Practical-Machine-Learning-Project-Activity-Recognition

📌 Project Overview

This project was completed as part of the Coursera Practical Machine Learning course.

The goal is to predict the manner in which participants performed barbell lifts using accelerometer data from wearable devices.
The target variable is classe, which represents five different execution styles (A–E), ranging from correct to incorrect exercise form.

---

📊 Data

The dataset comes from wearable sensors placed on:

• belt
• forearm
• arm
• dumbbell

Six participants performed barbell lifts in different ways, and sensor data was recorded.

---

🤖 Models Used

Two machine learning models were trained:

• Random Forest (final model)
• Gradient Boosting Machine - GBM (comparison)

---

📈Results

• Random Forest Accuracy: ~0.99
• GBM Accuracy: ~0.97

---

🎯 Final Prediction

The Random Forest model was used to predict 20 test cases provided in the assignment.

---

🧠 Files

• PML_project.Rmd – source code
• PML_project.html – final report
• pml-training.csv, pml-testing.csv – data

---

📌 How to run

Open PML_project.Rmd in RStudio and click Knit to HTML. Finish