Hierarchical Time Series Database for Parallel Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current time series forecasting systems face challenges in efficiently assembling, storing, and comparing numerous forecasts derived from large historical data sets, leading to delays in generating forecasts and inefficient processing of time series data.
Innovation Solution
A grid-computing system employing a hierarchical schema for distributed data storage and processing, utilizing volatile memory like RAM for quick access and parallelized computing to assemble and store time series data, enabling efficient generation of multiple forecasts simultaneously.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If time series data is assembled and processed using traditional computing systems, then forecasts can be generated, but time delays occur in generating forecasts and processing efficiency is reduced
Solution Approach 1:
The patent segments the time series data assembly process into multiple hierarchical levels (e.g., daily, weekly, monthly aggregations). Each level processes and stores pre-computed results independently, allowing parallel computation across levels and eliminating sequential processing bottlenecks that cause time delays in traditional systems.
Solution Approach 2:
The system performs preliminary assembly and aggregation of time series data at multiple hierarchical levels before actual forecast generation is needed. By pre-computing and storing aggregated statistics (means, variances, trends) at different time granularities, the system eliminates computation delays during forecast generation, enabling rapid forecast production when needed.
2Quantity of substance
If numerous forecasts are derived from large compilations of historical data, then comprehensive analysis is possible, but processing efficiency decreases and time delays increase
Solution Approach 1:
The patent introduces a hierarchical time dimension to organize historical data, structuring it by multiple levels of aggregation (intraday, daily, weekly, monthly). This dimensional transformation allows the system to process large volumes of historical data more efficiently by operating at appropriate aggregation levels rather than processing all raw data uniformly, thereby maintaining processing efficiency while handling comprehensive historical datasets.
Solution Approach 2:
The system performs preliminary aggregation and preprocessing of large historical data compilations into hierarchical structures before forecast analysis. By pre-organizing data into manageable hierarchical levels with pre-computed statistics, the system reduces the computational burden during actual forecast processing, maintaining high processing efficiency even when analyzing numerous forecasts from extensive historical data.
3Adaptability or versatility
If multiple time series are assembled from large data sets, then comprehensive forecasting is achieved, but assembly time and processing delays increase
Solution Approach 1:
The patent segments the comprehensive time series assembly into hierarchical components (different time granularities and data levels). Each segment is assembled and stored independently, allowing parallel processing and eliminating the need to assemble all comprehensive data sequentially. This segmentation maintains comprehensive forecasting capability while dramatically reducing assembly time through parallel execution of segmented assembly tasks.
Solution Approach 2:
The system performs preliminary assembly of time series data at multiple hierarchical levels before comprehensive forecast analysis is required. By pre-assembling and storing time series at different aggregation levels (daily, weekly, monthly), the system maintains comprehensive forecasting adaptability while reducing assembly time during actual use, as the comprehensive structure is already prepared in advance.
Data Source
AI summary
This disclosure describes methods, systems, and computer-readable media for accessing information that describes a hierarchical schema for assembling multiple time series of data in a distributed manner. The hierarchical schema associates each of the time series with a particular level of the hierarchical schema and prescribes a structure of relationships between time series assigned to different levels of the hierarchical schema. Multiple time series associated with a lowest level of the hierarchical schema are assembled by inventorying a portion of a data set. Multiple time series associated with an intermediate level of the hierarchical schema are assembled by aggregating the time series associated with the lowest level based on the structure of nested relationships. Also, multiple additional time series that are associated with the intermediate level and which were assembled by other grid-computing devices are received. After the time series are assembled, they are made available for processing to facilitate parallelized forecasting.