Model fusion method and system, electronic equipment and medium

Abstract

The invention discloses a model fusion method and system, electronic equipment and a medium, and the method comprises the steps: carrying out the multi-round training of a model, obtaining a probability vector of a training model, carrying out the processing of the probability vector, obtaining an average probability vector, and carrying out the entity labeling of the average probability vector to obtain a second entity labeling result set; after new entities and old entities are distinguished according to the first entity labeling result set, selecting a new entity labeling result set and an old entity labeling result set corresponding to the new entities and the old entities from the second entity labeling result set; after a first credibility threshold value and a second credibility threshold value are preset, selecting a third entity labeling result set from the second entity labeling result set according to the first credibility threshold value, and selecting a fourth entity labeling result set from the second entity labeling result set according to the second credibility threshold value; and selecting the third entity labeling result set and the fourth entity labeling result set according to the total number of the training models to obtain a final entity labeling result.

Description

technical field [0001] The present application relates to the technical field of knowledge graphs, in particular to a model fusion method, system, electronic equipment and media. Background technique [0002] In the field of machine learning, the same problem can usually be solved using multiple models with different parameters and different structures, and a method is needed to integrate different models into a robust model. It is also necessary to ensure that the integrated model outperforms the underlying sub-models. The current model fusion methods include Bagging (bootstrap aggregating), Boosting, and Stacking. Bagging is a method of predicting votes using N models for classification problems, and a method of predicting the average of N models for regression problems; Boosting is to assign equal weights to each training example at the beginning of training, and then use this algorithm to train the training set for t rounds , after each training, assign a larger weight...

AI Technical Summary

Problems solved by technology

[0004] The embodiment of the present application provides a model fusion method, system, electronic equipment and medium, so as to at least solve the problem of ignoring task characteristics when targeting specific tasks in the process of model fusion through the present invention, resulting in reduced pertinence to specific tasks in the process of model fusion And other issues

Images