Amazon is announcing that now its time-series machine learning based forecasting service Amazon Forecast can run what-if assessments to determine how different business scenarios can affect demand estimates. What-if analysis is an effective business technique for simulating hypothetical scenarios and stress testing on planning assumptions by recording potential outcomes.
According to Amazon, this launch brings together the predictive strength of Amazon Forecast with a seamless experience to support answering hypothetical questions and quantifying the influence of scenarios on forecasts.
A what-if analysis is a tool to help investigate and explain how different scenarios might affect the baseline forecast created by Amazon Forecast. The baseline forecast is the forecast that is created by Amazon Forecast based on the original related time series. A what-if analysis creates a series of what-if forecasts based on how one chooses to modify the related time series. Those what-if forecasts are compared and contrasted with the baseline forecast to help understand how specific changes might impact the model.
There are two methods for creating a modified related time series: provide a modified related time series in an Amazon S3 path, or specify a set of transformations to the existing related time series.
Amazon Forecast uses machine learning to generate more accurate demand forecasts with just a few clicks, without requiring any prior ML experience. Amazon Forecast includes algorithms that are based on over twenty years of forecasting experience and developed expertise used by amazon.com bringing the same technology used at Amazon to developers as a fully-managed service, removing the need to manage resources.
Amazon Forecast provides six different comprehensive accuracy metrics to help understand the performance of forecasting models and compare it to previous forecasting models created that may have looked at a different set of variables or used a different period of time for the historical data.
In addition, it automatically tracks the model accuracy over time as new data is imported. It can systematically quantify the model’s deviation from initial quality metrics and make more informed decisions about keeping, retraining, or rebuilding the model as new data comes in.
Amazon Forecast can be easily imported into common business and supply chain applications, such as SAP and Oracle Supply Chain. This makes it easy to integrate more accurate forecasting into the existing business processes with little to no change. Forecast results might view for each scenario in a single graph or export the data in bulk for offline analysis.
Time series forecasting will become more automated in the future and cloud platforms like Azure Time Series, Facebook’s Prophet, SAP Cloud Analytics, IBM Cognos Analytics with Watson among others are developing tools to help address these challenges.