Forecasting Datasets Using Blended Temporal and Grouped Aggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud service providers face challenges in accurately forecasting future values due to errors or reduced accuracy in initial forecasted values, which propagate to subsequent forecasts, especially when dealing with multiple time series data.
Innovation Solution
The implementation of a blended approach that combines temporal aggregation and grouped aggregation techniques to manipulate data points from multiple time series, creating new time series for improved forecasting accuracy by aligning and summing data points, and aggregating adjacent points, allowing for more accurate future value predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If initial forecast values are used for subsequent predictions, then forecasting process can proceed, but accuracy deteriorates due to error propagation
Solution Approach 1:
The patent segments the forecasting process into multiple independent time series, each processed separately through aggregation and forecasting steps. By dividing the overall forecasting task into smaller, manageable segments (different time series with different aggregation levels), the system avoids compounding errors that would occur in a single sequential forecast chain, thereby maintaining both productivity and accuracy.
Solution Approach 2:
The patent applies preliminary aggregation to time series data before forecasting. By pre-processing the data through temporal and/or grouped aggregation to create aggregated time series, the system establishes a more stable foundation for forecasting that reduces error propagation in subsequent predictions, while still allowing the forecasting process to proceed efficiently.
2Measurement precision
If multiple time series are processed individually, then forecasting accuracy for each series is maintained, but processing time increases
Solution Approach 1:
The patent merges multiple time series processing operations by implementing aggregation mechanisms that combine data from multiple series. Through temporal aggregation (combining data points within time periods) and grouped aggregation (combining data from multiple series), the system processes multiple time series more efficiently while maintaining the ability to generate accurate forecasts for each series, thus reducing overall processing time.
3Measurement precision
If aggregated time series are created through temporal and grouped aggregation, then forecasting accuracy is improved, but data complexity increases
Solution Approach 1:
The patent applies local quality by allowing different time series to have different aggregation characteristics. Each time series can be processed with appropriate temporal and/or grouped aggregation based on its specific properties and requirements. This targeted approach improves forecast accuracy for each series while avoiding the unnecessary complexity that would arise from applying uniform complex aggregation to all series.
Data Source
AI summary
Techniques are described herein for forecasting datasets using blend of temporal aggregation and grouped aggregation. An example method can include a device accessing a first and second time series, comprising a first data point associated with a first time step and a first value and a second data point associated with a second time step and a second value. The method can further include the device determining a grouped aggregated data point using the first and second time series by aligning the first and second data point. The method can further include the device determining the grouped aggregated data point by summing the first and second value. The method can further include determining a grouped aggregated time series. The method can further include the device determining a first set of input values for a machine learning model. The method can further include the device determining a first forecasted future value.