Multi-Target Event Prediction Using Target-Specific Historical Feature Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-target prediction methods are inaccurate due to the direct use of historical behavior data sequences without considering differences between prediction targets, leading to incomplete and inaccurate prediction results.
Innovation Solution
A multi-target event prediction method that generates historical feature data for each prediction target based on correlations between historical behavior codes and prediction targets, allowing for more accurate prediction results by considering differences between targets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If historical behavior data sequence is directly used as a whole to predict service effect, then the prediction process is simple, but the prediction accuracy is low
Solution Approach 1:
The patent segments the historical behavior data sequence into multiple discrete behavior codes (e.g., browsing, cart addition, purchase). Each behavior code is independently encoded and processed, allowing the system to capture specific behavioral patterns rather than treating the sequence as a monolithic entity. This segmentation enables more precise feature extraction for prediction.
Solution Approach 2:
The patent applies different encoding strategies and feature extraction methods to different types of behavior codes based on their specific characteristics. For example, browsing behaviors may be weighted differently from purchase behaviors, and each behavior type undergoes specialized processing to highlight its unique predictive value for the specific prediction target.
2Measurement precision
If target-specific historical feature data is generated for each prediction target, then the prediction accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent pre-computes and stores the encoding mappings for all possible behavior codes in advance. During the prediction phase, the system simply looks up pre-defined encoding rules and applies them to the historical behavior sequence, avoiding the need to perform complex computations in real-time. This preliminary preparation significantly reduces online computational complexity.
Solution Approach 2:
The patent dynamically adjusts encoding parameters and feature extraction configurations based on the specific prediction target. Different prediction targets (e.g., click-through rate, conversion rate, purchase amount) utilize different parameter settings and weighting schemes, allowing the system to optimize prediction accuracy for each target without requiring completely separate processing pipelines.
Data Source
AI summary
This application provides a multi-target event prediction method performed by a computer device. The method includes: obtaining a historical behavior code sequence according to a historical behavior data sequence of a target object. Prediction results of the predicted event associated with the target object are respectively obtained from dimensions of the various prediction targets on the basis of event information of the predicted event in combination with the historical feature data separately corresponding to the various prediction targets. Differences in the historical feature data of the target object between different prediction targets are considered during the obtaining of the prediction results, so that the obtained prediction results can reflect differences of the target object for the predicted event under the different prediction targets, thereby improving the accuracy and comprehensiveness of the prediction results.