CC BY
13ТЕХНИЧЕСКИЕ НАУКИ
APPLICATION TECHNOLOGY OF FUZZY LOGIC INFERENCE SYSTEMS
*MAMMADOVA K.A., ALIYEVA E.N.
1 Candidate of tech. sciences, lecturer at the Department of Computer Engineering, Azerbaijan State
Oil and Industry University, Baku, Azerbaijan
2 Candidate of tech. sciences, lecturer at the Department of Computer Engineering, Azerbaijan State
Oil and Industry University, Baku, Azerbaijan
Summary. The Matlab environment was used for the purpose of description, simulation and analysis of the fuzzy inference system, which is an important part of fuzzy logic. In most practical applications, such systems implement accurate nonlinear representation, which is represented as fuzzy rules that encode expert or common sense knowledge about the problem at hand. The research used a lot of data downloadedfrom the website. This is a actual problems lately. Our implementation of a fuzzy inference system is based on two inputs, positive and negative, and nine rules, each of which depends on solving several different fuzzy linguistic sets of the inputs.
The approach proposed in this paper can facilitate the research and development of big data analytics due to its simple and clear solution method. Application of fuzzy inference system in Matlab environment is to analyze large amount of data faster than traditional methods. The purpose of this work is to construct a fuzzy logic inference system, which is an important part of fuzzy logic, to analyze the obtained results and to find an optimal solution.
Keywords: fuzzy logic, fuzzy inference, fuzzy logic operator, implication application method
1. introduction
Big data analysis and research is the most important part of the research to be solved. A review of a number of scientific research studies revealed many simulation environments and network simulators available for measuring network performance. The creation and study of a fuzzy logical inference system is considered in the research work. A logical inference system helps determine the degree of correspondence between fuzzy input and rules. Determines which rules should be applied based on the input data. After that, the applicable rules are combined to prepare management activities. In this study, Matlab, a high-level technical computing language and interactive environment, was chosen to create a fuzzy logic inference system. Because it has a simple and clear solution method, it is more efficient for algorithm development and data visualization, data analysis, and numerical computation. By using the Matlab product, it is possible to analyze large volumes of data faster than traditional methods [1,8].
Matlab can be used in a wide range of applications, including signal and image processing, communication, control design, testing and measurement, financial modeling and analysis, and computational biology. Additional toolboxes (a set of custom Matlab functions) further expand the Matlab environment to solve specific categories of problems in different applications.
In order to establish and share of the work, a number of functions are included to the Matlab environment. Matlab has a high-level language for technical calculations, as well as code, file and data management development environment, interactive tools for iterative search, design and problem solving, linear algebra, statistics, Fourier analysis functions, mathematical functions for filtration, optimization and numerical integration, Data which is converted by 2-D and 3-D graphics functions for visualization, and a library of tools for building custom graphical user interfaces. Matlab code has the advantage of integrating with other languages and programs, for example, Matlab-based
algorithms with foreign programs and languages such as C, C ++, FORTRAN, Java ™, COM and Microsoft Excel [2,3].
The Matlab environment was used in the research to describe, simulate and analyze the fuzzy logic inference system, which is an important part of fuzzy logicIn most practical applications, that kind of systems provide an accurate nonlinear description, which is presented in the form of fuzzy rules that encode expert or general knowledge of an existing problem.
The research used a large amount of data downloaded from the website [1,2]. This is a topic that has become topical recently. It has a high data volume: 23 posts and 10755 comments. All comments for each post were collected within the first 30,000 seconds using the Facebook Graph API. A straight table format (for example, a CSV file) was used because the data was easy to store in a distributed file system. The title of the CSV file has a table with the following columns: Data entry time, Topic name, Post, Comment, Positive, Negative. The chosen topic is a trend topic recently "2021 US Presidential Elections" shown in Figure 1.
2. Application technology of fuzzy logic inference systems
Fuzzy inference is the process of forming input / output images using fuzzy logic. Fuzzy Logic Toolbox software has a command line and linear functions for creating Mamdani and Sugeno fuzzy systems. [4-6].
Fuzzy inference systems have been successfully applied in a number of areas, such as automatic control, data classification, decision analysis, expert systems, and computer vision [5]. Fuzzy inference systems have the following architecture (Figure 1):
Fuzz)' logic database
Figure 1. General Architecture of Fuzzy logic inference system
Fuzzy substantiation (fuzzy logic inference operations on fuzzy IF - THEN rules) performed by Fuzzy logic inference systems is performed in the following stages:
1. Obtain affiliation values for each selected linguistic function of each linguistic variable. (This stage is often called fuzzification.)
2. To combine the values of belonging in the main part by the operations of a minimum or maximum logic to determine the strength of each rule.
3. Create appropriate results (fuzzy or fuzzy (accurate)) or rules, depending on the degree of implementation.
4. Collect relevant results to obtain an inaccurate (fuzzy) result.
5. Defuzzification [4,5].
Establishment of fuzzy logic tool systems. The Fuzzy Logic toolbox contains five basic GUI (graphical user interface) tools for setting up, editing, and monitoring fuzzy inference systems: fuzzy inference system or FIS (fuzzy inference system) Editor, affiliation of function editor, rule viewer and surface viewer. These GUIs are dynamically interconnected. This, changes in the FIS
when using one of them can affect any of the other open GUI. Any or all of them can be made open
to any system.
Figure 2. To set up the system with Fuzzy Logic ToolBox
FIS Editor solves high-level issues for the system: How many input and output variables are there? What are their names? Fuzzy Logic Toolbox does not limit the number of entries. However, the number of entries may be limited by the available memory of your machine. If the number of inputs is large or the number of affiliation functions is very large, it may also be difficult to analyze the FIS using other GUI tools.
Membership Function Editor is used to define the forms of all affiliation functions associated with each variable.
Rule Editor is used to edit the list of rules that define the operation of the system.
The Rule Viewer and The Surface Viewer are used to control FIS, as opposed to editing. They are basically just readable tools. The Rule Viewer is a Matlab-based display of fuzzy inference diagrams. Used as a diagnostic, it can show which rules are active or how different forms of affiliation function affect results. The Surface Viewer is used to show the dependence of one of the outputs on any one or two inputs, so it creates a map of the output surface for the system.
Applied technology with a fuzzy logic inference system. Using the Matlab environment in the application, we included a matrix of Negative and Positive columns from the set of downloadable data to apply and simulate a fuzzy logic inference system, as shown in Figure 2.
EJ FIS Editor QSMNC.modeli
XX
XX
Figure 3. Application of FIS
ОФ "Международный научно-исследовательский центр
— flEfrr ■
23ГЯБ r=v — — ~ \
imük
fers*3 ■" ■ ■
•juT £23 — \ А /\ /\ Г~ xYxX
— — Ё. .... —
--' " ; -
Figure 4. Membership Function Editor
Fuzzification of inputs. The first step is to accept the inputs and determine the degree to which they belong to each of the relevant fuzzy sets through their affiliation functions, the input is always a crisp numeric value. The output has a fuzzy degree of belonging to the relevant linguistic set. [6].
Our application of the fuzzy inference system is based on two inputs: positive and negative (as shown in Figure 4); The nine rules and each of these rules depend on the solution of several different fuzzy linguistic sets of inputs: Low Positive, Medium Positive, High Positive, and there will be a negative score Low Negative, Medium Negative, and High Negative. We have a set of five fuzzy linguistic variables for outputs: Too Rejoicing, Rejoicing, Neutral, Angry, Too Angry. Before evaluating the rules, the inputs and outputs must be fuzzification in according to each of these linguistic sets.
The rules are based on minimum operations as follows:
l.If negative is LowNegative and positive is LowPositive then Result will be TooAngry;
2.If negative is LowNegative and positive is MediumPositive then Result will be Rejoicing;
3.If negative is LowNegative and positive is HighPositive then Result will be TooRejoicing;
4.If negative is MeduimNegative and positive is LowPositive then Result will be Angry;
5.If negative is MediumNegative and positive is MediumPositive then Result will be Neutral;
6.If negative is MediumNegative and positive is MediumPositive then Result will be Rejoicing;
7.If negative is HighNegative and positive is LowPositive then Result will be TooAngry;
8.If negative is HighNegative and positive is MediumPositive then Result will be Angry;
9. If negative is HighNegative and positive is HighPositive then Result will be Neutral (table 1).
Table 1.
ОФ "Международный научно-исследовательский центр "Endless Light in Science"
Table o ? Rules
^^-^Positive LowPositive MediumPositive HighPositive
Negative^--^
LowNegative TooAngry Rejoicing Too Rejoicing
MediumNegative Angry Neutral Rejoicing
HighNegative TooAngry Angry Neutral
Application of Fuzzy Operator. Any number of well-defined methods can be applied to the AND or OR operation
By typing any function in FIS and defining it as the method of your choice, custom methods can be created for And and OR.
In order to calculate the output, AND operator is shown as min which evaluates the definition of Rule 5. Two different parts of the determination (Low Negative and Medium Positive) set fuzzy affiliation values of 0.3 and 0.5, respectively. The fuzzy AND operator simply selects a minimum of two values, 0.5, and completes the fuzzy operation for rule 5. There is a possibility that the AND method will again result in 0.5.
negative is HighNegative =0.701 AND positive is LowPositive=0.337, Result of implication
Figure 5. Fuzzy Logic Operator
Rule 3. If negative is LowNegative and positive is HighPositive then Result is TooRejoicing
3. Application of the implication method
Before applying the implication method, the rule weight must be determined. Each rule has a weight (it can be given values from 0 to 1), which is applied to the value given by the implication. In general, this weight is 1 and therefore has no effect on the implication process. However, it is possible to reduce the effect of one rule on another by changing the weight value from one to another. [4-6].
Figure 6. implication method
Defuzzification. Defuzzification is the process of obtaining a crisp number from the output of a fuzzy set. It is used to convert a fuzzy result into an accurate (crisp) result.
There are five defuzzification methods: central mean, bisector, mean of maximum (mean maximum value of output set), maximum of maximum, and minimum of maximum. Perhaps the most popular method of defuzzification is to calculate the central mean returning to the center of the area under the curve, as shown below:
Ike result ofdefuzzificnrion
D 10
Figure 7. Operation of Defuzzification
Fuzzy inference diagram. A fuzzy inference diagram is a combination of all the small diagrams presented so far. It shows all the parts of the fuzzy inference process that you are researching at the same time. As shown in the figure below, the data goes through a fuzzy inference diagram [7,9]:
Figure 8. Rule Viewer
The following figure shows a surface diagram describing the dependence of the outputs and
inputs.
Figure 9. Editor Surface Viewer
CONCLUSiON
This article describes a system of fuzzy logic inference, which is an important part of fuzzy logic. In most practical applications, such systems provide an accurate nonlinear description, which is defined as fuzzy rules that encode expert knowledge of an existing problem.
REFERENCES
1. MathWorks, "MATLAB Getting Started Guide", 2011.
2. K. Srividya and A. M. Sowjanya, "Sentiment analysis of facebook data using naive bayes classifier Assistant Professor , Department of Computer Science and Engineering , Assistant Professor , Department of Computer Science and systems engineering , AU College of Engineering ( A ), Andhra ," vol. 15, no. 1, pp. 179-186, 2017.
3. Zadeh, L.A., "Fuzzy Logic," Computer, Vol. 1, No. 4, pp. 83-93, 1988
4. D. S. Hooda Vivek Raich. Fuzzy Logic Models and Fuzzy Control/ Alpha Science International LTD. Copyright © 2017 408p.
5. Fuzzy Logic toolbox - MathWorks./ User's Guide. 2016, 388p
6. Mamdani, E.H., "Applications of fuzzy logic to approximate reasoning using linguistic synthesis," IEEE Transactions on Computers, Vol. 26, No. 12, pp. 1182-1191, 1977
7. Zainab N. Al-Qudsy. Fuzzy Inference Systems. 2019
8. Gayatri Kapil, Alka Agrawal, and R. A. Khan. A Study of Big Data Characteristics. Conference: 2016 International Conference on Communication and Electronics Systems (ICCES), October, 2016
9. https://www.mathworks.com/help/fuzzy/fuzzy-inference-process.html, (Fuzzy Inference Process), April, 2019.
