Interpretation method and device for entity object classification results

The invention discloses an interpretation method and device for an entity object classification result. The method comprises the steps that service data input into a classification model and a classification result of an entity object corresponding to the service data are acquired, the model is a multi-layer neural network model, an activation function of a hidden layer of the multi-layer neural network model is a partwise linear activation function, and the service data are feature data of the entity object; for each feature of the service data, solving a gradient of the feature according tothe classification result to obtain a feature gradient of the feature; and explaining the classification result of the entity object according to the feature gradient of each feature. 说明书披露一种实体对象分类结果的解释方法和装置，所述方法包括：获取输入分类模型的业务数据，以及所述业务数据对应的实体对象的分类结果，所述模型是多层神经网络模型，所述多层神经网络模型隐藏层的激活函数为分段线性激活函数，所述业务数据是所述实体对象的特征数据；针对所述业务数据的每个特征，根据所述分类结果对所述特征求梯度，得到该特征的特征梯度；根据各个特征的特征梯度对所述实体对象的分类结果进行解释。