Intersection-based automatic feature generation method

Abstract

The invention relates to an intersection-based automatic feature generation method, which comprises the steps of S1, for a to-be-processed data table, performing binarization according to the type ofdata in the data table and converting binary features; S2, performing iterative feature intersection on the generated binary features to generate intersection features, including S21, calculating a plurality of feature evaluation indexes based on the binary features; S22, according to the specified feature generation number and the specified iteration round number, calculating the number m of features needing to be reserved in each round, the number n of cross features and the number k of features to be crossed; S23, selecting k binary features from the generated binary features, and selectingn cross features from the cross features generated by the last iteration; S24, performing cross operation on the k binary features and the n cross features in pairs to generate new cross features; and S25, selecting m cross features from the newly generated cross features, and reserving the m cross features as the cross features generated by the current iteration. Feature development efficiency of the user is greatly improved.

Description

technical field [0001] The present invention relates to the technical field of machine learning, and more specifically, relates to an automatic feature generation method based on intersection. Background technique [0002] With the emergence of massive data, people tend to use machine learning techniques to build models to solve practical problems. The basic process of training a machine learning model mainly includes: 1) clarifying the modeling goal and collecting available data; 2) feature generation and feature selection; 3) building a model; 4) evaluating the effect of the model. In the above process, the feature generation process is very important, and the quality of feature generation determines the upper limit of the model. [0003] Currently, the feature generation methods are as follows: [0004] 1) Artificial feature generation [0005] Technicians develop features through the cleaning and screening of underlying data, the design of feature logic, and the devel...

AI Technical Summary

Problems solved by technology

It does not require technicians to develop code, but requires technicians to have a deep understanding of business and data. The effect of features depends on the business experience of technicians, and when faced with a large amount of data, it still takes a lot of time

Images