Threat assessment modeling method based on small dataset and Bayesian network

Abstract

The invention provides a threat assessment modeling method based on a small dataset and a Bayesian network. The existence constraint of a Bayesian network edge is introduced into a structure modeling stage by aiming at a threat assessment modeling problem under a small dataset condition, a parameter monotone constraint is introduced into the parameter modeling stage to make up a problem that modeling accuracy is poor since information contained in the small dataset is insufficient, a feasible solution is provided for the threat assessment modeling problem under the small dataset condition, in addition, a referable method is provided for a small dataset modeling problem in other fields, and the method has a wide application prospect.

Description

technical field [0001] The invention relates to a modeling method for threat assessment using machine learning technology. Background technique [0002] Completing a fast and accurate threat assessment of threat sources is a key link in actual combat. On the one hand, in a complex and ever-changing combat environment, factors such as enemy interference and sensor performance limitations lead to insufficient information or observation data; mission, threat assessments and operational decisions have to be made when data is not yet sufficient. Therefore, there is an urgent need for a method that can perform threat assessment modeling under insufficient data conditions, which is referred to as a small data set here. The main idea of ​​the current threat assessment modeling method is to select threat factors according to the assessment object, and then synthesize these threat factors. The main methods are neural network, support vector machine, Bayesian network, vague set, exp...

AI Technical Summary

Problems solved by technology

In the literature "Research on Algorithms for Threat Level Assessment Based on Bayesian Networks", Yu Zhouyi et al. studied the inference model and inference algorithm based on Bayesian networks, but the Bayesian network models involved are subjectively given based on experience. , lacks the process of combining data for model building, and does not have the ability to handle small data sets

Qin Zheng and others studied the data fuzzification process, data extraction and reasoning process of threat assessment in the literature "Situational Threat Assessment Method Based on Fuzzy Dynamic Bayesian Network", but also did not involve the Bayesian network model of small data sets build process

Images