Lazy learningis a training set processing method that trains while receiving test samples, in contrast toEager to learn, which will start learning the samples during the training phase.

If the task data changes frequently, lazy learning can be used. No training is performed first, and probability valuation is performed based on the current data after receiving the prediction request. If the data continues to increase, incremental learning can be performed only on the attribute values of the new samples based on the existing valuations. It only needs to make technical corrections to the relevant probability valuations.