Dynamic Prefix Length Inversion Index for Search Input Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data processing methods for input recommendations in search engines result in a large number of redundant index entries due to the vast number of prefixes, leading to inefficient use of index entries and increased computational costs.
Innovation Solution
The method involves generating prefixes of varying lengths, combining prefixes with the same content, determining the number of entries each prefix points to, and truncating or updating prefix lengths to ensure that only relevant entries are included in the index, thereby reducing redundancy and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all prefixes for all recommended entries are listed to create a comprehensive inverted index, then the index coverage is improved, but the number of index entries becomes enormous and redundancy increases
Solution Approach 1:
The patent extracts only the necessary prefixes that meet the threshold criteria (entry count >= N and length <= max_prefix_length) from the complete set of all possible prefixes. This selective extraction removes redundant index entries while preserving the coverage needed for effective input recommendations.
Solution Approach 2:
The patent applies different quality criteria to different prefixes: only prefixes with sufficient entry coverage (>= N entries) and appropriate length (<= max_prefix_length) are included in the inverted index. This local quality filtering ensures that each included prefix entry serves a useful purpose, eliminating uniformly redundant entries.
2Adaptability or versatility
If a comprehensive inverted index with all prefixes is created, then all possible input recommendations can be supported, but the computational cost and processing time increase
Solution Approach 1:
The patent performs preliminary filtering of prefixes during the offline index generation phase, applying the N-threshold and max-prefix-length criteria before creating the inverted index. This preliminary action reduces the amount of data that needs to be processed and stored, thereby reducing computational cost and processing time while maintaining recommendation coverage.
Solution Approach 2:
The patent changes the parameters of the inverted index by introducing two key constraints: minimum entry count (N) and maximum prefix length. These parameter changes transform the index from a comprehensive but inefficient structure to a optimized structure that balances coverage with computational efficiency.
3Measurement precision
If longer prefixes are included in the inverted index, then more precise input recommendations can be provided, but the number of rarely used index entries increases
Solution Approach 1:
The patent applies a local quality constraint on prefix length (max_prefix_length) to ensure that only prefixes of appropriate length are included. This prevents the inclusion of overly long prefixes that would provide minimal additional precision but would consume excessive computational resources and create rarely used index entries.
Solution Approach 2:
The patent introduces the max_prefix_length parameter to control the maximum length of prefixes in the inverted index. This parameter change balances recommendation precision with computational efficiency by preventing the inclusion of excessively long prefixes that would be rarely used and consume unnecessary resources.
Data Source
AI summary
A data processing method of establishing an input recommendation is provided. The method establishes an inverted index having a variable prefix length based on entry number distribution, thus preventing the established inverted index from generating a number of index entries that are rarely used, and reducing the redundancy of the input recommendation index that is created.