Feature data-based attribution analysis method and apparatus, and electronic device
By generating training data and using the XGB tree model to analyze feature indicators, the key features of the swapped-in and swapped-out data are determined, which solves the problem of lack of attribution analysis in existing technologies and improves the reliability and accuracy of the model.
CN122262933APending Publication Date: 2026-06-23JINGDONG TECH HLDG CO LTD
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202411884750.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-19
- Publication Date
- 2026-06-23
AI Technical Summary
Technical Problem
Existing technologies lack accurate attribution analysis of swap-in and swap-out data, making it impossible to determine the characteristics that cause swap-in and swap-out phenomena.
Method used
By generating training data and using a pre-defined classification model such as the XGB tree model, the indicator data of the input features of the original decision model are analyzed, target features are selected, and key features are determined based on the target features, the input dataset, and the output dataset, thereby adjusting and monitoring the model.
Benefits of technology
It enables the analysis of key features of swapped-in and swapped-out data, reduces swapped-in and swapped-out phenomena, and improves the reliability and accuracy of decision-making models.
✦ Generated by Eureka AI based on patent content.
Abstract
The embodiment of the invention discloses an attribution analysis method and device based on feature data and electronic equipment. According to one specific embodiment, the method comprises the steps that training data are generated according to feature data, belonging to model-in features, in a swap-in dataset and a swap-out dataset, and the model-in features are model-in features of an original decision-making model with a swap-in and swap-out phenomenon; training a preset classification model through the training data, and determining index data of each model entering feature of the original decision model based on the trained preset classification model; according to the determined index data, selecting a set number of modulo-entering features from each modulo-entering feature of the original decision model to serve as target features; and based on the target features, the swap-in data set and the swap-out data set, key features enabling the original decision model to generate swap-in and swap-out data are analyzed. The implementation mode is related to a model bottom layer analysis technology, and key characteristics causing swap-in and swap-out are revealed by analyzing the in-model characteristics of an original decision model.
Need to check novelty before this filing date? Find Prior Art