Statistical Prediction Model Sharing for Campaign Clones
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content delivery campaign clones lack data on past performance, leading to potential worse or better initial performance compared to the original campaign, as they do not inherit the statistical prediction models used by the original campaign.
Innovation Solution
A system and method for leveraging statistical prediction models by associating a clone of a content delivery campaign with the prediction model of the original campaign, allowing the prediction model to generate a predicted user selection rate for the clone, and sharing this model under specific criteria such as time, targeting criteria, and charging models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a campaign is cloned to enable quick deployment, then deployment speed is improved, but the clone lacks past performance data leading to reduced prediction accuracy
Solution Approach 1:
The system performs preliminary action by pre-generating statistical prediction models for the original campaign before cloning. When a campaign is cloned, the pre-existing prediction model is automatically associated with the clone, eliminating the need for the clone to start from scratch and collect its own performance data. This allows the clone to immediately benefit from the original campaign's historical performance insights.
Solution Approach 2:
The system applies the copying principle by replicating the statistical prediction model from the original campaign to the cloned campaign. Instead of creating a new prediction model for each clone, the system copies the existing model's structure, parameters, and learned patterns, allowing the clone to inherit the predictive capabilities developed during the original campaign's performance.
2Stability of the object's composition
If prediction models are shared across campaign clones, then performance consistency is improved, but the risk of malicious boosting increases
Solution Approach 1:
The system implements feedback mechanisms by continuously monitoring campaign performance and using this information to dynamically adjust or update prediction models. When anomalies or suspicious patterns are detected in cloned campaigns, the feedback loop allows the system to identify potential malicious boosting and respond by adjusting the model association or triggering validation processes.
Solution Approach 2:
The system applies dynamics by making the prediction model association flexible and adaptive rather than static. The system can dynamically determine whether to share prediction models based on campaign performance, detect changes in campaign behavior, and adjust the level of model sharing accordingly. This allows the system to maintain performance consistency while responding to changing conditions that may indicate malicious activity.
Data Source
AI summary
Techniques for leveraging existing statistical prediction models are provided. A first statistical prediction model is generated for a content item. An instruction is received to create a clone from the content item. In response to receiving the instruction, the clone is created based on attributes of the content item. A second statistical prediction model that is different than the first statistical prediction model is generated for the clone. In response to receiving a request for content, the clone is identified as relevant to the first request. A similarity between (1) first content of the content item and (2) second content of the clone is determined. If the similarity exceeds a similarity threshold, then the first statistical prediction model is used to generate a prediction of an entity user selection rate associated with the clone. Otherwise, the second statistical prediction model is used to generate the prediction.