Fuzzy ARTMAP Network Research Articles

A Modified LBP Operator-Based Optimized Fuzzy Art Map Medical Image Retrieval System for Disease Diagnosis and Prediction.

Medical records generated in hospitals are treasures for academic research and future references. Medical Image Retrieval (MIR) Systems contribute significantly to locating the relevant records required for a particular diagnosis, analysis, and treatment. An efficient classifier and effective indexing technique are required for the storage and retrieval of medical images. In this paper, a retrieval framework is formulated by adopting a modified Local Binary Pattern feature (AvN-LBP) for indexing and an optimized Fuzzy Art Map (FAM) for classifying and searching medical images. The proposed indexing method extracts LBP considering information from neighborhood pixels and is robust to background noise. The FAM network is optimized using the Differential Evaluation (DE) algorithm (DEFAMNet) with a modified mutation operation to minimize the size of the network without compromising the classification accuracy. The performance of the proposed DEFAMNet is compared with that of other classifiers and descriptors; the classification accuracy of the proposed AvN-LBP operator with DEFAMNet is higher. The experimental results on three benchmark medical image datasets provide evidence that the proposed framework classifies the medical images faster and more efficiently with lesser computational cost.

An Approach to Enhance the Design of Protein Sequence Classifier Using Data Mining

Analysis of protein plays a major role in bio-informatics domain. Protein sequence classification is most vital in this analysis of protein. In this scenario, different researchers proposed various techniques to classify protein sequence which is reviewed here at first. Different features extraction procedure from the protein sequence is a common part of all proposed classification model. But selection and arrangement of features are the most challenging task for classifying unknown protein sequences into its known families with high accuracy and preferable computational time. Analysis of recent trend points that traditional database system fails to classify a large amount of biological data which is overcome by data mining approach. New classification model involving four phases has been proposed here for classifying the unknown sequences into its family, generating knowledge-based as well as data analysis procedure. Here a hybrid algorithm with the combination of Fuzzy ARTMAP and neural network based algorithm is used. Neighbourhood analysis is also done in the final stage. This proposed model has implemented and validated with 497 different test sequences. Finally, it is also compared with the previous model and produce high accuracy level with low computation time than previous.

A framework to reduce category proliferation in fuzzy ARTMAP classifiers adopted for image retrieval using differential evolution algorithm

Image classifiers are largely adopted to categorize a pool of images or patterns in a databank, match category of a query image and to retrieve similar images to query from the category. Fuzzy ARTMAP (FAM) architecture have been widely included for pattern classification in various applications. The major constraint that limits the application of FAM network is category proliferation problem. That is the architecture has the tendency to increase the network size. The issue is because of noisy data, order of presenting training data and/or overlapping categories. In this paper, we propose a new methodology, DE-FAM to handle category proliferation problem by reducing the quantity of categories in the trained FAM architectures. The enhanced generalized performance, reduction in network size and influence of the proposed algorithm in computational cost is demonstrated by adopting the algorithm for image classification and retrieval. Furthermore the comparison of DE-FAM with other algorithms that address the category proliferation problem illustrate the advantages of DE-FAM.

An improved fuzzy ARTMAP and Q-learning agent model for pattern classification

The Fuzzy ARTMAP (FAM) network is an online supervised neural network that operates by computing the similarity level between the new sample and those prototype nodes stored in its network against a threshold. In our previous study, we have developed a multi-agent system consisting of an ensemble of FAM networks and Q-learning, known as QMACS, for data classification. In this paper, an Improved QMACS (IQMACS) model with trust measurement using a combination of Q-learning and Bayesian formalism is proposed. A number of benchmark and real-world problems, i.e., motor fault detection and human motion detection, are conducted to evaluate the effectiveness of IQMACS. Statistical features are extracted from real-world case studies and utilized for classification with IQMACS, QMACS, and their constituents. The experimental results indicate that IQMACS produces better classification performance by combining the outcomes of its constituents as compared with those of QMACS and other related methods.

Incremental Class Learning for Hierarchical Classification

Objects can be described in hierarchical semantics, and people also perceive them this way. It leads to the need for hierarchical classification in machine learning. On the other hand, when a new data that belongs to a new class is given, the existing classification methods should be retrained for all data including the new data. To deal with these issues, we propose an adaptive resonance theory-supervised predictive mapping for hierarchical classification (ARTMAP-HC) network that allows incremental class learning for raw data without normalization in advance. Our proposed ARTMAP-HC is composed of hierarchically stacked modules, and each module incorporates two fuzzy ARTMAP networks. Regardless of the level of the class hierarchy and the number of classes for each level, ARTMAP-HC is able to incrementally learn sequentially added input data belonging to new classes. By using a novel online normalization process, ARTMAP-HC can classify the new data without prior knowledge of the maximum value of the dataset. By adopting the prior labels appending process, the class dependency between class hierarchy levels is reflected in ARTMAP-HC. The effectiveness of the proposed ARTMAP-HC is validated through experiments on hierarchical classification datasets. To demonstrate the applicability, ARTMAP-HC is applied to a multimedia recommendation system for digital storytelling.

A hybrid FAM‐CART model for online data classification

AbstractIn this paper, an online soft computing model based on an integration between the fuzzy ARTMAP (FAM) neural network and the classification and regression tree (CART) for undertaking data classification problems is presented. Online FAM network is useful for conducting incremental learning with data samples, whereas the CART model prevails in depicting the knowledge learned explicitly in a tree structure. Capitalizing on their respective advantages, the hybrid FAM‐CART model is capable of learning incrementally while explaining its predictions with knowledge elicited from data samples. To evaluate the usefulness of FAM‐CART, 2 sets of benchmark experiments with a total of 12 problems are used in both offline and online learning modes. The results are examined and compared with those published in the literature. The experimental outcome positively indicates that the online FAM‐CART model is useful for tackling data classification tasks. In addition, a decision tree is produced to allow users in understanding the predictions, which is an important property of the hybrid FAM‐CART model in supporting decision‐making tasks.

Use of Savitzky-Golay Filter for Performances Improvement of SHM Systems Based on Neural Networks and Distributed PZT Sensors.

A considerable amount of research has focused on monitoring structural damage using Structural Health Monitoring (SHM) technologies, which has had recent advances. However, it is important to note the challenges and unresolved problems that disqualify currently developed monitoring systems. One of the frontline SHM technologies, the Electromechanical Impedance (EMI) technique, has shown its potential to overcome remaining problems and challenges. Unfortunately, the recently developed neural network algorithms have not shown significant improvements in the accuracy of rate and the required processing time. In order to fill this gap in advanced neural networks used with EMI techniques, this paper proposes an enhanced and reliable strategy for improving the structural damage detection via: (1) Savitzky–Golay (SG) filter, using both first and second derivatives; (2) Probabilistic Neural Network (PNN); and, (3) Simplified Fuzzy ARTMAP Network (SFAN). Those three methods were employed to analyze the EMI data experimentally obtained from an aluminum plate containing three attached PZT (Lead Zirconate Titanate) patches. In this present study, the damage scenarios were simulated by attaching a small metallic nut at three different positions in the aluminum plate. We found that the proposed method achieves a hit rate of more than 83%, which is significantly higher than current state-of-the-art approaches. Furthermore, this approach results in an improvement of 93% when considering the best case scenario.

Kappa-PSO-FAN based method for damage identification on composite structural health monitoring

Recently much research has been conducted towards finding fast and accurate pattern classifiers applied to Structural Health Monitoring (SHM) systems. In this way, researchers have proposed new methods based on Fuzzy ARTMAP Network (FAN) in order to enhance the success rate for structural damage classification applied to SHM applications. Conversely, the performance of methods based on FAN is very dependent of its setup parameters. In several SHM approaches in the literature, authors have proposed selecting those parameters by using several attempts (empirical and manual selection) and keeping them fixed for all cases in the resulting analysis, hampering the success rate of the neural network. To overcoming that, this paper introduces a new strategy for enhancement of structural damage identification focusing on supervised learning of FAN by using Particle Swarm Optimization (PSO) for selecting optimal setup parameters automatically for the FAN algorithm. Also, the Kappa coefficient is used as an objective function to be maximized through the PSO algorithm. As a result, the optimum setup parameters improved the success rate while the damage identification is being carried out. Indeed this proposed method is certainly very promising and constitutes a novelty. The proposed method achieves more than 75% hit rate that is significantly higher than the state-of-the-art approaches as presented in this paper. Furthermore, this approach yields a 20% improvement when considering the worst case scenario. Hence, this approach shows a practical application of expert and intelligent systems applied to damage identification in SHM systems. To conclude, the proposed approach successfully identifies structural damage with accuracy and efficiency.

Performance analysis of simplified Fuzzy ARTMAP and Probabilistic Neural Networks for identifying structural damage growth

Structural Health Monitoring (SHM) has been advancing worldwide as evidenced by recent papers and practical applications. SHM systems increase safety and reduce maintenance costs in a variety of applications contributing significantly to prevent structural failures. Studies focusing on the structural damage identification have been proposed in literature by using methods based on Probabilistic Neural Networks (PNN). Although Fuzzy based methods are recurrent in SHM approaches, the Simplified Fuzzy ARTMAP Network (SFAN) has not been explored for analyzing structural damage. Furthermore, there is no evidence of comparative studies of these two methods when applied to the problem of progressing structural damage. Thus, this paper presents a comparative analysis of these two methods in the context of identifying structural damage growth. The comparison is made based on the factors of success rate and training/testing times. Additionally, this approach has carried out the performance analysis of suitable parameters setup for SFAN: choice parameter, training rate and vigilance parameter. As a practical case study, both methods were applied to a unidirectional composite plate containing four PZT (Lead Zirconate Titanate) patches where the damage growth scenarios were also simulated by loosening bolts for three different levels. In addition, the repaired structural condition was also considered by retightening bolts. The results have shown that both methods are suitable for the problem of damage growth, particularly to supporting decisions about the structural damage assessment. In short, the comparative analysis has shown that the SFAN method is better suited to the problem of damage growth especially in regard to training and testing times. Thus, this comparative study contributes to helping researchers and industrialist in choosing more effective approaches for structural damage growth.

Simplified fuzzy ARTMAP network-based method for assessment of structural damage applied to composite structures

This paper addresses a novel method for structural damage assessment based on the electromechanical impedance (EMI) technique applied to composite structures. This method is based on a high-frequency excitation range in order to overcome the difficulties caused when the low vibration modes are excited. A structure made up of unidirectional carbon fiber pre-impregnated with resin-epoxy is excited using a chirp signal which varies the frequency in a wide range through the PZT (Lead Zirconate Titanate) patches. From the structural response signal, the Euclidean Distance (ED) is computed from each structural condition considering the pristine condition as a reference (baseline). The ED is used as input to the Simplified Fuzzy ARTMAP Network (SFAN) which is responsible for identifying various different structural conditions. It is shown that the robustness provided by SFAN allows the evaluation of the progress of structural damage even under critical conditions. The structural damage was simulated by loosening bolts in a structure. Furthermore, the repaired condition was also considered by retightening bolts. The paper discusses the advantages and drawbacks of the approach in light of the experimental results.

Application of intelligent tools to detect and classify broken rotor bars in three‐phase induction motors fed by an inverter

A comprehensive study of intelligent tools used to classify broken rotor bars in induction motors, which operate with three different types of frequency inverters, is presented. The diagnosis of defective rotor bars is a critical issue for the predictive maintenance of induction motors. A proper classification of these defects in their early stages of evolution is necessary for preventing major machine failures and production downtime. The proposed approach is performed by analysing the amplitude of the stator current signal in the time domain, using a dynamic acquisition rate based on machine frequency supply. To assess classification accuracy under the various severity levels of the faults, the performance of four different learning machine techniques is investigated: (i) fuzzy ARTMAP network; (ii) support vector machine (sequential minimal optimisation); (iii) k-nearest neighbour; and (iv) multilayer perceptron network. Results obtained from 1274 experimental tests are presented in order to validate the study, which considers a wide range of load conditions and operating frequencies. Experimental results presented in this study validate the robustness and efficacy of the proposed approach.

Evaluation of stator winding faults severity in inverter-fed induction motors

Three-phase induction motor are one of the most important elements of electromechanical energy conversion in the production process. However, they are subject to inherent faults or failures under operating conditions. The purpose of this paper is to present a comparative study among intelligent tools to classify short-circuit faults in stator windings of induction motors operating with three different models of frequency inverters. This is performed by analyzing the amplitude of the stator current signal in the time domain, using a dynamic acquisition rate according to machine frequency supply. To assess the classification accuracy across the various levels of faults severity, the performance of three different learning machine techniques were compared: (i) fuzzy ARTMAP network; (ii) multilayer perceptron network; and (iii) support vector machine. Results obtained from 2.268 experimental tests are presented to validate the study, which considered a wide range of operating frequencies and load conditions.

Function Approximation with ARTMAP Architectures

We analyze function approximation (regression) capability of Fuzzy ARTMAP (FAM) architectures - well-known incremental learning neural networks. We focus especially on the universal approximation property. In our experiments, we compare the regression performance of FAM networks with other standard neu- ral models. It is the first time that ARTMAP regression is overviewed, both from theoretical and practical points of view.

Guest editorial: Special issue on advances in intelligent data processing and analysis

In today’s information era, data (especially digital data) are generated at an unprecedented rate in terms of volume and velocity. Data gathered from the real world are often multifaceted, e.g. spatiotemporal data, multivariate data, and multimodal data, and exist in a variety of formats, e.g. structured/unstructured text, time-/frequency-domain signals, static images, and dynamic video streams. The available data samples also need to be processed carefully, e.g. cleansing, filtering, and transforming, before useful information can be extracted for inferring conclusions and supporting decision-making. Intelligent techniques stemming from the domain of soft computing, which include neural, fuzzy, and evolutionary computing methodologies and other related techniques, provide a viable approach to either automatically or semiautomatically process and analyse a large volume of data. Neural computing techniques are useful for processing low-level data samples and extracting information out from data samples, while fuzzy computing techniques are beneficial for analysing high-level human linguistic variables and performing inference/reasoning with if–then rules elicited from human experts. On the other hand, evolutionary computing techniques are capable of searching in a high-dimensional space and finding solutions for optimization problems. In this special issue, a total of nine of articles reporting recent research findings in intelligent data processing and analysis from either theoretical or practical perspectives are presented. It should be noted that these articles represent only a small fraction of researches in this fast-moving domain. We aspire that this special issue could provide useful insight and stir further research in the area of intelligent data processing and analysis. A summary of each article is as follows. In Knauer et al., an ensemble of Radial Basis Function (RBF) neural networks with c-divergence-based similarity measures is first formulated. Fusion of the RBF outputs with decision trees is then performed. In addition, a selection scheme for subsets of RBF networks based on their relevance in the fusion process is proposed. A number of hyperspectral imaging data sets are deployed for evaluation. Different tree-based learners and combination strategies, which include AdaBoost with decision trees, random forests, and pruned decision trees, are experimented. The results demonstrate the usefulness of the proposed fusion tree approach to combining multiple RBF outputs in deriving an accurate classification system. The use of Genetic Algorithm (GA) to design an appropriate structure of the Fuzzy ARTMAP (FAM) neural network is described in Loo et al. The network parameters and the order of training data are first determined by a GA. Then, an ensemble of FAM networks is optimized with another GA to improve its performance in undertaking data classification problems. Probabilistic voting is employed to combine the predictions from multiple FAM networks. The usefulness of the proposed model is demonstrated using benchmark problems from the UCI Machine Learning repository. To reduce patient waiting time and to improve quality of healthcare, Wang et al. adopt the Fuzzy Min–Max (FMM) C. P. Lim Deakin University, Waurn Ponds, Victoria, Australia

A Hybrid Intelligent Model for Deterministic and Quantile Regression Approach for Probabilistic Wind Power Forecasting

With rapid increase in wind power penetration into the power grid, wind power forecasting is becoming increasingly important to power system operators and electricity market participants. The majority of the wind forecasting tools available in the literature provide deterministic prediction, but given the variability and uncertainty of wind, such predictions limit the use of the existing tools for decision-making under uncertain conditions. As a result, probabilistic forecasting, which provides information on uncertainty associated with wind power forecasting, is gaining increased attention. This paper presents a novel hybrid intelligent algorithm for deterministic wind power forecasting that utilizes a combination of wavelet transform (WT) and fuzzy ARTMAP (FA) network, which is optimized by using firefly (FF) optimization algorithm. In addition, support vector machine (SVM) classifier is used to minimize the wind power forecast error obtained from WT+FA+FF. The paper also presents a probabilistic wind power forecasting algorithm using quantile regression method. It uses the wind power forecast results obtained from the proposed hybrid deterministic WT+FA+FF+SVM model to evaluate the probabilistic forecasting performance. The performance of the proposed forecasting model is assessed utilizing wind power data from the Cedar Creek wind farm in Colorado.

Locating Pd in Transformers through Detailed Model and Neural Networks

Abstract In a power transformer as one of the major component in electric power networks, partial discharge (PD) is a major source of insulation failure. Therefore the accurate and high speed techniques for locating of PD sources are required regarding to repair and maintenance. In this paper an attempt has been made to introduce the novel methods based on two different artificial neural networks (ANN) for identifying PD location in the power transformers. In present report Fuzzy ARTmap and Bayesian neural networks are employed for PD locating while using detailed model (DM) for a power transformer for simulation purposes. In present paper PD phenomenon is implemented in different points of transformer winding using threecapacitor model. Then impulse test is applied to transformer terminals in order to use produced current in neutral point for training and test of employed ANNs. In practice obtained current signals include noise components. Thus the performance of Fuzzy ARTmap and Bayesian networks for correct identification of PD location in a noisy condition for detected currents is also investigated. In this paper RBF learning procedure is used for Bayesian network, while Markov chain Monte Carlo (MCMC) method is employed for training of Fuzzy ARTmap network for locating PD in a power transformer winding and results are compared.

A Novel Hybrid Approach Based on Wavelet Transform and Fuzzy ARTMAP Networks for Predicting Wind Farm Power Production

This paper presents a novel hybrid intelligent algorithm based on the wavelet transform (WT) and fuzzy ARTMAP (FA) network for forecasting the power output of a wind farm utilizing meteorological information such as wind speed, wind direction, and temperature. The prediction capability of the proposed hybrid WT +FA model is demonstrated by an extensive comparison with a benchmark persistence method, other soft computing models, and hybrid models as well. The test results show a significant improvement in forecasting error through the application of a proposed hybrid WT +FA model. The proposed hybrid wind power forecasting strategy is applied to real-life data from Kent Hill wind farm located in New Brunswick, Canada.

Wind speed forecast model for wind farm based on a hybrid machine learning algorithm

This paper presents a new strategy for wind speed forecasting based on a hybrid machine learning algorithm, composed of a data filtering technique based on wavelet transform (WT) and a soft computing model based on the fuzzy ARTMAP (FA) network. The prediction capability of the proposed hybrid WT+FA model is demonstrated by an extensive comparison with some other existing wind speed forecasting methods. The results show a significant improvement in forecasting error through the application of a proposed hybrid WT+FA model. The proposed wind speed forecasting strategy is applied to real data acquired from the North Cape wind farm located in PEI, Canada.

An Embedded Simplified Fuzzy ARTMAP Implemented on a Microcontroller for Food Classification

In the present study, a portable system based on a microcontroller has been developed to classify different kinds of honeys. In order to do this classification, a Simplified Fuzzy ARTMAP network (SFA) implemented in a microcontroller has been used. Due to memory limits when working with microcontrollers, it is necessary to optimize the use of both program and data memory. Thus, a Graphical User Interface (GUI) for MATLAB® has been developed in order to optimize the necessary parameters to programme the SFA in a microcontroller. The measures have been carried out by potentiometric techniques using a multielectrode made of seven different metals. Next, the neural network has been trained on a PC by means of the GUI in Matlab using the data obtained in the experimental phase. The microcontroller has been programmed with the obtained parameters and then, new samples have been analysed using the portable system in order to test the model. Results are very promising, as an 87.5% recognition rate has been achieved in the training phase, which suggests that this kind of procedures can be successfully used not only for honey classification, but also for many other kinds of food.

A novel hybrid approach using wavelet, firefly algorithm, and fuzzy ARTMAP for day-ahead electricity price forecasting

This paper presents a novel hybrid intelligent algorithm utilizing a data filtering technique based on wavelet transform (WT), an optimization technique based on firefly (FF) algorithm, and a soft computing model based on fuzzy ARTMAP (FA) network in order to forecast day-ahead electricity prices in the Ontario market. A comprehensive comparative analysis with other soft computing and hybrid models shows a significant improvement in forecast error by more than 40% for daily and weekly price forecasts, through the application of a proposed hybrid WT+FF+FA model. Furthermore, low values obtained for the forecast mean square error (FMSE) and mean absolute error (MAE) indicate high degree of accuracy of the proposed model. Robustness of the proposed hybrid intelligent model is measured by using the statistical index (error variance). In addition, the good forecast performance and the rapid adaptability of the proposed hybrid WT+FF+FA model are also evaluated using the PJM market data.