Inspiration
Modern technology is advancing more rapidly than ever, yet crude, inaccurate lie detector devices are still commonplace in law enforcement in many countries. We wanted to apply state-of-the-art machine learning algorithms to see if we can build a better lie detector.
What it does
A muse headband gathers EEG data from the subject while they answer yes/no questions. A Gaussian Naive-Bayes machine learning algorithm searches this data to find patterns in eye, brain, and muscle activity that indicates lying. The computer then reports its guess: is the subject honest, or deceiving?
How I built it
We first had to hack the Muse headband to get the raw data into Python. Then we experimented with various models until we found one that worked best for a large number of features. Finally, we trained the model on training data from our fellow hackers! Thanks fellow hackers!
Accomplishments that I'm proud of
In just a weekend, with a relatively cheap EEG headband, we managed to come close to the industry standard for professional human and mechanical lie detectors.
What's next for Machine Learning Lie Detector
Big data and larger training datasets will help Machine Learning Lie detector will make (EEG) waves in the world of lie detection!
Built With
- data-science
- machine-learning
- muse-eeg-headband
- pandas
- python
- scikit-learn
- scipy
Log in or sign up for Devpost to join the conversation.