Classification prediction method and device and prediction model training method and device

Abstract

The invention provides a method and a device for performing classification prediction based on a prediction model, and a method and a device for training the prediction model. The classification prediction method includes: starting from a first prediction sub-model in the prediction sub-model chain, executing the following prediction process aiming at at least one to-be-predicted sample until a predetermined condition is met: inputting the current to-be-predicted sample into the current prediction sub-model to obtain a prediction classification result and a corresponding prediction confidencecoefficient of each to-be-predicted sample in the current to-be-predicted sample; taking a prediction classification result of the to-be-predicted sample with the prediction confidence not lower thana prediction confidence threshold as a prediction classification result of the current prediction submodel and outputting the prediction classification result; and taking the obtained to-be-predictedsample with the prediction confidence lower than the prediction confidence threshold as the current to-be-predicted sample of the next prediction submodel. By means of the method, another different prediction sub-model can be used for predicting the data which is not easy to predict at present again, and therefore the classification prediction efficiency and accuracy can be improved.

Description

technical field [0001] The present disclosure generally relates to the field of computer technology, and in particular, relates to a method and device for classification prediction based on a prediction model, a method and device for abnormal transaction prediction based on a prediction model, and a method and device for training a prediction model. Background technique [0002] With the popularization and development of artificial intelligence and machine learning technology, more and more companies are trying to use machine learning technology to solve business problems, and machine learning technology is widely used in various tasks. Taking the classification model as an example, it has a wide range of applications in tasks such as user profiling, abnormal user discovery, and abnormal transaction mining. At the same time, a large amount of data can be collected, creating conditions for the use of machine learning techniques and improving the performance of machine learnin...

AI Technical Summary

Problems solved by technology

This means more storage overhead, higher resource consumption and longer training time

Another way is to increase the complexity of the model. Taking models such as random forest or gradient boosting decision tree (GBDT) as examples, the common method is to increase the number of trees in the model, which will also bring more More resource consumption and longer training time

In addition, the above methods can bring very limited performance improvements to machine learning models.

Images