18061899. Selecting Influencer Variables in Time Series Forecasting simplified abstract (SAP SE)
Contents
Selecting Influencer Variables in Time Series Forecasting
Organization Name
Inventor(s)
Selecting Influencer Variables in Time Series Forecasting - A simplified explanation of the abstract
This abstract first appeared for US patent application 18061899 titled 'Selecting Influencer Variables in Time Series Forecasting
Abstract: Optimizing a time series forecasting model involves selecting a subset of original influencer variables. The process includes receiving an original time series forecasting model with an original set of influencer variables, calculating the contributions of these variables to the model, and excluding variables falling below a cumulative contribution threshold. A new time series forecasting model is then created from the remaining variables, and if validated, further iterations may occur to refine the model. If not validated, the cumulative contribution threshold is adjusted to generate another new model. Ultimately, a subset of original influencer variables is selected for the new time series forecasting model.
- Simplified Explanation:
- The patent involves optimizing a time series forecasting model by selecting a subset of influencer variables based on their contributions to the model. - The process includes excluding variables with low contributions and refining the model through iterations.
- Key Features and Innovation:
- Selection of influencer variables based on their contributions to the forecasting model. - Iterative process to refine the model by adjusting the subset of variables.
- Potential Applications:
- Time series forecasting in various industries such as finance, retail, and healthcare. - Demand forecasting, stock market prediction, and resource planning.
- Problems Solved:
- Efficient selection of influencer variables for accurate time series forecasting. - Improved model performance by excluding irrelevant variables.
- Benefits:
- Enhanced accuracy in time series forecasting. - Reduction in model complexity and computational resources.
- Commercial Applications:
- "Optimized Time Series Forecasting Model for Enhanced Predictions in Financial Markets"
- Prior Art:
- Researchers have explored variable selection techniques in time series forecasting models. - Studies have focused on the impact of influencer variables on forecasting accuracy.
- Frequently Updated Research:
- Ongoing research on automated variable selection methods in time series forecasting. - Studies on the integration of machine learning algorithms for improved forecasting accuracy.
Questions about Time Series Forecasting Model Optimization: 1. How does the process of selecting influencer variables impact the accuracy of time series forecasting models? 2. What are the potential challenges in implementing an iterative approach to refining forecasting models?
Original Abstract Submitted
Optimizing a time series forecasting model, selects a subset of original influencer variables. An original time series forecasting model comprising an original set of influencer variables, is received. Contributions of the influencer variables to the model are calculated (optionally including regularization). Variables falling below a cumulative contribution threshold, are excluded. A first new time series forecasting model, created from the remaining variables, is stored. If the first new time series forecasting model is validated based upon a performance horizon, iteration occurs to further reduce a number of influencer variables and generate another new time series forecast model. If the first new time series forecasting model is not validated under the performance horizon, the cumulative contribution threshold is lowered to exclude fewer of the original set of influencer variables and generate another new model. A subset of original influencer variables ultimately selected for a new time series forecasting model, is output.
