Selection-Based Information Recommendation Without User History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information recommendation systems struggle to accurately recommend content to users without prior history data, resulting in ineffective information delivery.
Innovation Solution
A method and apparatus that dynamically adjust recommended information based on user selection, generating selection recommendation pages by determining selected object information and associated recommendation objects, and presenting these dynamically to enhance relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional information recommendation systems are used without user history data, then the system can operate with simple structure, but the recommendation accuracy deteriorates
Solution Approach 1:
The system performs preliminary actions by pre-calculating and storing similarity relationships between different types of objects (songs, singers, albums, playlists) in a knowledge graph. This advance preparation enables accurate recommendations even for new users without requiring complex real-time analysis, thus improving recommendation accuracy while maintaining relatively simple system structure during operation.
Solution Approach 2:
The patent introduces a knowledge graph as an intermediary structure that stores pre-computed similarity relationships between various music objects. This intermediary enables the recommendation system to accurately connect users with relevant content without requiring complex real-time computation or extensive user history data, resolving the contradiction between accuracy and complexity.
2Reliability
If dynamic recommendation pages are generated based on user selection, then the recommendation effectiveness improves, but the processing time increases
Solution Approach 1:
The system pre-calculates and stores similarity scores between all pairs of music objects in a knowledge graph before runtime. When generating dynamic recommendation pages, the system only needs to query these pre-computed similarities rather than performing complex real-time analysis, thus maintaining high recommendation effectiveness while minimizing processing time for dynamic page generation.
Solution Approach 2:
The patent applies local quality by computing similarities locally for specific object pairs in advance and storing them in the knowledge graph. This localized pre-computation approach allows the system to quickly retrieve relevant similarity data during dynamic recommendation generation without performing global, time-consuming analysis, thereby balancing effectiveness and processing speed.
3Ease of operation
If multiple selection recommendation controls are generated, then the user interaction quality improves, but the device complexity increases
Solution Approach 1:
The patent segments the recommendation system into distinct functional modules: a knowledge graph construction module, a similarity calculation module, and a recommendation page generation module. Each module handles specific tasks independently, making the overall system easier to manage and maintain despite generating multiple complex recommendation controls. The segmentation allows parallel processing and reduces inter-module dependencies.
Solution Approach 2:
The knowledge graph serves as a universal data structure that stores similarity relationships between multiple types of music objects (songs, singers, albums, playlists). This multi-functional structure enables the system to generate various types of selection recommendation controls using the same underlying mechanism, improving user interaction quality without proportionally increasing system complexity through code duplication.
Data Source
AI summary
The disclosure discloses a method, an apparatus, an electronic device and storage medium for information recommendation. The method of information recommendation includes: in response to a detected object selection operation, determining selected object information corresponding to a selected object identification and an associated recommendation information of at least one associated recommendation object; generating a plurality of selection presenting controls based on the selected object information and the associated recommendation information, determining a plurality of selection recommendation controls based on the plurality of selection presenting controls and a current display control in a current display page; and generating, based on the plurality of selection recommendation controls, a selection recommendation page to be dynamically presented, and presenting the selection recommendation page dynamically.