Content Delivery Strategy Selection Using Effect Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for optimizing the selection of styles for digital contents, such as advertisements, rely on Exploration-Exploitation, click-through-rate, or conversion rate-based prediction models, which fail to utilize user and content features effectively, leading to poor generalization and inability to quantify the incremental impact of styles on user behavior.
Innovation Solution
A method for content delivery that involves receiving a request to deliver a target content to a target object, determining observed samples based on target features, calculating an effect metric for each content delivery strategy, and selecting the optimal strategy based on these metrics to refine and personalize content delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If Exploration-Exploitation or click-through-rate based prediction models are used for optimizing content styles, then the implementation is simple, but the generalization performance is poor and user features are not utilized effectively
Solution Approach 1:
The patent transforms the optimization approach from simple prediction models to a structured framework that incorporates multiple parameters including user features, content features, and contextual information. The effect metric calculation changes the evaluation parameters from basic click-through rates to comprehensive metrics that capture incremental impact, thereby improving generalization performance while maintaining implementation feasibility through systematic parameter organization.
2Device complexity
If conventional prediction models are used, then the system complexity is low, but the ability to quantify incremental impact of styles on user behavior is insufficient
Solution Approach 1:
The patent introduces an effect metric as an intermediary that bridges the gap between simple prediction models and comprehensive impact analysis. This intermediary component processes user features, content features, and contextual information to produce a quantified measure of incremental impact, thereby enhancing measurement precision without requiring complete system redesign.
Solution Approach 2:
The patent segments the content delivery optimization into distinct components: user feature extraction, content feature extraction, contextual information processing, and effect metric calculation. This segmentation allows each component to be developed and optimized independently, managing system complexity while achieving precise impact quantification through the integrated effect metric.
3Speed
If content delivery optimization does not utilize user and content features, then the processing speed is fast, but the personalization effectiveness is poor
Solution Approach 1:
The patent performs preliminary extraction and processing of user features and content features before the actual content delivery decision. By pre-processing these features and organizing them into structured formats, the system maintains fast processing speed during deployment while achieving effective personalization through the comprehensive effect metric that leverages these pre-extracted features.
Data Source
AI summary
A method, device and storage medium for content delivery are provided. In the method, a request to deliver a target content to a target object is received. For a content delivery strategy of a plurality of content delivery strategies, at least one group of observed samples are determined based on target features of the target content and the target object, an observed sample comprising reference features of a reference content and a reference object, an indication of whether the content delivery strategy is applied to deliver the reference content and an object response. An effect metric is determined for the content delivery strategy based on the at least one group of observed samples. A target content delivery strategy for delivering the target content to the target object is selected from the plurality of content delivery strategies based on respective effect metrics for the plurality of content delivery strategies.