Data feature screening method and device and computer equipment

Abstract

According to a data feature screening method and device and computer equipment provided by the embodiment of the invention, most features with relatively large trend fluctuation are eliminated by utilizing the correlation coefficient during screening, and the whole process only relates to correlation calculation and is not required to be visualized, so that the overall speed is higher, and the efficiency is higher. Through screening of the whole process, the number of variables finally entering the model is reduced, the complexity of the model is simplified, the overall service cost is reduced, and the interpretability of the model is improved. Therefore, the correlation between a binning value and the corresponding positive example ratio is calculated by utilizing a binning result table in the IV calculation process, the characteristic trend can be measured, and then the characteristic trend is combined with the IV value to perform characteristic screening, so that the characteristictrend to be molded is good, the measurement is accurate and the interpretability is high under the condition of reducing the time consumption and the computer resource consumption,.

Description

technical field [0001] The present invention relates to the technical field of data feature processing, in particular, to a data feature screening method, device and computer equipment. Background technique [0002] In the process of building a risk control model, an important step is feature engineering. Feature engineering refers to the process of using professional background knowledge and skills to process data to generate features that can better describe the data, and using these features to make machine learning algorithms play a better role. The process includes modules such as feature extraction, feature construction, and feature screening. [0003] The linear model represented by logistic regression is widely used in the industry as a model with strong interpretability. In the feature screening module of linear model modeling, the usual practice is to first filter the features through EDA (exploratory data analysis), such as filtering out high missing rate, const...

AI Technical Summary

Problems solved by technology

[0004] However, the above technologies have the problems of taking too long when screening feature combinations, occupying too many resources, and measuring a single feature inaccurately.

Images