Fuzzy ARTMAP Neural Network Research Articles

Development of a Short-Term Electrical Load Forecasting in Disaggregated Levels Using a Hybrid Modified Fuzzy-ARTMAP Strategy

In recent years, electrical systems have evolved, creating uncertainties in short-term economic dispatch programming due to demand fluctuations from self-generating companies. This paper proposes a flexible Machine Learning (ML) approach to address electrical load forecasting at various levels of disaggregation in the Peruvian Interconnected Electrical System (SEIN). The novelty of this approach includes utilizing meteorological data for training, employing an adaptable methodology with easily modifiable internal parameters, achieving low computational cost, and demonstrating high performance in terms of MAPE. The methodology combines modified Fuzzy ARTMAP Neural Network (FAMM) and hybrid Support Vector Machine FAMM (SVMFAMM) methods in a parallel process, using data decomposition through the Wavelet filter db20. Experimental results show that the proposed approach outperforms state-of-the-art models in predicting accuracy across different time intervals.

BepFAMN: A Method for Linear B-Cell Epitope Predictions Based on Fuzzy-ARTMAP Artificial Neural Network.

The public health system is extremely dependent on the use of vaccines to immunize the population from a series of infectious and dangerous diseases, preventing the system from collapsing and millions of people dying every year. However, to develop these vaccines and effectively monitor these diseases, it is necessary to use accurate diagnostic methods capable of identifying highly immunogenic regions within a given pathogenic protein. Existing experimental methods are expensive, time-consuming, and require arduous laboratory work, as they require the screening of a large number of potential candidate epitopes, making the methods extremely laborious, especially for application to larger microorganisms. In the last decades, researchers have developed in silico prediction methods, based on machine learning, to identify these markers, to drastically reduce the list of potential candidate epitopes for experimental tests, and, consequently, to reduce the laborious task associated with their mapping. Despite these efforts, the tools and methods still have low accuracy, slow diagnosis, and offline training. Thus, we develop a method to predict B-cell linear epitopes which are based on a Fuzzy-ARTMAP neural network architecture, called BepFAMN (B Epitope Prediction Fuzzy ARTMAP Artificial Neural Network). This was trained using a linear averaging scheme on 15 properties that include an amino acid ratio scale and a set of 14 physicochemical scales. The database used was obtained from the IEDB website, from which the amino acid sequences with the annotations of their positive and negative epitopes were taken. To train and validate the knowledge models, five-fold cross-validation and competition techniques were used. The BepiPred-2.0 database, an independent database, was used for the tests. In our experiment, the validation dataset reached sensitivity = 91.50%, specificity = 91.49%, accuracy = 91.49%, MCC = 0.83, and an area under the curve (AUC) ROC of approximately 0.9289. The result in the testing dataset achieves a significant improvement, with sensitivity = 81.87%, specificity = 74.75%, accuracy = 78.27%, MCC = 0.56, and AOC = 0.7831. These achieved values demonstrate that BepFAMN outperforms all other linear B-cell epitope prediction tools currently used. In addition, the architecture provides mechanisms for online training, which allow the user to find a new B-cell linear epitope, and to improve the model without need to re-train itself with the whole dataset. This fact contributes to a considerable reduction in the number of potential linear epitopes to be experimentally validated, reducing laboratory time and accelerating the development of diagnostic tests, vaccines, and immunotherapeutic approaches.

Application of Deep Learning Technology in Predicting the Risk of Inpatient Death in Intensive Care Unit.

The Intensive Care Unit (ICU) is an important unit for the rescue of critically ill patients in hospitals, and patient mortality is an important indicator to measure the level of ICU treatment. Currently, a variety of clinical scoring systems are used to evaluate the patient's condition and predict survival, but these systems require a lot of resources. However, due to the rapid development of artificial intelligence and deep learning, machine learning based methods have been used to study the survival prediction of ICU patients. Additionally, these methods have made significant progress, but there is still a distance from clinical application, and equally metric interpretability of the deep learning method is not very mature. Therefore, in this paper, we have proposed a predicting model for the life and death of ICU patients, which is primarily based on the Fuzzy ARTMAP model. With a thorough analysis of the existing ICU patient condition assessment and life and death prediction methods, we have observed that patient's ICU monitoring information performs integrated analysis and extracts features according to the clinical characteristics of physiological indicators. Finally, fuzzy ARTMAP neural network is used to predict the life and death of patients. Likewise, prediction results are combined with the clinical scoring system and logistic regression, artificial neural network, support vector machine, and AdaBoost. Experimental results of these algorithms were compared, which verifies that the proposed method has outperformed the existing model. The main purpose of the proposed mode is to design a life and death prediction method for ICU patients, which has high predictive performance and is an acceptable method for clinical medical staff, where ICU monitoring data is used. Experimental results show that the method proposed has achieved better prediction performance and accuracy ratio, which provide theoretical reference for clinical application.

Fault Classification in Power Distribution Systems Using Multiresolution Analysis and a Fuzzy-ARTMAP Neural NetworkAnalysis and a Fuzzy-ARTMAP Neural Network

This paper presents a tool for the detection and classification of short-circuits in electric power distribution systems, which is based on the combined use of multi-resolution analysis and Fuzzy-ARTMAP neural network. The multiresolution analysis allows the identification of singularities in the waveforms and the ART family network guarantees to the classifier the ability to continuously learn without losing the previously acquired knowledge. The entire diagnosis procedure is performed in a single step, reducing the computational effort. The efficiency of the system is verified by a direct accuracy analysis, and by a comparison between the Fuzzy-ARTMAP and the Euclidean-ARTMAP. Results show that the system is efficient and robust, being able to detect and classify 100% of the electrical faults considering the test that were performed.

A new approach to online training for the Fuzzy ARTMAP artificial neural network

The evolution of internet resources has led to an increase in the flow of data and, consequently, the need for classification or forecasting models that support online learning. The Fuzzy ARTMAP neural network has been used in the most areas of knowledge; however, few have explored real-time applications that require continuous training. In this work, a Fuzzy ARTMAP neural network with continuous training is proposed. This new network can acquire knowledge via classification or prediction. Modifications made to the architecture and learning algorithm enable online learning from the first sample of data and perform the classification or forecasting at any time during training. To validate the proposed model, three experiments were performed, one for forecasting and two for classification. Each experiment used benchmark databases and compared its final results with the results of the original Fuzzy ARTMAP neural network. The results demonstrate the ability of the proposed model to acquire knowledge from the first data samples in a stable and efficient way. Thus, this study contributes to the evolution of the Fuzzy ARTMAP neural network and introduces the continuous training method as an effective alternative to real-time applications.

Fuzzy ARTMAP-Based Fast Object Recognition for Robots Using FPGA

Fast object recognition and classification is highly important when handling operations with robots. This article shows the design and implementation of an invariant recognition machine vision system to compute a descriptive vector called the Boundary Object Function (BOF) using the FuzzyARTMAP (FAM) Neural Network. The object recognition machine is integrated in the Zybo Z7-20 module that includes reconfigurable FPGA hardware and a RISC processor. Object encoding, description and prediction is carried out rapidly compared to the processing time devoted to video capture at the camera’s frame rate. Benefiting from parallel computing, we calculated the object’s centroid and boundary points while acquiring the progressive image frame; all that was done with the intention of readying it for neural processing. The remaining time was devoted to recognising the object, this caused low latency (1.47 ms). Our test-bed also included TCP/IP communication to send/receive part location for grasping operations with an industrial robot to evaluate the approach. Results demonstrate that the hardware integration of the video sensor, image processing, descriptor generator, and the ANN classifier for cognitive decision on a single chip can increase the speed and performance of intelligent robots designed for smart manufacturing.

Hyperspectral Unmixing Based on Spectral and Sparse Deep Convolutional Neural Networks

Hyperspectral unmixing refers to the process of obtaining endmembers and abundance vectors through linear or nonlinear models. The traditional linear unmixing model assumes that each mixed pixel can be represented by a linear combination of endmembers. Considering real-world situations, a sparse constraint is normally added to the linear unmixing model. However, the linear model does not take into account that the spectrum of mixed pixels is not simply linearly mixed. To fully study the mixing characteristics of ground object spectra before being imaged by the sensor, we propose a supervised unmixing architecture based on one-dimensional convolutional neural network (CNN) by considering the spectral information and the sparse characteristics in the mixed pixel. Since one-dimensional CNN only consider feature learning, we combine the traditional root mean square error (RMSE) and l_"1" regularization in its loss function to minimize training error. The performance of our proposed unmixing model is assessed by comparing the unmixing results with three traditional linear sparse unmixing algorithms, and the Fuzzy-ARTMAP neural network in a simulated dataset and three real datasets. The RMSE was used to verify the unmixing accuracy of the different methods. The results showed that the RMSE obtained by our proposed CNN-based method was the lowest among the methods on all three real datasets, proving the effectiveness and stability of CNN in unmixing tasks.

Impact wave predictions by a Fuzzy ARTMAP neural network

Impact waves caused by landslides, rock blocks, and avalanches in lakes and dam reservoirs have provoked several disasters since the dawn of time; further, they have caused considerable number of deaths and property loss. Understanding the behavior and characteristics of these waves, principally their height, propagation velocity, and energy, are key for determining engineering parameters, and therefore developing alert systems and evacuation plans. Several studies have investigated impact waves using experimental, mathematical, and numerical models; however, few have used artificial neural networks. Considering the learning and prediction characteristics of neural networks, this work explores their application in the prediction of impact waves. The main objective is to verify the capability of a fuzzy ARTMAP neural network in predicting the evolution of the maximum wave height, one of the main parameters of impact waves. The experimental data used in this paper are from the work developed by Huber (1980) when studying the impact waves generated by dropping deformable material (granular) in a channel. This study uses different forms of normalization, as well as different training parameters. The Fuzzy ARTMAP neural network predicts adequately the evolution of the maximum wave heights, becoming useful for this kind of application.

A new formulation of multinodal short-term load forecasting based on adaptive resonance theory with reverse training

A multinodal intelligent predictive method for electrical power systems has been developed. Knowing the electrical load accurately and in advance is essential for conducting studies in regard to the system operations, and to create strategies that improve the quality of the energy-supply for commercial, industrial, and residential consumers. The proposed method employs a supervised Fuzzy-ARTMAP neural network, using the new concept of reverse training, to forecast the global demand and load of several nodes of an electric network (multinodal load forecasting) up to 24 h ahead. To evaluate and test the proposed system, an application is presented that considers real historical data from a company in the electric sector. Results show that the reverse training reduces the error of the neural network, making the forecast more accurate, reliable, and very fast.

Intelligent mobile robot navigation using a neuro-fuzzy approach

This paper introduces an intelligent navigation system allowing a car-like robot to attain its destination autonomously, intelligently and safely. Based on a neuro-fuzzy (FNN) approach, the applied technique permits the robot to avoid all encountered obstacles and seek for its target's location in a local manner referring to the concepts of learning and adaptation. It uses two fuzzy Artmap neural networks, a reinforcement trial and error neural network and a Mamdani fuzzy logic controller (FLC). Experimental results in the Generator of modules (GenoM) robotics architecture, in an unknown environment, shows the FNN effectiveness for the autonomous mobile robot Robucar.

Intelligent mobile robot navigation using a neuro-fuzzy approach

This paper introduces an intelligent navigation system allowing a car-like robot to attain its destination autonomously, intelligently and safely. Based on a neuro-fuzzy (FNN) approach, the applied technique permits the robot to avoid all encountered obstacles and seek for its target's location in a local manner referring to the concepts of learning and adaptation. It uses two fuzzy Artmap neural networks, a reinforcement trial and error neural network and a Mamdani fuzzy logic controller (FLC). Experimental results in the Generator of modules (GenoM) robotics architecture, in an unknown environment, shows the FNN effectiveness for the autonomous mobile robot Robucar.

Open-set single-sample face recognition in video surveillance using fuzzy ARTMAP

Single-sample face recognition has been investigated by a few researches over the past few decades. However, due to the demand of identity of interest searching from video surveillance in recent years, this system has been expanded to open-set face recognition (OSFR) scheme. The OSFR system provides the identity of registered subjects and rejects the unregistered ones using only single-sample reference. This is important in video surveillance applications in which both database members and non-members are expected to appear in the scene. In this paper, we propose to use fuzzy ARTMAP neural network to solve the problem of open-set single-sample face recognition in real-world video surveillance scenario. Our proposed approach can recognize faces in near-frontal views under various illumination and facial expression conditions. Facial features are extracted using histograms of oriented gradients and Gabor wavelets and then fused using canonical correlation analysis to yield feature vectors that are robust against the aforementioned conditions. The fuzzy ARTMAP classifier has been trained using only single sample per person. We have conducted experiments on three challenging benchmark datasets, namely AR, FRGC, and ChokePoint. The experimental results have shown that the proposed approach has a superior performance than the state-of-the-art approaches.

Multinodal load forecasting for distribution systems using a fuzzy-artmap neural network

This work proposes a predictor system (multinodal forecasting) considering several points of an electrical network, such as substations, transformers, and feeders, based on an adaptive resonance theory (ART) neural network family. It is a problem similar to global forecasting, with the main difference being the strategy to align the input and output of the data with several parallel neural modules. Considering that multinodal prediction is more complex compared to global prediction, the multinodal prediction will use a fuzzy-ARTMAP neural network and a global load participation factor. The advantages of this approach are as follows: (1) the processing time is equivalent to the processing required for global forecasting (i.e., the additional time processing is quite low); and (2) Fuzzy-ARTMAP neural networks converge significantly faster than backpropagation neural networks (improved benchmark in precision). The preference for neural networks of the ART family is due to the characteristic stability and plasticity that these architectures have to provide results in a fast and precise way. To test the proposed forecast system, the results are presented for nine substations from the database of an electrical company.

Assessment of the fuzzy ARTMAP neural network method performance in geological mapping using satellite images and Boolean logic

Currently, in executive comparative studies and even in the research studies of natural resources, the use of maps produced by the geological survey forms the basis of geological studies. However, in recent years, remote sensing technology has been introduced as a new and efficient tool for geological studies which in addition to the accuracy has other benefits such as access to the arduous mountainous areas or inaccessible that can help us to prepare geological maps with more accuracy. The purpose of this study is to compare geological survey geological maps with a scale of 1: 100,000, and produced and modified by Google Earth maps by using fuzzy ARTMAP artificial neural network method. For this purpose, the geological map of a part of Yazd-Shirkooh watershed prepared by using Landsat 7 image by fuzzy ARTMAP artificial neural network method with Kappa coefficient of 89%. Also, geological survey and Google Earth map prepared. Graphic map obtained from the visual interpretation of Google Earth images with the ground control was considered as the base map and comparing it with other maps evaluated by using Boolean logic. By using the sampling network created 100 points which 62 points were in the study area and 56 cases in the fuzzy ARTMAP map and 52 cases in the geological survey map were consistent with the base map. Due to reduced cost and time and no limits of time and space, fuzzy ARTMAP is a suitable method for the preparation of geological maps.

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.

Object-wise joint-classification change detection for remote sensing images based on entropy query-by fuzzy ARTMAP

The pixel-wise post-classification comparison (PCC) method is widely used in remote sensing images change detection. However, it is affected by the significant cumulative error caused by single image classification error. What’s more, the pixel-wise change detection method always produces “salt and pepper” effect. To solve the excessive evaluation of changed types and quantity caused by cumulative error and “salt and pepper” effect, a novel remote sensing image change detection method called entropy query-by fuzzy ARTMAP object-wise joint classification comparison (EQFAM-OBJCC) is presented in this article. Firstly, entropy query-by measurement of active learning is integrated with the fuzzy ARTMAP neural network to choose training samples which contain large amounts of information to improve the classification accuracy. Secondly, joint classification comparison is introduced to obtain the pixel-wise classification results. Finally, the object-wise classification and change detection results are produced by superpixel segmentation method, majority voting rule, and comparison of each superpixels. Experimental results demonstrate the validity of the proposed method. The classification and change detection results show that the proposed method can reduce the cumulative error with an average classification accuracy of 94.12% and a total detection error of 27.03%, and effectively resolve the “salt and pepper” problem. The proposed method was used to monitor the reclamation status of Liaohe estuary wetland via 10 time series remote sensing images from 1987 to 2014.

The investigation of spatiotemporal variations of land surface temperature based on land use changes using NDVI in southwest of Iran

Land use changes can bring about changes in land surface temperature (LST) which is influenced by climatic conditions and physical characteristics of the land surface. In this study, spatiotemporal variations of land surface temperature have been investigated in the desert area of Dasht-e-Abbas, Ilam, based on a variety of land use changes. The investigated periods for the study include 1990, 2000 and 2010 using Landsat image data. First, in mapping land use we used the Fuzzy ARTMAP Neural Network Classification method followed by determination of the NDVI Index to estimate land surface temperature. The results show an increase in LST in areas where degradation, land use and land cover changes have occurred. In 1990, 2000 and 2010, the average land surface temperature of the Fair Rangelands was 26.72 °C, 30.06 °C and 30.95 °C, respectively. This rangeland has been reduced by about 5%. For poor rangelands, the average LSTs were 26.95, 32.83 and 34.49 Cº, respectively which had a 18% reduction. In 1990, 2000 and 2010, the average land surface temperatures of agricultural lands were 24.31 °C, 27.87 °C and 28.61 °C, respectively which has been an increasing trend. The reason can be attributed to changes in cropping patterns of the study area.

Classification of Characters Using Multilayer Perceptron and Simplified Fuzzy ARTMAP Neural Networks

Classification of Characters Using Multilayer Perceptron and Simplified Fuzzy ARTMAP Neural Networks

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.

Integrated Fault Location and Power-Quality Analysis in Electric Power Distribution Systems

This paper presents a methodology for automated disturbance analysis and fault location on electric power distribution systems using a combination of modern techniques for network analysis, signal processing, and intelligent systems. New algorithms to detect, classify, and locate power-quality disturbances are developed. The continuous process of detecting these disturbances is accomplished through statistical analysis and multilevel signal analysis in the wavelet domain. The behavioral indices of the current and voltage signals are extracted by employing the discrete wavelet transform, multiresolution analysis, and the concept of signal energy. These indices are used by a number of independent Fuzzy-ARTMAP neural networks, which aim to classify the fault type and the power-quality events. The fault location is performed after the classification process. A real life three-phase distribution system with 134 nodes—13.8 kV and 7.065 MVA—was used to test the proposed algorithms, providing satisfactory results, attesting that the proposed algorithms are efficient, fast, and, above all, intelligent.