Text Mining System Partial Result Preview
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems fail to provide fast and interactive querying of large document sets with unstructured or semi-structured text, leading to prolonged wait times for users to receive high-quality results, as they rely on computationally expensive text mining processes that deliver results sequentially rather than interactively.
Innovation Solution
The system provides partial text mining results to users by initiating multiple query sub-processes in parallel, using a text mining index to evaluate query constraints and sending progress notifications to display partial results as they become available, allowing users to see the query progress and refine their queries sooner.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text mining is performed on large document sets, then comprehensive and accurate results are obtained, but the query response time increases significantly
Solution Approach 1:
The patent applies partial action by displaying partial text mining results to users before the complete query processing is finished. The system shows intermediate results as they become available during parallel processing, allowing users to see and interact with partial findings rather than waiting for all results to be computed. This resolves the contradiction by providing timely partial accuracy rather than complete accuracy delayed by long wait times.
Solution Approach 2:
The patent segments the text mining query process into multiple parallel query sub-processes that can execute simultaneously. By dividing the large document set into smaller segments processed in parallel, the system maintains comprehensive coverage (complete action) while reducing overall query response time through concurrent execution, thus resolving the time-accuracy contradiction.
2Measurement precision
If text mining queries are executed sequentially to ensure accuracy, then result quality is maintained, but user interaction and query refinement are delayed
Solution Approach 1:
The system displays partial text mining results during the ongoing query process, enabling users to interact with and refine their queries based on intermediate findings. This maintains result quality through complete processing while improving operational ease by allowing iterative query refinement without waiting for full query completion, directly resolving the contradiction between accuracy and interaction speed.
Solution Approach 2:
The patent implements dynamic query processing where the system adapts to user interactions during the query execution. Users can refine queries based on partial results, and the system continues processing while incorporating refined query parameters, creating a dynamic feedback loop that maintains accuracy while enabling rapid iteration and improvement of query results.
3Loss of information
If all query results are processed and displayed before user interaction, then complete information is provided, but the iterative query refinement process is slowed down
Solution Approach 1:
The patent displays partial query results during the processing workflow rather than waiting for complete results. This provides sufficient information for users to perform meaningful query refinements while the remaining results continue to be processed in the background, maintaining information completeness while dramatically improving query iteration efficiency through overlapping computation and interaction.
Solution Approach 2:
The system maintains continuous useful action by allowing query refinement to occur during the ongoing text mining process. Rather than stopping to display all results before allowing interaction, the system continues processing while users refine queries, ensuring that computational work is not wasted and that iterative refinement can proceed without idle waiting periods, thus improving productivity while maintaining result completeness.
Data Source
AI summary
The system and methods described herein provide results previewing for an interactive text mining system in order to feedback partial query results to users before all results that are responsive to a query have been found. These partial results allow the user to see the progress of their text mining query much sooner.