Method and device for selecting features for a constructed machine learning model
A machine learning model and model technology, applied in the computer field, can solve problems such as low efficiency and achieve the effect of improving effectiveness
Active Publication Date: 2021-06-04
ADVANCED NEW TECH CO LTD
View PDF9 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
This feature selection method needs to calculate AUC for each feature one by one through the single-feature AUC evaluation module when screening features. If the number of features is large, this feature selection method has low timeliness
At the same time, the scope of application of this feature selection method has certain limitations, such as the inability to calculate the feature AUC value for regression problems, etc.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0031] The solutions provided in this specification will be described below in conjunction with the accompanying drawings.
[0032] figure 1It is a schematic diagram of an application scenario of the embodiment of this specification. In this application scenario, it mainly includes data module, feature selection module and model training module. The method of selecting features for the constructed machine learning model provided in this manual is mainly applicable to figure 1 The feature selection module in . The data module, feature selection module and model training module can be set on the same computing platform (such as the same server or server cluster), or they can be set on different computing platforms, which is not limited here. Wherein, the data module may include, for example, various storage media for storing training data sets.
[0033] The basic idea of the embodiments of this specification is based on random disturbance. It can be understood that if a f...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The embodiment of this specification provides a method and device for selecting features for a constructed machine learning model. According to an embodiment of the method, m sample data pairs are first obtained, and then m sample data pairs are randomly perturbed to analyze feature importance. Specifically, on the one hand, m sample data pairs are used to train the machine learning model, so as to obtain the first importance degree of the first feature through the trained machine learning model; on the other hand, the sample labels of the m sample data pairs are randomly exchange, and use the m sample data pairs after the sample labels are randomly exchanged to train the machine learning model, so as to obtain the second importance degree of the first feature through the trained machine learning model. Further, the first importance degree and the second importance degree of each feature are compared, and features are selected for the constructed machine learning model according to the comparison result. This embodiment can improve the effectiveness of feature selection.
Description
technical field [0001] One or more embodiments of this specification relate to the field of computer technology, and in particular to a method and device for selecting features for a machine learning model constructed by a computer. Background technique [0002] In order to build a machine learning model with optimal performance, it is usually necessary to manually select many dimensions of features (also called variables) based on business experience and understanding of data. If the features selected by this process are inappropriate, they may not have much value for the machine learning model to be built, or even have a negative effect. Therefore, in the process of building a machine learning model, it is necessary to continuously experiment and perform feature screening to build a better machine learning model. For models that have been running online for a period of time, new elements may also be added, making the model unable to predict correctly, resulting in model d...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62
CPCG06F18/214
Inventor 易灿许辽萨王维强
Owner ADVANCED NEW TECH CO LTD