确定业务预测模型的方法和装置
By acquiring new data blocks to generate a second base classifier group and detecting the error rate, an ensemble classifier is formed, which solves the problem of reduced prediction accuracy in business prediction models for concept drift types, and achieves rapid model adaptation and improved accuracy.
CN117093837BActive Publication Date: 2026-07-17JINGDONG TECH HLDG CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- JINGDONG TECH HLDG CO LTD
- Filing Date
- 2023-08-21
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
When dealing with business scenarios characterized by concept drift, existing technologies suffer from declining accuracy in business forecasting models over time, slow update speeds, and difficulty in adapting to drastic changes in business data.
Method used
By acquiring new data blocks for the target business, a second set of base classifiers is generated, and the error rate of each base classifier is detected to form a first or second ensemble classifier to adapt to changes in new business data and improve prediction accuracy.
Benefits of technology
This enables the business forecasting model to adapt to new business data changes in a timely manner, thereby improving forecast accuracy.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117093837B_ABST
Abstract
本公开提出一种确定业务预测模型的方法和装置，涉及大数据挖掘领域。该方法包括：获取目标业务在最新时间段的新数据块；获取至少一个第一基分类器组，每个第一基分类器组基于一个非最新时间段的旧数据块生成，且包括多个第一基分类器；基于新数据块生成第二基分类器组，其包括多个第二基分类器；利用新数据块检测每个基分类器预测的错误率；在错误率不满足要求的基分类器的总数量超过数量阈值的情况下，基于第二基分类器组形成第一集成分类器，作为目标业务的业务预测模型。在判断出业务数据分布剧烈变化的情况下，全部基于新基分类器形成业务预测模型，使得业务预测模型能够及时适应新的业务数据变化，提高预测准确性。
Need to check novelty before this filing date? Find Prior Art