅ Compare models using metrics like Accuracy, F1-score, Recall, Precision
✅ Use pipelines to ensure clean, modular, and reproducible ML workflows

---

Contents

The solution is presented in the Notebook

Task 1 – Regression Modeling

• Load and split dataset from task1_data.csv
• Train and evaluate linear regression model using:
  • MSE, RMSE, MAE, R²
• Perform polynomial regression and select optimal degree using GridSearchCV
• Visualize and compare model performance

Task 2 – Classification Pipeline

• Load pokemon_modified.csv and perform:
  • Missing value imputation
  • One-hot encoding for categorical features
  • Feature scaling
• Train and tune:
  • Logistic Regression
  • K-Nearest Neighbors (KNN)
  • Naive Bayes
• Evaluate models using:
  • Accuracy, Precision, Recall, F1-score
• Select best-performing model

---

Technologies

• Python (Jupyter Notebook)
• numpy, pandas, matplotlib, scikit-learn
• Data preprocessing pipelines
• Cross-validation with GridSearchCV
• Classification and regression metrics

---

Author

Valeria Neganova
Focus: Practical supervised learning, evaluation & model selection
📫 Valerochka.neganova@mail.ru