Fuzzy Neural Networks

Provides an abstract of the presentation and a brief professional biography of the presenter. The complete presentation was not made available for publication as part of the conference proceedings.

Summary form only given. Interpolative reasoning plays a key role in both fuzzy logic and neural network theory. The basic approaches to interpolative reasoning in both fuzzy logic and neural networks were surveyed, and their differences and similarities were analyzed. An important issue in interpolative reasoning in fuzzy logic relates to the solution of a system of fuzzy algebraic equations. Various approaches to this problem, including fuzzy Lagrangian interpolation and the use of FA-Prolog, were described and analyzed. Among other issues discussed were the compression of a system of fuzzy if-then rules and the induction of rules from observations. >

In a recent article in Cancer, Cho and colleagues1 suggested that the Cancer of the Liver Italian Program (CLIP) system was the best prognostic system for predicting the survival of patients with hepatocellular carcinoma (HCC) who were undergoing transcatheter arterial chemoembolization, and further validation studies will be performed. We agree with the authors. Predicting mortality is an important component of the decision-making process, and scoring systems (Child-Pugh, CLIP, Model for Endstage Liver Disease, Barcelona Clinic Liver Cancer, etc) long have been accepted for establishing patient outcomes in HCC. Although these models perform well in predicting the mortality or morbidity in the original cohorts on which the models were developed, they may underestimate or overestimate mortality or morbidity in different populations.1 However, in the midst of such uncertainty, fuzzy logic may play an important role in the decision-making process. Fuzzy logic is the science of reasoning, thinking, and inference that recognizes and uses the real-world phenomenon that everything is a matter of degree. In the simplest terms, fuzzy-logic theory is an extension of binary theory that does not use crisp definitions and distinctions. Instead of assuming that everything must be defined in to black and white (binary view), fuzzy logic is a method that captures and uses the concept of fuzziness in a computationally effective manner. Neural networks and fuzzy logic are 2 complementary technologies. Neural networks can learn (ie, adapt) from data and feedback, but understanding the pattern learned by neural networks is difficult. Conversely, fuzzy rule-based models are easy to comprehend, because they use linguistic terms and the structure of if/then rules. Unlike neural networks, fuzzy logic does not come with a learning algorithm (ie, the system cannot learn from feedback). The combination of neural networks and fuzzy logic has created a new term, a neurofuzzy system. Neural networks, fuzzy systems, and the combination of both already have been applied successfully to computer-aided diagnosis—eg, detection of microcalcification, automatic detection of distorted plethysmogram pulses in neonates and pediatric patients, detection of erythematosquamous diseases, and lung-nodule detection—and have been useful for predicting the presence of prostate cancer.2 In conclusion, slight variations in population characteristics can change predictive values of the scoring systems. However, neurofuzzy systems can incorporate data from many clinical, biologic, and genetic variables and can provide better risk estimations for hepatocellular carcinoma. Mesut Tez MD*, Baris Zulfikaroglu MD , * Gazi University School of Medicine, Department of Surgery, Ankara, Turkey, Ankara Numune Hospital, Department of Fifth Surgery, Ankara, Turkey.

A pattern recognition approach to data fusion in Intelligent Vehicle Highway Systems : Palacharla, Prasad V., Ph.D. University of Illinois at Chicago, 1995. 78 pp.

Fuzzy logic provides a method for representing analog processes in a digital framework. Processes that are implemented through fuzzy logic are often not easily separated into discrete segments and may be difficult to model with conventional mathematical or rule-based paradigms that require hard boundaries or decisions, i.e., binary logic where elements are either a member of a given set or they are not. Consequently, fuzzy logic is valuable where the boundaries between sets of values are not sharply defined or there is partial occurrence of an event. In fuzzy set theory, an element's membership in a set is a matter of degree. This chapter describes the concepts inherent in fuzzy set theory and applies them to the solution of the inverted-pendulum problem and a Kalman-filter problem. Fuzzy and artificial neural network concepts may be combined to form adaptive fuzzy neural systems where either the weights and/or the input signals are fuzzy sets. Fuzzy set theory may be extended to fuse information from multiple sensors as discussed in the concluding section.

Fuzzy and Neuro Fuzzy logic concepts have been used for variety of applications, where machine intelligence can be used for the system. The fuzzy logic concept has been used for a number of biomedical applications also. We have used fuzzy and neuro fuzzy approach for tackling surgery procedures of a patient suffering from stress urinary incontinence (SUI). We have further extended the application of Fuzzy and Neuro Fuzzy logic for predicting the approximate recovery time of a patient under physical therapy suffering from stroke, paraplegia, quadriplegia etc. Here, we have outlined methods for above mentioned biomedical applications based on a knowledge base created from the patients of different age groups. We have shown our simulation results with input data and possible outcomes with the Fuzzy and Neuro Fuzzy system.

Identifying a strategy for legacy software system evolution requires analysis and assessment activities. Information on performance and costs of software systems must be considered when making decisions on the most suitable strategy to be applied. Many approaches have been defined for supporting this task, and several authors have proposed decision frameworks for aiding the selection of evolution strategy. These approaches often lack of techniques for the management of uncertainty, traditionally considered as unscientific and as a source of errors and arising from the confidence of the answers provided by respondents. An approach previously proposed is analyzed and extended with fuzzy logic concepts. The approach uses a measurement framework based on the Goal-Question-Metric (GQM) paradigm and a set of critiquing tables. Fuzzy logic principles have been introduced in both components, for obtaining a better insight of the analyzed software system quality and an indication of the risks to be assumed when one selected strategy is adopted instead of another.

When combining fuzzy logic and neural networks it is possible to get a hybrid system that can process uncertain values and can be trained. Fuzzy logic elements can be regarded as fuzzy-neural networks. In order to present a set of fuzzy values the ternary encoding is used. Three fuzzy neural networks on linear neurons are proposed. The first operates as a fuzzy logical minimum element, the second does as a fuzzy logical maximum element, the third - as a fuzzy logical complement element.

Ambiguity is always present in any realistic process. This ambiguity may arise from the interpretation of the data inputs and in the rules used to describe the relationships between the informative attributes. Fuzzy logic provides an inference structure that enables the human reasoning capabilities to be applied to artificial knowledge-based systems. For efficient working the artificial knowledge-based systems depend upon algorithms which are cumbersome to implement and require extensive computational time. On the other hand, the human brain which performs approximate reasoning employs simple information processing elements called neurons. The paradigm of artificial neural networks, developed to emulate some of the capabilities of the human brain, has demonstrated a great potential in terms of learning and adaptation for various applications such as system identification and control, pattern recognition, prediction, etc. They provide low-level computations and embodies salient features such as learning, fault-tolerance, parallelism and generalization. On the other hand, fuzzy logic provides a means for converting linguistic strategy into control actions and thus offering a high-level computation. Although, fuzzy logic and artificial neural networks are both functionally and structurally different, it is envisaged that the synthesis of these two areas will give rise to a new paradigm called fuzzy neural networks. The latter have the potential to capture the benefits of both the fields, fuzzy logic and neural networks, into a single paradigm. The objective of this paper is to describe the basic concepts of fuzzy neural networks. Towards this goal, a fuzzy neural structure based on the notion of T-norm and T-conorm is developed. A fuzzy cellular neural network as applied to image enhancement is also described in this paper.

PurposeThis paper aims to optimize supply chain information decision-making systems to better manage complex, high-dimensional and uncertain information through the integration of fuzzy logic and neural network technology.Design/methodology/approachA framework based on fuzzy logic reasoning is developed to address empirical issues in traditional supply chain systems. Subsequently, an innovative radial basis function-dynamic fuzzy neural network (RBF-DFNN) model is constructed, enhancing the system’s capability to interpret uncertain information. This model retains the advantages of traditional dynamic fuzzy neural networks (DFNN) while introducing an anti-fuzzy layer and optimizing the membership function and T-paradigm layers.FindingsThe RBF-DFNN model leads to the creation of a high-dimensional information decision-making model for supply chains. Experimental results indicate that this model effectively utilizes the K-medoids clustering algorithm to accurately capture the high-dimensional characteristics and intrinsic correlations of supply chain data. Parameter optimization significantly improves the model’s performance, with the root mean squared error (RMSE) and mean absolute error (MAE) enhanced, resulting in coefficients of determination rising from 95.6 and 97.8–99.1% compared to STPF-AIMM and ANFIS networks.Originality/valueThis study contributes to the advancement of supply chain management by developing a highly intelligent and refined decision-making model, enhancing the intelligence level of intelligent storage systems and promoting more sophisticated supply chain operations.

The fuzzy logic and neural networks are combined in this paper, setting up the fuzzy neural network (FNN); meanwhile, the distinct differences and connections between the fuzzy logic and neural network are compared. Furthermore, the algorithm and structure of the FNN are introduced. In order to diagnose the faults of nuclear power plant, the FNN is applied to the nuclear power plant, and the intelligence fault diagnostic system of the nuclear power plant is built based on the FNN. The fault symptoms and the possibility of the inverted U-tube break accident of steam generator are discussed. In order to test the system’s validity, the inverted U-tube break accident of steam generator is used as an example and many simulation experiments are performed. The test result shows that the FNN can identify the fault.

Fuzzy logic system based prediction effort: A case study on the effects of tire parameters on contact area and contact pressure

Over the last decade, an extensive research has been carried out in the areas of fuzzy logic and neural networks. Fuzzy logic has emerged as a mathematical tool to deal with the uncertainties in human perception and reasoning. It also provides a framework for an inference mechanism that allows for approximate human reasoning capabilities to be applied to knowledge-based systems. On the other hand, artificial neural networks have emerged as fast computation tools with learning and adaptivity capabilities. Recently, these two fields have been integrated into a new emerging technology called fuzzy neural networks which combines the benefits of each field. The objective of the paper is to establish the similarities and differences between fuzzy systems and neural networks and to discuss possible models for fuzzy neural networks which can be applied to system modeling and control.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Fuzzy Logic (Fuzzy Logic) is a branch of science based on thinking like human beings and solving them with mathematical functions. Fuzzy logic theory is a mathematical theory. Based on fuzzy set theory, it also uses intermediate values. The fuzzy logic that emerged in 1965 is used in many fields. In the production of pacemakers, in the production of artificial organs, in many electronic devices, company efficiency estimation, etc. situations are used. Fuzzy logic, which is frequently used in the solution of problems that occur in uncertain situations such as quality assessment in recent years, is one of the artificial intelligence methods. With the help of machines, people-specific data and experiences are studied using the fuzzy logic approach. In this study, by using Matlab Fuzzy Toolbox, it was aimed to design a system that gives information about the breeding performances of cows. The expert system was designed based on the optimal values under the ideal conditions specified in the literature. The architecture of the system presented in this paper is designed as three input parameters and one output. The designed system was tested with 100 sample values. Afterwards, expert results were evaluated and system decisions were compared. The success of the decision support system was 94%. As a result, the reproductive efficiency of cows can be determined with this designed system. With this determination, the handling or disposal of cows can be determined.

This paper presents a FPGA based approach for a modular architecture of Fuzzy Neural Networks (FNN) to embed with easily different topologies set up. The project is based on a Takagi - Hayashi (T-H) method for the construction and tuning of fuzzy rules, this is commonly referred as neural network driven fuzzy reasoning. The proposed architecture approach consists of two main configurable modules: a Multilayer Perceptron - MLP with sigmoidal activation function that composes the first module to determine a Fuzzy membership function; the second employs an MLP with pure linear activation function to define the consequents. The DSPBuilder® software along the Simulink® is used to connect, set and synthesize the Fuzzy Neural Network desired. Other hardware components employed in the architecture proposed cooperate to the system modularity. The system was tested and validated through a control problem and an interpolation problem. Several papers proposed different hardware architecture to implement hybrid systems by using Fuzzy logic and Neural Network. However, there is no approach with this specific neural network driven fuzzy reasoning by T-H method and the aim to be embedded. The Self-Organizing Map (SOM) and Levenberg-Marquardt backpropagation were used to train the FNN proposed off-line.