Content Recommendation Using Pre-generated Sequences for Low-Click Users
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content recommendation methods struggle with recommending content to users who have limited click behaviors, as their historical click data is insufficient for effective recommendations.
Innovation Solution
A content recommendation method and device that uses feedback information from users to determine a sequence of preferred contents, employing an average recommendation probability and the roulette wheel selection algorithm to recommend content, even for users with less click behavior, by collecting and processing real-time feedback data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If content recommendation is performed based on historical click data, then recommendation accuracy can be improved for users with sufficient click history, but the method becomes ineffective for random users who have limited click behaviors
Solution Approach 1:
The system pre-generates a sequence of preferred contents for each user before actual recommendation needs arise. This preliminary action stores potential recommendation candidates based on user profiles and content characteristics, so when a user with limited click history needs recommendations, the system can immediately provide pre-prepared content sequences without relying on insufficient historical data.
Solution Approach 2:
The patent introduces an intermediary mechanism (the pre-generated sequence of preferred contents) that bridges the gap between users with limited click history and the recommendation system. This intermediary stores potential recommendation matches that don't require extensive user interaction data, allowing the system to serve both user groups effectively.
2Measurement precision
If the system waits for sufficient user feedback data before making recommendations, then recommendation quality can be improved, but user engagement and system productivity decrease due to delayed recommendations
Solution Approach 1:
The system performs preliminary generation of content sequences in advance, storing them for immediate retrieval. This eliminates the need to wait for sufficient user feedback before making recommendations, as the preferred content sequences are already prepared based on user profiles and content characteristics, enabling both high quality and fast delivery.
Solution Approach 2:
The system prepares recommendation sequences beforehand as a cushion or buffer, so when recommendation requests occur, pre-computed results are immediately available. This beforehand cushioning ensures that recommendation quality is maintained while eliminating delivery delays.
Data Source
AI summary
The present disclosure relates to a content recommendation method, device and system. The method includes: recommending a content in a content set to a user based on an average recommendation probability; collecting feedback information on the recommended content from the user's client, wherein the feedback information includes display information and click information, the display information including displaying times and displaying timing of the recommended content on the client, and the click information including clicking times and clicking timing of the recommended content on the client; and determining a sequence of preferred contents from the contents in the content set according to the feedback information, so as to recommend a content to the user based on the sequence of preferred contents. Respective aspects of the present disclosure recommend contents based on a user's feedback so that it is possible to perform reasonable and effective content recommendation for random users who have less click behaviors.