Analysis Model Preparing System with Intermediate Data Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data analysis model preparation methods are inefficient due to the need for trial and error in selecting and tuning analysis models, leading to lengthy processing times and potential errors in restoring previous model states, especially when using caching technologies that fail to maintain accurate intermediate data relationships.
Innovation Solution
An analysis model preparing system that includes an execution unit, partial execution unit, cache, cache processing unit, maintaining unit, storage unit, and general processing unit to manage intermediate data and maintain mapping information between data and models, allowing for efficient partial execution and data retrieval without re-executing unchanged processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If caching technology is used to store intermediate data, then execution efficiency is improved, but accuracy of intermediate data relationships deteriorates when models are restored
Solution Approach 1:
The patent introduces mapping information as an intermediary data structure that connects cached intermediate data with their corresponding analysis models and process positions. This mapping information includes identifiers for the analysis model, process identifiers, and position information, allowing the system to accurately retrieve the correct intermediate data when restoring models without re-executing unchanged processes.
Solution Approach 2:
The system performs preliminary actions by caching intermediate data during the initial execution of the analysis model and storing the mapping relationships between this data, the model version, and process positions. When the model is later restored or modified, the system uses this pre-established mapping to quickly identify and retrieve applicable cached data, avoiding unnecessary re-execution.
2Measurement precision
If trial and error process is used to select and tune analysis models, then model precision is improved, but preparation time increases
Solution Approach 1:
The patent implements a mechanism to discard intermediate data when the analysis model is modified and recover it when the model is restored to a previous state. The system tracks modifications to the analysis model and automatically determines whether to invalidate cached intermediate data or reuse it, enabling efficient trial and error processes without unnecessary re-computation.
Solution Approach 2:
The system maintains continuity of useful action by preserving cached intermediate data across model restorations and modifications. When the analysis model is restored to a previous state, the system continues to use the cached intermediate data from that state, allowing the trial and error process to proceed without interrupting useful computational work.
3Reliability
If entire analysis model is re-executed after model changes, then data accuracy is maintained, but execution time increases
Solution Approach 1:
The patent segments the analysis model into individual processes with unique identifiers, allowing the system to track which specific processes have changed and which have remained unchanged. This segmentation enables selective re-execution of only the modified processes while reusing cached results from unchanged processes, maintaining data accuracy while reducing execution time.
Data Source
AI summary
An analysis model execution unit executing a part of an analysis model, an analysis model partial execution unit partially executing the analysis model based on intermediate data generated during execution of the analysis model, external storage storing the intermediate data and mapping information which is corresponding relationship between the intermediate data and the analysis model, and an analysis model general processing unit generating the mapping information by associating the intermediate data with the analysis model and reading the intermediate data associated with the analysis model from the external storage based on the mapping information are provided.