Human Activity Recognition

The system predicts human activity with logistic regression, to help record an athlete’s activity and to make human–computer interaction possible.

• Explore and clean the data: visualise it and deal with null values and outliers.
• Train a logistic-regression model to predict the activity from the training data.
• Get results quickly by keeping only the features that affect the prediction most.

Problem statement

Fitness and sports

Sports, and fitness and running especially, are where activity recognition has resonated most. Athletes often need to track their activity to choose the next step in training, but recording it remains a problem.

Human–computer interaction

Games have improved a great deal, but people cannot interact with computers physically. Only by recognising what a user is doing can the computer understand them and respond to their actions, which makes that interaction possible.

Run it locally

git clone https://github.com/himanshu010/human-activity-recognition.git
cd human-activity-recognition
pip install -r requirements.txt

Then open Logistic regression.ipynb in Jupyter Notebook.