Machine Learning with MATLAB

Overview

This repository contains a curated collection of MATLAB-based machine learning projects focused on practical data analysis, supervised learning, signal-feature classification, and reinforcement learning for control-oriented simulation. The projects demonstrate how machine learning workflows can be implemented, evaluated, and documented using MATLAB and related toolboxes.

Technical Focus

Machine learning with MATLAB emphasizes end-to-end experimentation: importing data, preparing features, training models, evaluating predictive performance, and interpreting results with clear visual and numerical summaries.

The projects in this repository cover:

• supervised classification and regression
• data cleaning and missing-value handling
• feature engineering and dimensionality reduction
• cross-validation and holdout evaluation
• model comparison using quantitative metrics
• confusion matrix and classification-performance analysis
• regression evaluation using RMSE, MAE, and R-squared
• predictor importance and correlation analysis
• wireless-signal feature extraction
• reinforcement learning for continuous control
• Simulink-based environment modeling

Skills & Technologies

Core skills: MATLAB, Machine Learning, Classification, Regression, Feature Engineering, Data Cleaning, Cross-Validation, Model Evaluation, Confusion Matrix, RMSE, MAE, R-squared, PCA, KNN, Naive Bayes, LDA, Decision Tree, Ensemble Learning, SVM, Reinforcement Learning, TD3, Simulink, Control Systems, Technical Documentation

Project Directory

No.	Project	Topic	Purpose	Main Concepts	Skills / Tags
01	01-secom-classification	SECOM Manufacturing Classification	Build and evaluate classification models for semiconductor manufacturing quality prediction	missing-value imputation, PCA, Naive Bayes, LDA, Decision Tree, Ensemble learning, imbalanced classification	MATLAB, Classification, PCA, Naive Bayes, LDA, Decision Tree, Ensemble, Confusion Matrix
02	02-auto-mpg-regression	Auto MPG Regression	Predict vehicle fuel efficiency from numerical vehicle attributes and compare regression approaches	data cleaning, correlation analysis, regression modeling, model comparison, predictor importance	MATLAB, Regression, Linear Regression, Regression Tree, SVM, Bagged Trees, RMSE, Feature Importance
03	03-wireless-channel-classification	Wireless Channel Classification	Classify wireless channel measurements using extracted statistical features	signal feature extraction, KNN classification, cross-validation, train-test split, confusion matrix	MATLAB, KNN, Feature Engineering, Signal Data, Cross-Validation, Classification Metrics
04	04-water-tank-reinforcement-learning	Water Tank Reinforcement Learning	Train a reinforcement learning agent for water-level control in a Simulink environment	actor-critic networks, TD3 agent, continuous control, reset randomization, simulation-based training	MATLAB, Simulink, Reinforcement Learning, TD3, Control Systems, Actor-Critic

Skill Coverage Map

Skill / Concept	01 SECOM Classification	02 Auto MPG Regression	03 Wireless Channel Classification	04 Water Tank RL
MATLAB scripting	✅	✅	✅	✅
Data cleaning	✅	✅	✅	✅
Missing-value handling	✅	✅
Feature engineering	✅	✅	✅	✅
Supervised classification	✅		✅
Regression modeling		✅
Reinforcement learning				✅
Cross-validation	✅	✅	✅
Holdout test evaluation	✅	✅	✅	✅
Model comparison	✅	✅	✅
Confusion matrix analysis	✅		✅
Correlation analysis		✅
Predictor importance analysis	✅	✅
PCA / dimensionality reduction	✅
Signal feature extraction			✅
Simulink environment modeling				✅
Continuous-control modeling				✅
Result visualization	✅	✅	✅	✅
Technical documentation	✅	✅	✅	✅

Repository Structure

machine-learning-matlab/
├── README.md
├── .gitignore
├── 01-secom-classification/
├── 02-auto-mpg-regression/
├── 03-wireless-channel-classification/
└── 04-water-tank-reinforcement-learning/

Typical project folders may use the following structure:

project-folder/
├── README.md
├── src/
├── scripts/
├── results/
├── data/
└── model/

Some folders may omit directories that are not required for that specific project.

Reproducibility Notes

Most projects are designed to run in MATLAB with project-specific scripts or live scripts. A typical workflow is:

% Open MATLAB from the project folder
% Then run the main script for the selected project
run("src/main.m")

For projects that use MATLAB Live Scripts, open the corresponding .mlx file in MATLAB and run the sections in order. For Simulink-based reinforcement learning, open the referenced .slx model before starting training or simulation.

Recommended MATLAB products:

• MATLAB
• Statistics and Machine Learning Toolbox
• Deep Learning Toolbox
• Reinforcement Learning Toolbox
• Simulink