A default user probability prediction method based on sparse feature embedding
A sparse feature and probability prediction technology, applied in special data processing applications, data processing applications, resources, etc., can solve problems such as unsatisfactory sparse data processing effects, and achieve benefits for learning and processing, improving processing capabilities, and reducing dimensions Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach
[0060] In this part, the original features obtained through data cleaning in the first step are processed by feature engineering and converted into training data for the model. Firstly, the original features are converted into new features through the traditional feature engineering method, and then part of the sparse features in the new features are converted into one-dimensional variables through the multi-category variable method proposed by the present invention, which can be directly used as the training data of the model. The specific implementation is as follows:
[0061] 2.1 Feature Engineering
[0062] The variables in the original features are subjected to feature extraction and variable derivation according to the time class, amount class, address class, and phone number class. The time field includes authentication time, loan time, shopping time, etc. 00-24:00) and early morning (00:00-6:00), and count the number and proportion of orders in these four time period...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com