Search Query Tag Extraction for Diverse Content Recommendation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content recommendation systems rely on traditional methods that result in homogenized and less diverse recommended content items, failing to accurately reflect user preferences due to reliance on object names and interaction metrics.
Innovation Solution
A method that extracts tags from search queries associated with a target object, using machine learning models and clustering algorithms to determine relevant recommended content items based on these tags, ensuring better matching with user interests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional recommendation methods based on object names and interaction metrics are used, then the recommendation system is simple to implement, but the recommended content items become homogenized and fail to accurately reflect user preferences
Solution Approach 1:
The patent introduces search queries as an intermediary element between user behavior and content recommendation. Instead of directly using interaction metrics, the system extracts tags from search queries that users perform on or related to target objects. These tags serve as a mediator that captures genuine user intent and preferences, thereby improving recommendation accuracy while maintaining reasonable system complexity through established NLP techniques.
Solution Approach 2:
The patent transforms the recommendation approach by changing the parameter used for matching from traditional interaction metrics to tags extracted from search queries. This parameter change enables the system to capture semantic meaning and user intent more accurately. The extraction process converts unstructured search query text into structured tags that can be efficiently matched with content attributes, resolving the contradiction between accuracy and complexity.
2Adaptability or versatility
If tags are extracted from search queries using machine learning models and clustering algorithms, then the diversity and quality of recommended content items improve, but the processing time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-processing and storing tags extracted from search queries in advance. Instead of performing complex machine learning extraction and clustering at recommendation time, the system pre-processes search query data to generate and store relevant tags. This preliminary preparation significantly reduces the computational burden and processing time during actual recommendation operations, while still achieving diverse and high-quality content recommendations.
Data Source
AI summary
According to embodiments of the disclosure, a method, apparatus, device, medium, and program product for content searching are provided. The method includes: obtaining a set of search queries associated with a target object; extracting a set of tags for the target object from the set of search queries, each of the set of tags indicating a keyword related to the target object in a corresponding search query; and determining, from recommended content items associated with the target object, one or more recommended content items matching one or more tags of the set of tags, to be provided to a target user group.