Kernel Probability Density Estimation in Solving the Problem of Classification of the Technical Condition of Complex Systems
摘要
Abstract
To improve the versatility of modeling for recognizing the technical condition of a complex system, a solution to the problem of its statistical classification is proposed. The fact that the current condition belongs to a specific class is evaluated by confirming the hypothesis using a decision function based on the inductive behavior concept. The confirmation is made by estimating the probability of the object’s current parameters falling within a 2D parallelepiped of the joint density function determined using the kernel probability density estimation method.