Abstract
This paper offers an approach for evaluating the significance of individual characteristics of recognized objects. The scope of this approach is not the subject area where objects and characteristics of these objects are specified, but a trained ∑Π - neural network that works correctly on the specified subject area. In this paper, we propose a method for constructing a crucial function based on the weight characteristics of a correctly functioning ∑Π - neuron. A logical derivative is used to evaluate the significance of object characteristics. This makes it possible to track how the decision function will change its value if one or more object characteristics change their value. This will allow us to draw a conclusion about the most important properties of the subject area under consideration.
Highlights
Today, neural networks are one of the most popular tools for solving poorly formalized tasks
Available structure and weight characteristics, which acquired a neural network as a result of training
If the query does not coincide with the values of the variables that are in the training set, for example (0,1,0), the result may be incorrect or it may not exist at all
Summary
Neural networks are one of the most popular tools for solving poorly formalized tasks. As a result of training according to the table, the -neuron will look like: Any query corresponding object. If the query does not coincide with the values of the variables that are in the training set, for example (0,1,0), the result may be incorrect or it may not exist at all. It could be an object, b-4, or c-6, in cases where there are inaccuracies, noise, interference in the data. When constructing the decisive function, you may not know the training set; it is enough to know the value of weights and the structure of the neuron. The decision function for our example will look like this: That is, the most important features for the source data will use the logical derivative
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call