Neural Network Projects with Python

Predicting Taxi Fares with Deep Feedforward Nets

In Chapter 3, Predicting Taxi Fares with Deep Feedforward Nets, we used a deep feedforward neural network in a regression prediction problem, where the task was to predict the dollar amount of a NYC taxi fare, based on features such as pick-up and drop-off locations. In this project, we had to work with a noisy dataset that included missing data and data anomalies. We saw how data visualization is essential to help us identify outliers in the dataset, and to discover important trends in the dataset.

This project was also the first project where feature engineering was done. Using the existing features, we created other features that improved the accuracy of our neural network. Finally, we ...