Click Probability Modeling for Accidental Click Content Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to accurately distinguish between intentional and accidental click events when determining click probabilities for content items, leading to inaccurate content selection and presentation.
Innovation Solution
A machine learning model is trained using event information that differentiates between intentional and accidental click events, incorporating accidental click probabilities to improve the accuracy of click probability determination and content selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing systems treat all click events as intentional, then content selection is simplified, but accuracy of click probability determination deteriorates
Solution Approach 1:
The patent segments click events into distinct categories (intentional clicks, accidental clicks, and skip events) based on user interaction patterns. By dividing the homogeneous set of all clicks into heterogeneous subgroups with different characteristics, the system can apply different weighting and analysis methods to each segment, thereby improving click probability determination accuracy while managing complexity through structured classification.
Solution Approach 2:
The patent introduces machine learning models as intermediary components that analyze event information and determine whether a click is intentional or accidental. These models act as mediators between raw click data and content selection decisions, automatically classifying events based on learned patterns from training data, thus improving accuracy without requiring complex manual classification rules.
2Measurement precision
If accidental click events are included in training data without differentiation, then training data volume increases, but model accuracy deteriorates
Solution Approach 1:
The patent extracts and separates accidental click events from the overall training dataset, creating distinct subsets for intentional clicks and accidental clicks. By removing the contaminating effect of unlabeled mixed data and creating purified training subsets with known characteristics, the model can learn distinct patterns for each event type, significantly improving prediction accuracy while maintaining sufficient data volume through targeted sampling.
Solution Approach 2:
The patent performs preliminary classification and labeling of training events before model training, pre-processing the data to assign accurate labels indicating whether each click is intentional or accidental. This preliminary action ensures that the training data is properly structured and labeled before being fed to the model, preventing accuracy degradation from mislabeled or ambiguous data while maintaining comprehensive data coverage.
Data Source
AI summary
In an example, sets of event information associated with events may be identified. The events may include intentional click events, accidental click events and/or skip events. Accidental click probabilities associated with the accidental click events and/or the skip events may be determined. Machine learning model training may be performed, using the sets of event information associated with the events and labels associated with the events, to generate a first machine learning model. The labels may include second labels associated with the intentional click events and/or third labels associated with the accidental click events and/or the skip events. The second labels may correspond to an intentional click classification. The third labels may be based upon the accidental click probabilities. Click probabilities associated with content items may be determined using the first machine learning model. A content item may be selected for presentation via a client device based upon the click probabilities.