Adaptive inference-based learning and rule generation algorithms in Fuzzy Neural Network for failure prediction

Applications for power modules under harsh environmental conditions have gained importance in recent years, e.g. in offshore wind turbines. One of the critical conditions that leads to failures in these applications is humidity. Therefore, it is important to subject the power electronics to appropriate load tests in advance, in order to generate these failures. An established test method for the qualification of power electronics is the High Humidity High Temperature Reverse Bias (H3 TRB)-test. This paper deals with the prediction of failures in H3 TRB-tests. The goal in this paper is to use the currents noise to improve the method of forecasting time to failure in H3 TRB-testing. This is based on the assumption that there may be a correlation between the time to failure and the point in time at which the electrical parameters begin fluctuating.

For common people, stock investing is one popular way to manage their property. As Information Technology (IT) has risen in recent years, every security company has analyzed computer systems for their customers by developing their own investing. Taiwan is an island nation, and the economy relies on international trade deeply. The fluctuations of international stock markets will impact Taiwan stock market to a certain degree. Therefore, the use of fluctuations of other stock markets as forecasting factors for forecasting the Taiwan stock market is a practical way. In this paper, the proposed model uses the fluctuations of other national stock markets as forecasting factors, employs different discretization methods (Fuzzy C-means clustering, Subtractive Clustering and Cumulative Probability Distribution Approach) to discretize stock data, utilizes a fuzzy inference system to produce understandable rules, and applies an adaptive neural network to optimize model parameters to reach the best forecasting accuracy. To evaluate the forecasting performances, the proposed model is compared with two different models, Chen’s model and Yu’s model. The experimental results indicate that the proposed model is superior to the listing methods in terms of RMSE (root mean squared error). Key words: Data mining, Adaptive Neuro-Fuzzy Inference System (ANFIS), stock forecasting, subtractive clustering, cumulative probability distribution approach.

Cerebellar model articulation controller (CMAC) is a useful neural network learning technique. It was developed two decades ago but yet lacks adequate learning algorithm especially when it is used in a hybrid-type controller. Part I of this work was devoted to introduce a new CMAC adaptive learning algorithm. Part II will be directed to experimental application of new learning algorithm of a CMAC based hybrid-type real time controller. The proposed controller is applied for the trajectory tracking of a piezoelectric actuated tool post. It has been proven that the piezoelectric actuated tool post has hysteretic behavior. Extensive experiments have been carried out on the experimental setup to evaluate the proposed adaptive learning algorithm of CMAC. Only few experiments and their results are being presented. In these experiments, the performance of the piezoelectric actuated tool post has been examined and evaluated using different types of control algorithms and applying external load disturbance. The control performance of the proposed controller is compared with those of conventional controllers (PI controller and the conventional CMAC based controller). The experimental results showed that performance of the hybrid-type controller using the proposed learning algorithm is stable and more effective than that of the conventional controllers. Testing and comparing the learning ability of the proposed learning algorithm with that of the conventional CMAC learning algorithm indicated the effectiveness of the learning ability of the proposed algorithm. Finally, the response using the proposed hybrid-type controller is slightly better than using the conventional PI controller under the effect of external load disturbance.

Creating an applicable and precise financial early warning model is highly desirable for decision makers and regulators in the financial industry. Although Business Failure Prediction (BFP) especially banks has been extensively a researched area since late 1960s, the next critical step which is the decision making support scheme has been ignored. This paper presents a novel model for financial warning which combines a fuzzy inference system with the learning ability of neural network as a Fuzzy Neural Network (FNN) to predict organizational financial status and also applies reasoning capability of Fuzzy Case-Based Reasoning (FCBR) to support decision makers measuring appropriate solutions. The proposed financial warning model generates an adaptive fuzzy rule base to predict financial status of target case and then if it is predicted to fail, the FCBR is used to find similar survived cases. Finally according similar cases and a fuzzy rule base, the model provides financial decisions to change particular features as company goals in upcoming year to avoid future financial distress.

Cerebellar model articulation controller (CMAC) is a useful neural network learning technique. It was developed two decades ago but yet lacks an adequate learning algorithm, especially when it is used in a hybrid- type controller. This work is intended to introduce a simulation study for examining the performance of a hybrid-type control system based on the conventional learning algorithm of CMAC neural network. This study showed that the control system is unstable. Then a new adaptive learning algorithm of a CMAC based hybrid- type controller is proposed. The main features of the proposed learning algorithm, as well as the effects of the newly introduced parameters of this algorithm have been studied extensively via simulation case studies. The simulation results showed that the proposed learning algorithm is a robust in stabilizing the control system. Also, this proposed learning algorithm preserved all the known advantages of the CMAC neural network. Part II of this work is dedicated to validate the effectiveness of the proposed CMAC learning algorithm experimentally.

Medical image has now been widely used as the object in research particularly in artificial intelligent. Classification automation on medical image can assist to provide information or as the second opinion for the paramedics in doing a medical action and giving a diagnosis for the patients. One of the algorithms as the classifier is the adaptive neuro-fuzzy inference system (ANFIS) method - a hybrid method combining the fuzzy logic and neural network. ANFIS algorithm is the fuzzy inference system (FIS) implemented into the adaptive fuzzy neural network framework. This method combines the explicit knowledge as the representation from FIS and learning ability from the artificial neural network. This paper presents the discussion and the review of the adaptive neuro-fuzzy inference system (ANFIS) algorithm as the classifier in medical image classification. A number of research that have been conducted on the medical image object are evaluated and discussed in terms of their strengths and weaknesses.

In the classic approaches to design a fuzzy inference system, the fuzzy rules are determined by a domain expert a priori and then they are maintained unchanged during the learning. These fixed fuzzy rules may not be appropriate in real-time applications where the environment or model often meets unpredicted disturbances or damages. Hence, poor performance may be observed. In comparison to the conventional methods, fuzzy inference systems based on neural networks, called fuzzy-neural systems, have begun to exhibit great potential for adapting to the changes by utilizing the learning ability and adaptive capability of neural networks. Thus, a fuzzy inference system can be built using the standard structure of neural networks. Nevertheless, the determination of the number of fuzzy rules and the adjustment of the parameters in the if-then fuzzy rules are still open issues. A sequential adaptive fuzzy inference system (SAFIS) is developed to determine the number of fuzzy rules during learning and modify the parameters in fuzzy rules simultaneously. SAFIS uses the concept of influence of a fuzzy rule for adding and removing rules during learning. The influence of a fuzzy rule is defined as its contribution to the system output in a statistical sense when the input data is uniformly distributed. When there is no addition of fuzzy rules, only the parameters of the “closest” (in a Euclidean sense) rule are updated using an extended Kalman filter (EKF) scheme. The performance of SAFIS is evaluated based on some function approximation problems, via, nonlinear system identification problems and a chaotic time-series prediction problem. Results indicate that SAFIS produces similar or better accuracies with lesser number of rules compared to other algorithms.

In this paper, an optimization method based on adaptive Neuro-Fuzzy inference system (ANFIS) for determining the parameters used in the design of a coplanar capacitive coupled probe fed rectangular microstrip antenna. The antenna was analyzed in the 2-10GHz range to demonstrate universal working of the proposed model. Here, an expert knowledge of fuzzy inference system (FIS) and the learning capability of artificial neural network (ANN) have been embedded (ANFIS). By calculating and optimizing the patch dimensions of a rectangular microstrip antenna with air gap, this paper shows that ANFIS produces good results that are in agreement with the mathematical analysis of the design parameters of antenna. Of the parameters considered for optimization, the error difference (average) between the proposed model and the calculated data is 0.21% for L, 0.41% for W, and 0.2% for air gap which are less than 0.5% and acceptably low.

This paper introduces fuzzy neural network technology into the adaptive filter and makes further research on its structure and algorithms. At first, fuzzy rules are determined and the network structure is built by means of dividing fuzzy subspaces. Secondly, membership functions are chosen layers are defined and the network is trained by adaptive learning algorithm. Thirdly, training error is the minimum with repeating debugging. Finally, linking weight, the central value and width of the network membership function is adjusted by using experience of experts. The optimal performance of Adaptive Wiener Filter is realized based on Fuzzy Neural Networks.

In Section 4.5 the main features of adaptive fuzzy-neural systems were discussed. It follows that a fuzzy-neural network has a structure resulting from the combination of a fuzzy system and a neural network. It is the goal of this network to perform a required non-linear mapping from input to output spaces by using a neural network integrated with a fuzzy system (fuzzy-neural network) whose parameters are adjusted to improve performance by using a learning algorithm. Usually the design methods are hybrid design methods, originating from both fuzzy system and neural system concepts. Some of the main features are: Main aspects of neural-fuzzy system integration: the most useful features of the two system should be combined; combination should be performed in such a way that resulting integrated system should be most efficient.

Fuzzy neural network has been considered as an effective model to apply in many applications. However, due to the training mode based on minimizing the mean squared error, the typical fuzzy neural network suffers from poor robustness for disturbances. To overcome this problem, a robust fuzzy neural network with α-divergence is designed and analyzed in this paper. First, a cost function based on α-divergence is developed to describe the discrepancy between the real output and fuzzy neural network output. Then, a training mode, which minimizes the above function, can reduce the sensibility of disturbances to improve the robustness of fuzzy neural network. Second, an adaptive learning algorithm is employed to adjust the parameter of fuzzy neural network. Then, the proposed fuzzy neural network is able to obtain fast convergence in the learning process. Finally, some benchmarks are used to test the merits of fuzzy neural network. The simulation results illustrate that the proposed fuzzy neural network can achieve good robustness.

This paper presents a hybrid data mining mechanism to extract expert knowledge from historical data and extend expert systems' reasoning capabilities by using fuzzy neural network (FNN)-based learning & rule extraction algorithm. Our hybrid data mining mechanism is based on association rule extraction mechanism, FNN learning and fuzzy rule extraction algorithm. Most of traditional data mining mechanisms are depended ()n association rule extraction algorithm. However, the basic association rule-based data mining systems has not the learning ability. Therefore, there is a problem to extend the knowledge base adaptively. In addition, sequential patterns of association rules can`t represent the complicate fuzzy logic in real-world. To resolve these problems, we suggest the hybrid data mining mechanism based on association rule-based data mining, FNN learning and fuzzy rule extraction algorithm. Our hybrid data mining mechanism is consisted of four phases. First, we use general association rule mining mechanism to develop an initial rule base. Then, in the second phase, we adopt the FNN learning algorithm to extract the hidden relationships or patterns embedded in the historical data. Third, after the learning of FNN, the fuzzy rule extraction algorithm will be used to extract the implicit knowledge from the FNN. Fourth, we will combine the association rules (initial rule base) and fuzzy rules. Implementation results show that the hybrid data mining mechanism can reflect both association rule-based knowledge extraction and FNN-based knowledge extension.

Machine Learning in Engineering Analysis and Design: An Integrated Fuzzy Neural Network Learning Model

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.

An interval-valued fuzzy cerebellar model neural network (IV-FCMNN) is proposed for the identification and control of uncertain systems. It is a more general model that uses the framework of a cerebellar model neural network (CMNN) and Atanassov intuitionistic fuzzy sets (AIFSs), so that the mathematical representation of a fuzzy event is more complete and the fuzzy neural network is more general. In some special cases, this neural network can be reduced to an interval-valued fuzzy neural network (IV-FNN), a fuzzy neural network (FNN), a fuzzy cerebellar model neural network (FCMNN) or a CMNN. Since the interval-type input data are used to realize this algorithm, the IV-FCMNN copes better with uncertainty and allows greater freedom of design. Therefore, the ability to learn, the approximation precision and the fuzzy semantic description of this network are much better than those of other models. In the proposed IV-FCMNN, a training algorithm that uses a gradient descent method is proposed to adjust the parameters and convergence is proved using the Lyapunov stability theorem. The variable learning rates are analyzed and the optimal learning rates are also determined. For demonstrating the effectiveness of the proposed IV-FCMNN, three types of applications, including multiple functions approximation, multi-dimensional classification and nonlinear dynamic system feedback control, are performed and the comparison with other models are also provided.