Partial Query Auto-Completion Clustering for Search Diversity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search systems provide limited and non-diverse autosuggestions based on partial query strings, leading to decreased search efficiency and user engagement, as they often rely solely on semantic similarity and fail to utilize the available search pane effectively, resulting in users ignoring suggestions until more characters are entered.
Innovation Solution
The system generates cluster groups of candidate suggestions based on similarity and popularity, displaying them in a hierarchical manner to increase diversity and visibility, allowing users to select cluster labels or sub-topics as complete query terms, thereby enhancing search efficiency and user interaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional autosuggest systems display limited suggestions based solely on semantic similarity, then the system complexity remains low, but the diversity and visibility of suggestions decrease, leading to reduced user engagement
Solution Approach 1:
The patent segments the autosuggestion system into multiple independent components: (1) multiple suggestion generation modules that create different types of suggestions (popular queries, related topics, autocomplete) independently, (2) a clustering module that groups suggestions by semantic similarity, and (3) a display module that presents clustered suggestions in an organized hierarchy. This segmentation allows each component to specialize and improves overall suggestion diversity without proportionally increasing system complexity.
Solution Approach 2:
The patent introduces a new dimension to the suggestion display by organizing suggestions into clustered groups with hierarchical structure (parent clusters and child suggestions). Instead of displaying suggestions as a flat list, the system creates a two-dimensional organization where suggestions are grouped by semantic categories, allowing users to explore multiple dimensions of search possibilities simultaneously. This dimensional change increases visibility and diversity of displayed suggestions.
2Adaptability or versatility
If the system displays more suggestions to increase diversity, then user engagement improves, but the effective use of search pane space becomes inefficient
Solution Approach 1:
The patent divides the suggestion display into segmented clustered groups rather than a single continuous list. Each cluster represents a semantic category and can be independently sized and positioned. This segmentation allows the system to pack more suggestions into the available search pane area by efficiently utilizing vertical and horizontal space through multiple compact clusters rather than one elongated list.
Solution Approach 2:
The patent transitions from a one-dimensional vertical list of suggestions to a two-dimensional clustered layout where suggestions are organized in groups that can be arranged both vertically and horizontally within the search pane. This dimensional change allows more suggestions to be displayed within the same area by utilizing the pane's width more effectively, improving space utilization while increasing the number of visible suggestions.
3Measurement precision
If users must enter more characters to get relevant suggestions, then suggestion accuracy improves, but search time increases
Solution Approach 1:
The patent implements preliminary action by generating and displaying multiple types of suggestions (popular queries, related topics, autocomplete suggestions) simultaneously as the user types, rather than waiting for a minimum character threshold. The system proactively provides relevant suggestions at each typing stage, clustered by semantic similarity, allowing users to select from accurate suggestions before they have fully formed their query, thereby reducing search time while maintaining accuracy.
Solution Approach 2:
The patent adds a temporal dimension to suggestion accuracy by providing multiple layers of suggestions at different stages of query formation. Instead of a single accuracy threshold, the system offers clustered suggestions that maintain relevance across different query lengths, allowing users to find accurate suggestions faster by selecting from organized groups rather than waiting for longer query input.
4Productivity
If the system uses hierarchical arrangement of cluster groups and sub-topics, then user selection efficiency improves, but the device complexity increases
Solution Approach 1:
The patent segments the suggestion display into hierarchical clusters with parent categories and child suggestions. This segmentation creates a structured organization where users can navigate through grouped suggestions rather than scanning a flat list. The hierarchical structure improves user selection efficiency by reducing cognitive load and enabling faster navigation through organized categories, while the modular nature of the segmentation keeps implementation complexity manageable.
Data Source
AI summary
Aspects of the present disclosure relate to providing, based on a partial query string, a plurality of autosuggestions that are diverse in nature such that the user is more likely to see the preferred complete query terms and therefore more likely to select one of the preferred suggestions which will increase search efficiency. As described herein, such functionality relates to generating cluster groups of candidate suggestions, each cluster including sub-topics, then performing the search based on a selected cluster or sub-topic. To generate the cluster groups, systems and methods, as described herein, analyze the similarity between candidate suggestions as well as the popularity of generated sub-topics. The cluster groups and sub-topics may be displayed visually, and in certain embodiments the cluster groups are ordered vertically from top to bottom and aligned to the left side of the display, while sub-topics are ordered horizontally from left to right following the cluster label.