Annotated Search Index via Historical Session Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engine technologies face challenges in providing complete and efficient search results due to the lack of direct association between search query terms and relevant resources, especially when resources do not contain textual references, leading to increased computing resources and response time.
Innovation Solution
Generating an annotated search index by retrieving historical search sessions, calculating an association parameter based on transitions and user interaction time, and associating non-indexed resources with indexed ones to enhance search result relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If an inverted index is created with all search terms from the data collection, then search query execution speed is improved, but computing resources and processing time increase significantly
Solution Approach 1:
The patent extracts only the necessary search terms from the data collection by analyzing historical search sessions and user behavior patterns. Instead of indexing all possible search terms, the system identifies and indexes only those terms that are actually used in historical queries, thereby reducing the index size and computing resources required while maintaining search efficiency.
Solution Approach 2:
The system performs preliminary analysis of historical search sessions to pre-identify relevant search terms and resources before actual search queries are executed. By analyzing past user behavior and extracting useful search term associations in advance, the system prepares an optimized index structure that reduces processing time during actual search operations.
2Device complexity
If resources without textual search term references are excluded from indexing, then indexing complexity is reduced, but search result completeness deteriorates
Solution Approach 1:
The patent introduces an intermediary association mechanism that links resources without direct textual search term references to relevant search terms through historical search session analysis. The system uses user behavior patterns and resource transition data as intermediaries to establish indirect associations, allowing resources lacking explicit textual references to be included in search results through these intermediary connections.
Solution Approach 2:
The system creates a multi-functional indexing approach where search terms serve multiple purposes: direct textual matching for resources containing the terms, and indirect association through historical search patterns for resources without explicit references. This universal approach allows the same indexing mechanism to handle both indexed and non-indexed resources, maintaining search result completeness without increasing indexing complexity.
3Reliability
If all resources are indexed with all possible search terms, then search result relevance is improved, but data storage requirements and processing time increase
Solution Approach 1:
The patent applies local quality by creating search term associations selectively based on local context and historical usage patterns. Instead of uniformly indexing all resources with all possible search terms, the system identifies specific local associations between search terms and resources based on actual user behavior in historical sessions, storing only these relevant local associations rather than all possible combinations.
Solution Approach 2:
The system uses partial action by indexing only the subset of search term-resource associations that are supported by historical search session data. Rather than attempting to index all possible associations (excessive action), the system selectively indexes only those associations that have been observed in actual user behavior, reducing storage requirements while maintaining sufficient relevance for practical search purposes.
Data Source
AI summary
There is disclosed a method and a system for generating an annotated search index. The method is executable at a server. The method comprises retrieving a portion of a historical search session for a first search query, the portion including a first resource and a second resource, both the first and the second resource being relevant to the first search query; generating an association parameter for the second resource, the association parameter being based on a first historical parameter and a second historical parameter; and, responsive to the association parameter for the second resource being above a predetermined threshold, associating the second resource with one or more of the first resource and the included search terms, thereby generating an annotated search index for the included search terms.