Unified Ranking Score for Native Apps and Web Resources
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Search engines face challenges in consistently ranking native applications and web page resources for user queries, as they often use different scoring signals, leading to inconsistent relevance scores for deep-linked content.
Innovation Solution
A method to determine universal ranking scores for native applications with and without corresponding web resources, allowing for a combined ranking that preserves individual rankings, by using position and total count in rankings, and optionally Fibonacci numbers, to merge search result rankings and provide more relevant search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If separate scoring signals are used for native applications with web resources and without web resources, then each type can be ranked according to its specific characteristics, but the relevance scores become inconsistent and cannot be directly compared across different native application types
Solution Approach 1:
The patent introduces a mediator component that converts different scoring signals from separate ranking systems into a common relevance score scale. This mediator translates scores from the first ranking system (for native applications with web resources) and the second ranking system (for native applications without web resources) into comparable relevance scores, enabling consistent comparison while preserving the adaptability of separate scoring approaches
Solution Approach 2:
The patent creates a universal relevance score that can represent both types of native applications on a common scale. This universal scoring mechanism allows the system to handle multiple types of search results (native applications with web resources, native applications without web resources, and web page resources) using a single consistent relevance metric, thereby resolving the inconsistency issue
2Productivity
If multiple search result rankings are merged into a single combined ranking, then users receive comprehensive search results covering all relevant resources, but the individual ranking integrity of each source may be compromised
Solution Approach 1:
The patent maintains separate ranking segments for different types of search results (native applications with web resources, native applications without web resources, and web page resources) while introducing a mapping mechanism that preserves the relative ordering within each segment. This segmentation approach allows comprehensive merging without compromising the integrity of individual ranking sources
3Adaptability or versatility
If different scoring signals are used for different types of native applications, then each type can be optimized for its specific characteristics, but the system complexity increases due to multiple ranking mechanisms
Solution Approach 1:
The patent introduces a mediator component that handles the complexity of multiple scoring signals by translating them into a common relevance score format. This intermediary layer allows the system to maintain specialized scoring mechanisms for different native application types while presenting a unified, simplified interface for combining and comparing results, thereby reducing overall system complexity
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for ranking native applications. In one aspect, a method includes determining, for a given query, a first ranking of a set of first native applications that each have a respective corresponding web resource. For each particular first native application in the first ranking, a first universal ranking score is determined based on a position of the particular first native application in the first ranking and a total number of first native applications in the first ranking. A second ranking of a set of second native applications that do not have a corresponding web resource is determined. For each particular second native application, a second universal ranking score is determined based on a position of the particular second native application in the second ranking and a total number of second native applications in the second ranking.