Web Query Classification Using Centroid and Bayesian Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Web query classification is challenging, especially for queries with few words, as they lack rich textual features, and existing methods can yield ineffective search results depending on the classification method used.
Innovation Solution
A method and system that integrate a reductionist and enrichment module to classify queries, where the reductionist module identifies a core term using a centroid algorithm and the enrichment module uses a Bayesian transformation function to determine category probabilities, with a feedback loop to enhance precision and recall.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single module (reductionist or enrichment) is used for query classification, then the system is simple, but the classification precision and recall are insufficient
Solution Approach 1:
The patent combines two distinct classification modules (reductionist module and enrichment module) into an integrated system. The reductionist module processes queries using a centroid algorithm to identify core terms, while the enrichment module uses Bayesian transformation to calculate category probabilities. By merging these modules and integrating their results through a feedback loop, the system achieves superior classification precision and recall compared to using either module alone, while maintaining manageable system complexity through modular architecture.
2Productivity
If queries with few words are classified using traditional methods, then the classification process is fast, but the search results are ineffective due to lack of rich textual features
Solution Approach 1:
The patent introduces an intermediary feedback loop mechanism that connects the reductionist and enrichment modules. For short queries with limited textual features, the reductionist module quickly identifies core terms, then the enrichment module uses Bayesian transformation to infer category probabilities based on these core terms. The feedback loop allows results from one module to enhance the other, enabling effective classification of short queries without sacrificing classification speed, as the modular architecture allows parallel processing.
3Ease of manufacture
If existing classification methods are used, then the implementation is straightforward, but the results are ineffective for certain query types
Solution Approach 1:
The patent segments the query classification task into two distinct functional modules: the reductionist module that focuses on identifying core terms using centroid algorithms, and the enrichment module that calculates category probabilities using Bayesian transformation. This segmentation allows each module to specialize in specific aspects of classification, improving overall effectiveness for different query types while maintaining ease of implementation through modular design where each module can be independently developed, tested, and maintained.
Data Source
AI summary
Implementing query classification includes receiving a core term from a search query responsive to execution of a first module. The first module searches a table for the core term and yields a first result. The query classification also includes receiving a second result from the search query responsive to execution of a second module. The second module searches an index of terms that are mapped to documents and corresponding categories in the index. The second result is indicative of one of the corresponding categories in the index based on a probability score. Upon determining the first result is a category associated with the core term in the table, the query classification also includes calculating a weighted average for the first result and the second result. The calculation yields a third result. The query classification further includes transmitting the third result to a computer device that generated the query.