effectiveness in practice. Currently, active research is underway in the field of implementing a network-centric

approach to improve the management of business entities. A network-centric control system is a distributed

control system in which its main components are integrated into a single information space. The purpose of this

research is to develop proposals for building a network-centric model for managing geographically distributed

crop areas based on a digital IoT-platform, which represents a universal info-communication environment. The

features of network-centric control using fuzzy modeling systems that provide analytical support for decision-

making under uncertainty are considered. As an example, it is considered a method of fuzzy modeling and

forecasting of segments of averaged sensor readings from web-devices, which can provide information support

for predictive and prescriptive analytical solutions. Proposed solutions can be used not only in the field of

precision agriculture, but also to build a digital network-centric management platform for any business entity.

Crop area management system includes sources and

consumers of information, telecommunications

facilities, as well as a center for processing input data

and preparing information for users. A network-

centric management system based on a unified IoT

information platform embodies the idea of creating

“Smart Agriculture”, which is a high-tech set of

solutions that allows for maximum automation of

agricultural enterprises. Crops, soil, irrigation

devices, agricultural equipment and web-devices that

monitor climate conditions, including temperature

and ground humidity, can accumulate, send and

process data, creating invisible images ready to be

multispectral analysis of data from multi-copters in

the form of vegetation maps.

circuits of the proposed NCMS is determined by the

sufficient set of tools that provide the required

adequacy of information support in the process of

collecting, storing and processing sensor readings

from web-devices and multispectral data from

remote monitoring of the current state of the crop

area, carried out by multi-copters in real time. The

disciplines [1], the tools for compiling knowledge in

solving management problems can and should be

subjected to significant revision. Its main essence

lies in a radical change in the point of view on the

role and place of modern intelligent technologies in

the organization of hierarchical management of

complex dynamic objects.

for processing sensor data to provide information

support for making agricultural decisions based on

model is to identify internal connections that

determine cause-effect relationships between sensor

readings throughout the entire observation period.

Depending on the number of prerequisites in the

fuzzy relation of the form “If <...>, then <...>”,

the principle: if the fuzzy set At is connected with Ap

presented in Table 3 and Table 4, respectively.

between sensor readings x64(74.86) and x65(66.79),

example, A47, A59A58 can be interpreted as the

fuzzy implication: “If xt is A47 and xt is A59, then xt+1

interpreted as the fuzzy implication:

“If xt is A59 and xt is A58, then xt+1 is A52 or xt+1 is A59”

The corresponding geometric interpretations of the

forecasting results are presented in Figure 5.

values of statistical criteria for assessing the

adequacy of predictive models are presented, [18]:

MSE (Mean Squared Error), MAPE (Mean Absolute

Percentage Error) and MPE (Mean Percentage

Error), which are calculated using the corresponding

formulas:

metrics of forecast errors, allows to evaluate the

accuracy of the forecast in absolute units of

deviation as a percentage. In particular, MAPE can

be useful for comparing the forecast accuracy of

different models processing different ranges of data.

orders

predictive models are equal to MSE1 = 49.16 and

MSE2 = 30.52, respectively. According to the MAPE

criterion, which demonstrates the percentage of the

forecast error in comparison with the actual values of

[6] S. M. Chen, and N. Y. Wang, “Fuzzy

[7] S. M. Chen, “Forecasting enrollments based

[10] Y. L. Huang, S. J. Horng, M. He, P. Fan, T.

[11] K. Huarng, “Heuristic models of fuzzy time

[14] J. R. Poulsen, Fuzzy Time Series Forecasting

[15] Q. Song, and B. S. Chissom, “Forecasting

[16] Q. Song, and B. S. Chissom, “Forecasting