Time Series forecasting is an important area of Machine Learning because there are so many prediction problems involving a time component.
However, while the time component adds additional information, it also makes time series problems more difficult to handle compared to many other prediction tasks.
Time series forecasting is basically looking for patterns and eventually spanning them over long sequences.
The Prophet package in R and Python can make predictions according to the previously encountered data. Prophet, developed by Facebook engineers, is very popular in time series. Data can be represented at different points populated by example from demand history of your ERP or CMMS.
Multiple Inputs and Outputs
Time Series forecasting often requires dealing with multiple inputs and forecasting multiple time steps. A neural network can be applied which allows for a fixed/multiple number of inputs for a mapping function. The neural networks support multivariate inputs and thereby supporting multivariate forecasting. Complex Time Series evaluation requires multivariate and multi-step forecasting. Neural networks support an arbitrary number of output values as well to help with multiple outputs in time series forecasting.
Deep Learning methods offers a lot of promise for Time Series forecasting, such as the automatic learning of temporal dependence and the automatic handling of temporal structures like trends and seasonality. With their quality of extracting patterns from the input data for long durations, they have the perfect applicability in forecasting. They can, therefore, deal with large amounts of data, multiple, complex variables and multi-step actions, which is what time series forecasting demands.
Jean Michaud P.Eng.