statistical data, weight coefficients.
I. I
NTRODUCTION
Neural network technologies is a branch of machine
learning dedicated to modeling human intellectual
activity. Today , it is an extensive area of research and
development of intelligent systems designed for work in
difficult-to-formalize areas of human activity. The
problems solved by the neural network method are
characterized by the presence of a large number of degrees
of freedom with the number of options for finding
solutions approaching infinity. In contrast to the rigidly
deterministic computer programs machine learning
systems themselves looking for ways to solve the
problem. At the same time, they can change their
parameters and structure, improve and develop [1,2].
Using a multi-layer neural network of direct
propagation, the classification problem is solved.
Classification models-describe rules or a set of rules
according to which a description of any new object can be
assigned to one of the classes. Such rules are built on the
basis of information about existing objects by dividing
them into classes [2-4].
II. E
ASE OF
U
SE
In this article, self-government bodies are considered
as the object of research [5], and the subject is decision –
making support based on machine learning using neural
network technologies. In this case the main task is the
following: based on machine learning parametric
assessment of educational, social and criminal state
authorities and the relevant decisions of the experts
expressed by the numbers [6], support decision-making in
the new government.
The structure of the training sample. In the
classification and regression problem, you need to
determine the value of the dependent variable of an object
based on the values of other variables that characterize
this object. Formally, the problem of classification and
regression can be described as follows. There are many
objects available:
𝑋 = {𝑋
1
, 𝑋
2
, … , 𝑋
𝑗
, … , 𝑋
𝑛
}
(1)
Here
𝑋
𝑗
is— the object under study. In our task,
objects are self-government bodies.
Each object is characterized by a set of variables,
that is, in the case of the task set, each self-government
body is characterized by parametric estimates obtained in
three areas of activity (education, social status and
intercession):
𝑋
𝑗
= {𝑥
𝑗1
, 𝑥
𝑗2
, … , 𝑥
𝑗ℎ
, … , 𝑥
𝑗𝑚
, 𝑦
𝑗
}
(2)
Here
𝑥
𝑗ℎ
is — independent variables whose
values are known and on the basis of which the value of
the dependent variable y is determined. In this example,
the independent variables are parametric estimates for the
fields of activity. Here,
𝑦
𝑗
is a dependent indicator and
contains numeric values of expert ratings. The training
sample has the following matrix representation:
TABLE I. Training Sample Matrix Representation
Objects
Parametric
estimation
of
educational
activities.
Parametric
estimation of
the social
environment
Parametric
estimation
of the state
of crime
Dostları ilə paylaş: |