Understanding time series forecasts with smart predict
Key Takeaways
⇨ Time Series Analysis (TSA) in SAP Analytics Cloud (SAC) is focused on predicting numerical variables ('signals') in the future by using historical data and considering 'candidate influencers' that might affect the signal.
⇨ SAC's Time Series Forecasting uses various modeling techniques, including trend/cycle/fluctuation decomposition and exponential smoothing, to find the most accurate predictive model through algorithmic approaches that test multiple models before selecting the best fit.
⇨ Understanding the accuracy of the forecasting model is crucial; users can evaluate model performance using metrics like Mean Absolute Percentage Error (MAPE) to assess how well the model predicts known values.
This article explains the mechanisms behind Time Series Forecasting in SAP Analytics Cloud, detailing how it employs mathematical concepts for accurate predictions based on historical data and candidate influencers, while also highlighting its advanced modeling techniques including Trend/Cycle decomposition and Exponential Smoothing.
