A neuron predicting application that uses machine learning to predict neuron type based on firing characteristics
A web application was created using machine learning to train a model on predicting neuron types based on firing characteristics. The data was aquired from NeuroElectro and was used to train a Gradient Boosting model from the scikit-learn library. After training the model on 80% of the data (filtered down to the top 6 firing characteristics with the most data), it predicted the correct neuron type with 86% accuracy.
A Heroku web application was created with this model and can be accessed here: Neuron Predictor
The landing page
After entering data into the 6 input fields, the app makes a prediction
