Document Ranking via User Performance Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search and recommendation systems fail to effectively rank enterprise documents, such as forms and manuals, which lack links, hindering employees' ability to find valuable information that improves productivity, as they do not account for the intrinsic characteristics of information content and its relation to worker productivity.
Innovation Solution
A method and apparatus that rank information content based on performance data of prior users by assigning scores using a regression model combining performance measures and social network metrics, such as graph entropy and revenue generated by related users, to identify valuable topics and improve search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If existing search systems use link-based scoring to rank documents, then the ranking is simple to compute, but enterprise documents without links cannot be effectively ranked
Solution Approach 1:
The patent changes the ranking parameters from link-based metrics to performance-based metrics. Instead of using link counts, the system uses performance measures (revenue, performance ratings) of users who accessed documents, combined with social network metrics (graph entropy, betweenness, outdegree) to compute document scores. This parameter transformation enables effective ranking of enterprise documents that lack links.
2Measurement precision
If search systems focus on social network topology, then network structure analysis is improved, but the correlation with productivity and information content value is insufficient
Solution Approach 1:
The patent merges social network topology analysis with performance data analysis. The system combines network metrics (graph entropy, betweenness, outdegree) with performance measures (revenue generated, performance ratings) of users who accessed documents. This integration ensures that both network structure precision and productivity correlation are captured in the document ranking process.
Solution Approach 2:
The patent creates a composite scoring mechanism that combines multiple types of data: social network metrics, performance measures, and content elements. This composite approach integrates diverse information sources to produce a comprehensive document score that reflects both network structure and productivity value.
3Measurement precision
If comprehensive performance data collection is implemented, then productivity measurement accuracy is improved, but system complexity increases
Solution Approach 1:
The patent implements self-service data collection by automatically gathering performance data from existing enterprise systems. The system automatically collects performance measures (revenue, performance ratings) and user access patterns without requiring manual input. This automation maintains measurement precision while reducing operational complexity.
Data Source
AI summary
Methods and apparatus are provided for ranking information content based on performance data of prior users of the information content. Information content is ranked by receiving a search request specifying search criteria; identifying a preliminary document list by searching a corpus using the search criteria; identifying content elements in documents in the preliminary document list; assigning a value to at least one document in the preliminary document list based on a score for each content element in the at least one document, wherein the score is based on a performance measure of one or more related users that accessed one or more documents having a given content element; and providing search results based on the assigned values. The score can be assigned, for example, based on a regression model between the performance measure and one or more of the content elements and/or one or more social network metrics.