Machine Learning Time Series Prediction w/ TensorFlow

RNNs & LSTMs have enjoyed great success in text generation algorithms, but their use in other fields has not been as widely studied. We will discuss our experiences & progress using Recurrent Neural Networks to make predictions on arbitrary multivariate time series data. Our first study used weather data from the JFK terminal over several years. We will discuss the issues related to tuning & validating this model, as well as how we migrated this model into the Model Asset Exchange, which is an IBM hosted API for making predictions on data using pre-trained neural network models. Our insight into tuning this model allowed us to provide another API via Watson Machine Learning, which is a hosted service that allows user defined data & models to be uploaded, trained, & tuned on GPU accelerated Watson Studio Notebooks on demand hardware using simple remote API calls.