Faceted Navigation Keyword Extraction via Clickstream Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing faceted navigation systems on websites rely primarily on product attributes for filtering and sorting, which may not fully capture the relevant keywords from product descriptions, leading to incomplete reflection of user intent.
Innovation Solution
A method that extracts candidate keywords from clickstream data and natural language processing of product text, filters them based on clickstream data, ranks them for relevance, clusters to remove near duplicates, and outputs them for display in faceted navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If keywords are mapped based on product attributes during product onboarding, then faceted navigation can filter and sort search results, but numerous relevant keywords in product descriptions are not captured
Solution Approach 1:
The system performs preliminary extraction of candidate keywords from clickstream data and product text before the actual search query is processed. This advance preparation ensures that relevant keywords are captured and stored for future use, preventing information loss while maintaining operational efficiency during search operations.
Solution Approach 2:
The patent introduces an intermediary processing layer that includes filtering, ranking, and clustering components. This intermediary system processes the raw extracted keywords through multiple stages: filtering removes irrelevant terms, ranking prioritizes important keywords, and clustering groups near-duplicates. This multi-stage intermediary process captures comprehensive keywords while managing system complexity through modular design.
2Adaptability or versatility
If all extracted keywords from product text are used for faceted navigation, then user intent is better reflected, but the system complexity and processing requirements increase
Solution Approach 1:
The system performs preliminary filtering, ranking, and clustering of keywords before they are made available for faceted navigation. This advance processing ensures that only high-quality, relevant keywords are stored and presented to users, improving adaptability while preventing the system from being overwhelmed by processing all possible keywords in real-time.
Solution Approach 2:
The patent applies parameter changes by transforming raw keywords through filtering (removing stop words and non-informative terms), ranking (prioritizing by relevance metrics), and clustering (grouping similar terms). These parameter transformations convert a large, unstructured set of keywords into a refined, organized set that better reflects user intent while reducing processing complexity.
3Measurement precision
If candidate keywords are extracted from both clickstream data and natural language processing, then more relevant keywords are found, but the filtering and ranking process becomes more complex
Solution Approach 1:
The patent segments the keyword extraction and processing system into distinct modular components: an extraction module that gathers keywords from clickstream data and product text, a filtering module that removes irrelevant terms, a ranking module that prioritizes keywords by relevance, and a clustering module that groups near-duplicates. This segmentation allows each component to specialize in one task, improving measurement precision while managing complexity through modular architecture.
Solution Approach 2:
The system introduces intermediary processing layers between data extraction and final keyword selection. The filtering, ranking, and clustering components act as intermediaries that systematically process the raw extracted keywords from multiple sources. This intermediary approach ensures high measurement precision by thoroughly evaluating keywords from diverse sources while managing complexity through structured intermediate processing steps.
Data Source
AI summary
A method includes extracting a set of candidate keywords from clickstream data and natural language processing of product text for a plurality of search queries. The set of candidate keywords are filtered based on the clickstream data. The set of candidate keywords as filtered are ranked based on the clickstream data. The set of candidate keywords as ranked are clustered to remove near duplicates. The set of candidate keywords as ranked for a respective search query is output.