Key to Quality: Decoding the Language of Keystrokes for Predictive Insight

Overview

This project explores the relationship between keystroke patterns and the quality of a user's writing. We seek to understand whether keystroke data can be used to predict the quality of a user's writing. This work is the final project for CSCI 567: Machine Learning at the University of Southern California.

Data

The data used for this project comes from the Linking Writing Processes to Writing Quality Kaggle Competition.

Repository Structure

• /Dataset: Includes the dataset used in the project.
• /Docs: Documentation files, including the project poster and any additional explanatory materials.
• /Models: Jupyter notebooks used for exploratory data analysis and model experimentation.

Usage

1. Clone the repository: git clone https://github.com/anuranjanpandey/Key-To-Quality.git
2. Navigate to the project directory: cd keystroke-analysis
3. Install all packages conda install --file requirements.txt
4. After downloading the dataset from Kaggle, run the notebooks in Dataset Folder to get the preprocessed dataframe and time series data to train the LSTM and Transformer models.
5. Use these data frames to train models in the Models Folder.

Happy coding!

Contributors

This project was created by the following people:

1. Anuranjan Pandey
2. Dhruv Maheshwari
3. Kriti Jha
4. Nesar Nanjundaswamy
5. Pratyush Bhatnagar