Search Recommendation Ranking Using Multi-Time Search Volume Sequences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information recommendation systems inaccurately determine target recommendation information due to high scores given to candidate features with low search volumes, leading to reduced search efficiency as users need to input multiple times to find desired information.
Innovation Solution
Determine sequence feature information of a candidate recommendation information set based on search volumes within different time limitations, using a trained target ranking model to accurately identify target recommendation information, enriching the sequence feature information across various time frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search recommendation systems prioritize information based on traditional ranking methods, then the system operation is simple, but the recommendation accuracy deteriorates due to high scores being given to low search volume content
Solution Approach 1:
The patent segments the search volume analysis into multiple time limitations (e.g., 24 hours, 7 days, 30 days). Instead of using a single overall search volume metric, the system divides the time dimension into discrete segments and calculates sequence feature information for each segment. This segmentation allows the system to capture temporal patterns in search behavior while maintaining manageable computational complexity through structured processing of divided time periods.
Solution Approach 2:
The patent introduces a temporal dimension by analyzing search volumes across different time limitations rather than using a single static metric. The sequence feature information captures the evolution of search volumes over time, transforming a one-dimensional ranking problem into a multi-dimensional analysis that incorporates time-based patterns. This dimensional expansion improves recommendation accuracy by considering when searches occur, not just how many times they occur.
2Measurement precision
If the system analyzes search volumes across multiple time limitations to improve recommendation accuracy, then the recommendation precision improves, but the calculation complexity increases
Solution Approach 1:
The patent changes the parameter representation by introducing sequence feature information that transforms raw search volume counts into normalized temporal patterns. Instead of directly using absolute search volume numbers across multiple time periods, the system calculates relative changes and sequences that are scaled and normalized. This parameter transformation reduces calculation complexity by working with standardized values rather than raw counts that vary widely in magnitude across different time limitations.
Solution Approach 2:
The patent creates simplified copies of the search volume data by generating sequence feature information that represents temporal patterns in a standardized format. Rather than processing the full complexity of raw search volume data across multiple time dimensions, the system creates derived copies (sequence features) that capture essential temporal relationships in a more manageable representation, reducing the computational burden while preserving the information needed for accurate recommendations.
3Productivity
If traditional ranking methods are used without temporal analysis, then the processing speed is fast, but the user search efficiency deteriorates due to multiple input attempts needed
Solution Approach 1:
The patent performs preliminary analysis of search volume patterns across multiple time limitations before generating recommendations. By pre-calculating sequence feature information that captures temporal trends in search behavior, the system prepares advance insights about which information items are likely to be relevant. This preliminary temporal analysis enables the system to make more accurate predictions about user needs, reducing the number of search attempts required and improving overall search efficiency.
Data Source
AI summary
An information recommendation method, an information recommendation apparatus, an electronic device, and a storage medium are provided. The information recommendation method includes: obtaining a candidate recommendation information set which matches current search information, the candidate recommendation information set including multiple candidate recommendation information; determining, according to search volumes of the multiple candidate recommendation information within different search time limitations, sequence feature information of the candidate recommendation information set under each search time limitation, the sequence feature information being used to reflect an overall search feature corresponding to the candidate recommendation information set; and determining target recommendation information based on a trained target ranking model and the sequence feature information of the candidate recommendation information set under each search time limitation.