NEURAL NETWORK METHOD FOR PREDICTING THE STATE OF INFORMATION OBJECTS

2021 ◽  
pp. 37-43
Author(s):  
A. D. Obukhov
Keyword(s):  
Neural Network ◽  
Short Term Memory ◽  
Learning Technologies ◽  
The State ◽  
Time Interval ◽  
Problem Solution ◽  
Neural Network Method ◽  
The Neural Network ◽  
Network Method ◽  
Information Objects

The analysis and process of not only the current states of information objects, but also the prediction of future states with a certain time interval presents a major significance for adaptive information systems. This allows improving the quality and reliability of these systems functioning, reducing the delay in response to external influences, preparing for operations, and increasing the responsiveness and speed of the system. In order to solve this problem, the article proposes a neural network method for forecasting the state of information objects based on the application of machine learning technologies. A formalized algorithm for its implementation in the notation of set theory is presented. A distinctive characteristic of the designed method is the automatic determination of the optimal structure of the neural network, depending on the type of information object. The method also covers the possibility of using the complex of the previous states of the object to improve the forecast accuracy. Practical researches on approbation of the neural network method showed its efficiency and high accuracy. The following popular datasets were used as input data: Individual household electric power consumption, HAR (Human Activity Recognition) accelerometer, as well as gathered data on human relocation. LSTM (Long Short-Term Memory) neural network was applied to conduct the forecasts. The comparison of the developed method with a similar software solution DEvol (DeepEvolution) showed comparable or better indicators in terms of accuracy and time for the problem solution (1.7 times faster on average).

scholarly journals Neural network method of data processing and transmission in adaptive information systems

2021 ◽  
Vol 31 (1) ◽  
pp. 149-164
Author(s):  
A.D. Obukhov ◽  
M.N. Krasnyansky
Keyword(s):  
Neural Network ◽  
Information Systems ◽  
Data Processing ◽  
Data Transfer ◽  
Learning Technologies ◽  
Practical Significance ◽  
Neural Network Method ◽  
The Neural Network ◽  
Network Method ◽  
Selection Of

The problem of automation of the processes of information transmission and processing in adaptive information systems is considered. An analysis of existing approaches to solving this problem showed the prospects of using neural network technologies. A neural network method for processing and transmitting information in adaptive information systems is formulated. The method includes a formalized description of a neural network data channel - a software tool for analysis, data processing and selection of data transfer protocol. The main stages of the proposed method are outlined: classification of the structures of the source data, their transformation, data processing, selection of the necessary protocol for transmitting information. Each of the stages is implemented through neural networks of various architectures. The theoretical rationale of the possibility of using the neural network method is given, obtained on the basis of the proof of a number of theorems. The novelty of the proposed method consists in the transition from an analytical solution of the problems of classification, processing and data transfer to an automated approach using machine learning technologies. The practical significance of the neural network method is to reduce the complexity of the implementation of information processing and transmission processes, to increase the level of automation in the organization of intermodular interaction. The implementation of the neural network method has been assessed using a number of software complexity assessment metrics. The application, virtues and failings of the developed method are analyzed.

Methods for evaluation the manufacturability of a vehicle

2020 ◽  
pp. 152-160
Keyword(s):  
Neural Network ◽  
Initial Data ◽  
Comparative Data ◽  
Training Algorithm ◽  
Neural Network Method ◽  
The Neural Network ◽  
Network Method ◽  
Calculation Results ◽  
Production And Operation

Methods for evaluation the manufacturability of a vehicle in the field of production and operation based on an energy indicator, expert estimates and usage of a neural network are stated. By using the neural network method the manufacturability of a car in a complex and for individual units is considered. The preparation of the initial data at usage a neural network for predicting the manufacturability of a vehicle is shown; the training algorithm and the architecture for calculating the manufacturability of the main units are given. According to the calculation results, comparative data on the manufacturability vehicles of various brands are given.

scholarly journals Comparative analysis of the geometrized histogram method and the neural network method for road markings recognition

2021 ◽  
pp. 1-22
Author(s):  
Aleksei Valerievich Podoprosvetov ◽  
Dmitry Anatolevich Anokhin ◽  
Konstantin Ivanovich Kiy ◽  
Igor Aleksandrovich Orlov
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Comparative Analysis ◽  
Computational Complexity ◽  
Video Sequences ◽  
Histogram Method ◽  
Neural Network Method ◽  
The Neural Network ◽  
Network Method ◽  
Accuracy Speed

This paper compares two approaches to determining road markings from video sequences, namely, the method of finding the markings using geometrized histograms and the method based on neural networks. An independent open dataset TuSimple is used to conduct a comparative analysis of the algorithms. Since the investigated methods have different architectures, their work is evaluated according to the following metrics: Accuracy, speed (relative FPS), general computational complexity of the algorithm (TFlops).

scholarly journals Porosity Prediction Using Neural Network Based on Seismic Inversion and Seismic Attributes

2019 ◽  
Vol 125 ◽  
pp. 15006
Author(s):  
Taufik Mawardi Sinaga ◽  
M. Syamsu Rosid ◽  
M. Wahdanadi Haidar
Keyword(s):  
Neural Network ◽  
Probabilistic Neural Network ◽  
Average Frequency ◽  
Seismic Attributes ◽  
Seismic Inversion ◽  
Carbonate Reservoir ◽  
Neural Network Method ◽  
Time Migration ◽  
The Neural Network ◽  
Network Method

It has done a study of porosity prediction by using neural network. The study uses 2D seismic data post-stack time migration (PSTM) and 2 well data at field “T”. The objective is determining distribution of porosity. Porosity in carbonate reservoir is actually heterogeneous, complex and random. To face the complexity the neural network method has been implemented. The neural network algorithm uses probabilistic neural network based on best seismic attributes. It has been selected by using multi-attribute method with has high correlation. The best attributes which have been selected are amplitude envelope, average frequency, amplitude weighted phase, integrated absolute amplitude, acoustic impedance, and dominant frequency. The attribute is used as input to probabilistic neural network method process. The result porosity prediction based on probabilistic neural network use non-linear equation obtained high correlation coefficient 0.86 and approach actual log. The result has a better correlation than using multi-attribute method with correlation 0.58. The value of distribution porosity is 0.05–0.3 and it indicates the heterogeneous porosity distribution generally from the bottom to up are decreasing value.

Sign in / Sign up

Export Citation Format

Share Document