SERP Ranking Adjustment via Pair-Specific Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional search engine ranking systems often fail to adequately rank documents that are relevant to user queries, leading to a decreased usefulness of Search Engine Results Pages (SERPs) as some useful search results are ranked low or not ranked correctly.
Innovation Solution
A method and server that utilize a rank-adjustment score to re-rank relevant documents on SERPs based on pair-specific values and pair-wise similarity with stored search pairs, boosting baseline ranking scores to improve the relevance and usefulness of search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional ranking algorithms are used to rank search results, then the ranking system can process queries efficiently, but some useful documents may be ranked low or not adequately ranked
Solution Approach 1:
The system pre-calculates and stores pair-specific values for query-document pairs in advance. When a new query comes in, the system retrieves these pre-computed values and uses them to adjust rankings, avoiding the need to re-evaluate all documents from scratch. This preliminary action enables both efficient processing and improved ranking accuracy.
Solution Approach 2:
The system computes pair-specific values that represent the usefulness of documents for specific queries, and uses these values as feedback to adjust the initial ranking. The rank adjustment module incorporates this feedback by adding the pair-specific value to the initial ranking score, thereby refining the final document ranking to better reflect actual usefulness.
2Reliability
If the system re-ranks all documents using pair-specific values, then the usefulness of SERP is improved, but the computational complexity increases
Solution Approach 1:
Instead of uniformly processing all documents with the same level of detail, the system applies pair-specific adjustments only to documents that have been pre-evaluated and stored in the database. This local quality approach allows the system to focus computational resources on documents where pair-specific values are available, improving SERP usefulness without proportionally increasing overall system complexity.
Solution Approach 2:
The system pre-computes and stores pair-specific values for query-document pairs in advance. When a new query comes in, the system retrieves these pre-computed values and uses them to adjust rankings, avoiding the need to re-evaluate all documents from scratch. This preliminary action enables both efficient processing and improved ranking accuracy.
3Measurement precision
If pair-wise similarity is computed between current and stored search pairs, then rank-adjustment scores can be generated accurately, but the time required for processing increases
Solution Approach 1:
The system pre-calculates and stores pair-specific values for query-document pairs in advance. When a new query comes in, the system retrieves these pre-computed values and uses them to adjust rankings, avoiding the need to re-evaluate all documents from scratch. This preliminary action enables both efficient processing and improved ranking accuracy.
Solution Approach 2:
The system computes pair-wise similarity between current and stored search pairs as a parameter to determine rank-adjustment scores. By using similarity as a parameter, the system can accurately estimate the usefulness of documents for new queries based on historical data, balancing precision with processing efficiency.
Data Source
AI summary
A method and a server for ranking documents in response to a current query are disclosed. The documents are to be presented on a SERP. A database stores stored search pairs in association with respective pair-specific values. The method comprises for the current query, ranking, by a MLA, relevant documents to be included in the SERP which have preliminary ranks. The current query and a respective relevant document form a current search pair. The method comprises, for a given current search pair, generating a rank-adjustment score associated with a stored search pair based on: the pair-specific value of the stored search pair, and a pair-wise similarity between the current search pair and the stored search pair. The method comprises, for the current query, re-ranking a relevant document of the current search pair on the SERP using the associated rank-adjustment score.