Method and device for determining target features for business data

Abstract

The embodiment of the invention provides a method and device for determining target features for business data, and the method comprises the steps: employing a piecewise linear model as an agent modelof a business model for processing the business data, and carrying out training in advance through a historical data processing result of the business model; for business data with target features tobe determined, obtaining a plurality of service features extracted by the business model for current business data; inputting the plurality of service features into a piecewise linear model and acquiring a neuron activation state corresponding to each hidden layer, determining each importance coefficient corresponding to each business feature according to the neuron activation state correspondingto each hidden layer and a weight matrix of the corresponding hidden layer, and determining a plurality of target features from a plurality of business features based on each importance coefficient.According to the embodiment, more accurate and effective explanation can be provided for the business model.

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 an apparatus for determining target features for business data by using a computer. Background technique [0002] With the development of machine learning technology, Deep Neural Network (DNN) is favored by those skilled in the art because it imitates the way of thinking of the human brain and has better results than simple linear models. A deep neural network is a neural network with at least one hidden layer that can provide modeling for complex nonlinear systems, thus improving the capabilities of the model. However, conventional deep neural network methods cannot give interpretability due to complex network structures. [0003] As an example, assume that in the lending scenario, the risk level of a user’s loan is judged. If the user’s risk is high, it will increase the risk of asset loss on the financial platform. In conventi...

AI Technical Summary

Problems solved by technology

In order to make the model interpretable, some schemes provide a new interpretable model structure, but for a model that runs stably online, changing the model will have a certain impact on stability. Therefore, consider using a proxy model to Online Models Explained

In conventional technology, the interpretability of the model often needs to rely indirectly on the simple model. If the complexity of the model to be explained is high, the simple model may not fit well

On the other hand, the proxy model usually keeps the input features and output results consistent with the online model. In the case of inconsistent hidden layers in the middle, the fidelity of the proxy model is poor, resulting in poor interpretation accuracy.

Images