Machine learning actually attempts to generalize from provided examples. These generalizations (that's what we call predictive models) can then be used to predict or forecast. In the business world, the need for creating a quality forecast is obvious. A time series forecast is a type of forecast model that is key to predicting future performance based on past events. Specifically, the principal goal of the use of time series forecasting is to observe historical, linear time series trends and to use those identified patterns to predict future values (future performance).
Many methods (AR, MA, ARIMA, SARIMA, AANs, and SVMs) exist and are used to create time series forecast-based models. Understanding the logic of the routines and the differences between them can be overwhelming. Fortunately, the Splunk ML Toolkit makes it a somewhat straightforward process to create a time series forecast model (a degree in statistics not required!).