Prostate cancer grading, using the International Society of Urological Pathology (ISUP) system, for treatment decisions is highly subjective and requires considerable expertise. Despite advances in computer-aided diagnosis systems, few have handled efficient ISUP grading on whole slide images (WSIs) of prostate biopsies based only on slide labels. In this scenario, TSOR is developed, where a novel task-specific self-supervised learning (SSL) framework is used for patch-level pretraining. This is fine-tuned using ordinal regression for WSI-level ISUP grading. One of the main challenges faced by deep learning (DL) in ISUP grading, is the learning of patch-level features based on slide labels. Though using models pretrained at patch-level using SSL or other paradigms is the most obvious choice here, the diversity of training samples plays a crucial role in effective pretraining. However, pretraining on a large database of different histopathology images becomes computationally expensive. Therefore, a patch-level dataset (relatively balanced with respect to the patch-level grades) is initially created for effective SSL-based pretraining. As stain-variation across centers leads to difficulty in generalization, an additional loss term is incorporated to effectively learn the stain-agnostic patch-level features. As it is desirable that misclassification be as close as possible to the actual grade, in medical images, we fine-tune the pretrained network for WSI-level ISUP grading using an ordinal regression-based approach. Experimental results on the most extensive prostate cancer grade assessment (PANDA) challenge dataset, and the SICAPv2 dataset, demonstrate the effectiveness of TSOR compared to state-of-the-art (SOTA) methods.

In this article, the memristor-based selective convolutional (MSC) circuit for salt-and-pepper (SAP) noise removal was proposed. In experiments, the MSC model was built and benchmarked against a ternary selective convolutional (TSC) model. Results show that the MSC model effectively restores images corrupted by SAP noise, achieving similar performance to the TSC model in both quantitative measures and visual quality at noise densities of up to 50%. In addition, this study proposes an enhanced MSC (MSCE) model based on MSC, which reduces power consumption by 57.6% compared with the MSC model while improving performance. The MSCE model maintains reliability when memristors experience conductance drift rates of less than 30% and yields greater than 89%.

Recent medical image segmentation methods extract the characteristics of anatomical structures only from the spatial domain, ignoring the distinctive patterns present in the spectral representation. This study aims to develop a novel segmentation architecture that leverages both spatial and spectral characteristics for better segmentation outcomes. This research introduces the wavelet-infused convolutional Transformer (WaveCoformer), a computationally effective framework to fuse information from both spatial and spectral domains of medical images. Fine-grained textural features are captured from the wavelet components by the convolution module. A transformer block identifies the relevant activation maps within the volumes, followed by self-attention to effectively learn long-range dependencies to capture the global context of the target regions. A cross-attention mechanism effectively combines the distinctive features acquired by both modules to produce a comprehensive and robust representation of the input data. WaveCoformer outperforms related state-of-the-art networks in publicly available Synapse and Adrenal tumor segmentation datasets, with a mean Dice score of 83.86% and 79%, respectively. The model is feasible for deployment in resource-constrained environments with rapid medical image analysis due to its computationally efficient nature and improved segmentation performance. The code is available at:https://github.com/duttapallabi2907/WaveCoformer.

The computerized delineation and prognosis of lung cancer is typically based on Computed Tomography (CT) image analysis, whereby the region of interest (ROI) is accurately demarcated and classified. Deep learning in computer vision provides a different perspective to image segmentation. Due to the increasing number of cases of lung cancer and the availability of large volumes of CT scans every day, the need for automated handling becomes imperative. This requires efficient delineation and diagnosis through the design of new techniques for improved accuracy. In this article, we introduce the novel Weighted Deformable U-Net (WDU-Net) for efficient delineation of the tumor region. It incorporates the Deformable Convolution (DC) that can model arbitrary geometric shapes of region of interests. This is enhanced by the Weight Generation (WG) module to suppress unimportant features while highlighting relevant ones. A new Focal Asymmetric Similarity (FAS) loss function helps handle class imbalance. Ablation studies and comparison with state-of-the-art models help establish the effectiveness of WDU-Net with ensemble learning, tested on five publicly available lung cancer datasets. Best results were obtained on the LIDC-IDRI lung tumor test dataset, with an average Dice score of 0.9137, the Hausdorff Distance 95% (HD95) of 5.3852, and Area Under the Receiver Operating Characteristic (ROC) Curve (AUC) of 0.9449.

Alzheimer's disease (AD) is the leading neurodegenerative disorder and primary cause of dementia. Researchers are increasingly drawn to automated diagnosis of AD using neuroimaging analyses. Conventional deep learning (DL) models excel in constructing learning classifiers in early-stage AD diagnosis. However, they often struggle with AD diagnosis due to uncertainties stemming from unclear annotations by experts, challenges in data collection, such as data harmonization issues, and limitations in equipment resolution. These factors contribute to imprecise data, hindering accurate analysis, interpretation of obtained results, and understanding of complex symptoms. In response, the integration of fuzzy logic into DL, forming fuzzy deep learning (FDL), effectively manages imprecise data and provides interpretable insights, offering a valuable advancement in AD. Therefore, exploring recent advancements in integrating DL with fuzzy logic is crucial for improving AD diagnosis. In this review, we explore the contributions of fuzzy logic within FDL models, focusing on fuzzy-based image preprocessing, segmentation, and classification. Moreover, in exploring research directions, we discuss the possibility of the fusion of multimodal data with fuzzy logic, addressing challenges in AD diagnosis. Leveraging fuzzy logic and membership while integrating diverse datasets, such as genomics, proteomics, and metabolomics may provide an effective development of a DL classifier. In addition, fuzzy explainable DL promises more accurate and linguistically interpretable decision support systems for AD diagnosis. The primary objective of this article is to serve as a comprehensive and authoritative resource for newcomers, researchers, and clinicians interested in employing FDL models for AD diagnosis.

Given the scarcity of sufficient annotated data, using small sets of labeled samples under semi-supervision in biomedical imaging becomes necessary. Despite being highly successful, deep learning algorithms demand plenty of data to obtain significant performance. Complex data models make the usage of these methods costly. Selecting the correct model and tuning the hyperparameters of a model are also difficult jobs. Hence, a novel approach namely affinity propagation-based semi-supervised segmentation (APSS) is proposed. Here, affinity propagation clustering is modified and integrated with the advanced learning techniques that can efficiently use limited training data by discarding the completely exploited labeled data points. Moreover, a novel affinity calculation method is proposed considering both the Euclidean and geodesic distances to compute the distance between the two points on the histogram. This twofold contribution is tested using the three standard datasets (the International Skin Imaging Collaboration (ISIC) dermoscopic image dataset, the retinal fundus image dataset, and the liver tumor segmentation (LiTS) dataset). Results are compared with the three standard semi-supervised algorithms and four supervised algorithms. The effectiveness of the APSS approach in finding and exploiting the relationship between the labeled and unlabeled datasets is demonstrated in terms of qualitative (subjective evaluation and visual inspection) and quantitative performance (objective evaluation and numerical measurements).

Volumetric medical image segmentation is an indispensable part of accurate diagnosis, treatment planning, and image-guided interventions. It entails the delineation of structures within 3D medical images. However, there are various challenges associated, including uncertain or intersecting boundaries, discrepancies in volume shapes and dimensions, variations among patients, and the need for considerable computational resources. We present here the new Volumetric global-Context integrated Attention Network (VoCANet), for segmenting anatomical structures from multi-dimensional medical images. Global contextual information, from different levels of the low-projection path, is utilized to efficiently capture features corresponding to anatomical structures of interest; which may often exhibit diverse shapes and sizes. An attention module is integrated into the network to enhance the range of activation responses for prioritizing pertinent features, thereby optimizing computational resources. The high-projection path increases the dimensionality of the feature volumes obtained from the low-projection path, to produce the final segmentation output. It incorporates multistage supervision and densely connected convolution kernels, for enhancing segmentation performance. The proposed deep network is applied to the task of multi-organ and Adrenocortical Carcinoma segmentation of the abdominal images from Synapse and Adrenal-ACC-Ki67-Seg datasets respectively. Experimental results demonstrate the superiority of our model, as compared to other state-of-the-art frameworks, in performing segmentation from multi-dimensional medical images.

The problem of efficiently classifying imbalanced data has become one of the most challenging tasks in machine learning. Some real world examples include medical image analysis, fraud detection, fault diagnosis, and anomaly detection. Although several data-level algorithms have been developed to ad dress imbalance, they are typically subject to some restrictions. We propose a novel variant of the k-NN family of classifiers, and name this as Density-based Adaptive-distance kNN (DAkNN). It can effectively handle data with skewed distributions and varying class-densities using the concept of adaptive distance. Comparative superiority is experimentally established over related data-level algorithms (SMOTE, ADASYN), using ten sets of two-class data, in terms of geometric mean (of the true positive and negative rates) and accuracy. Additionally, five sets of multi-class data are considered and compared with different variants of k-NN, which are currently very popular. Finally, DAkNN is successfully applied on the highly imbalanced Lower Grade Glioma (LGG) MR images, with an Average-Dice score of 0.9082 for delineating the tumor regions. The results demonstrate clear superiority over state-of-the-art algorithms.

Integration of various technologies to an Internet of Things (IoT) framework share the common goals of a consistent and structured data format that can be applied to any device, given the vast application scope of IoT. Additional goals include minimizing channel traffic and system energy consumption. In this paper, we propose to dismiss the requirement of certain seemingly crucial identifier fields from packets arriving through various sensor nodes in an agricultural IoT deployment. The proposed approach reduces packet size, thereby reducing channel traffic and energy consumption, as well as retaining the capability of identifying these originating nodes. We propose a method of a blind agricultural IoT node and sensor identification, which can be sourced and operated from a master node as well as a remote server. Additionally, this scheme has the capability of detecting the radio link quality between the master and slave nodes in a rudimentary form, as well as identifying the sensor nodes. We successfully trained and tested various multilayer perceptron-based models for blind identification, in real-time, using our implemented agricultural IoT implementation. The effect of changes in learning rate and momentum of the optimizer on the accuracy of classification is also studied. The projected cumulative energy savings across the network architecture, of our scheme, in conjunction with TCP/IP header compression techniques, are substantial. For a 100 node deployment using a combination of the proposed blind identification reduced sampling strategies over regular IPv4-based TCP/IP connection, an estimated annual saving of ≈99% is projected.

Brain tumors constitute one of the deadliest forms of cancers, with a high mortality rate. Of these, Glioblastoma multiforme (GBM) remains the most common and lethal primary brain tumor in adults. Tumor biopsy being challenging for brain tumor patients, noninvasive techniques like imaging play an important role in the process of brain cancer detection, diagnosis and prognosis; particularly using Magnetic Resonance Imaging (MRI). Therefore, development of advanced extraction and selection strategies of quantitative MRI features become necessary for noninvasively predicting and grading the tumors. In this paper we extract 56 three-dimensional quantitative MRI features, related to tumor image intensities, shape and texture, from 254 brain tumor patients. An adaptive neuro-fuzzy classifier based on linguistic hedges (ANFC-LH) is developed to simultaneously select significant features and predict the tumor grade. ANFC-LH achieves a significantly higher testing accuracy (85.83%) as compared to existing standard classifiers.

A novel similarity-based feature selection algorithm is developed, using the concept of distance correlation. A feature subset is selected in terms of this similarity measure between pairs of features, without assuming any underlying distribution of the data. The pair-wise similarity is then employed, in a message passing framework, to select a set of exemplars features involving minimum redundancy and reduced parameter tuning. The algorithm does not need an exhaustive traversal of the search space. The methodology is next extended to handle large data, using an inherent property of distance correlation. The effectiveness of the algorithm is demonstrated on nine sets of publicly-available data.

Radiographic-imaging modalities like computerized tomography, positron emission tomography, and magnetic resonance imaging are playing a major role in the diagnosis and prognosis of cancer. Gene and protein expression patterns, from the tumor genome, are seen to facilitate individualized selection of therapies. Along with breakthroughs in biotechnology, applicable within cancer radiation biology, a new research field called Radiogenomics has been born in radiation oncology. Associating genotypes with imaging phenotypes holds promise for personalized optimal treatment. Segmentation and feature selection from the region of interest in an image are followed by correlation with the gene expression profile of the tumor in order to determine its noninvasive surrogates. This paper highlights the roles of quantitative imaging, genomics, and radiogenomics for a patient-specific tumor management.

Knee osteoarthritis (OA) is a debilitating health condition affecting elderly population. Early diagnosis of the disease noninvasively will be very useful, and requires novel image analysis techniques to be developed for processing radiological and MRI scans of knee joints. In this paper, we propose novel micro texture based feature descriptors for modeling subtle variations in MRI T2 maps, so as to facilitate early detection of osteoarthritis. The experimental evaluation of the proposed micro texture descriptors on a publicly available OAI database show, that the proposed features allow a significant improvement in discriminating the textures for MRI T2 maps, corresponding to normal subjects as compared to the subjects with risk factors for developing osteoarthritis.

In this paper, we propose a Three Ensemble neural network rule extraction algorithm. Then we investigate Hayashi's first question, “Can the Ensemble-Recursive-Rule eXtraction (E-Re-RX) algorithm be extended to an ensemble neural network consisting of three or more MLPs and extract comprehensible rules?” The E-Re-RX algorithm is an effective rule extraction algorithm for dealing with data sets that mix discrete and continuous attributes. Using the experimental results, we consider the three MLP ensemble Re-RX algorithm from various points of view. Finally, we present provisional positive conclusions.

Gene expression data being high-dimensional and redundant, dimensionality reduction is of prime concern. We employ the algorithm Fuzzy Clustering Large Applications based on RAN-domized Search (FCLARANS) for attribute clustering and dimensionality reduction based on the study of gene ontology and differential gene expressions. The use of domain knowledge helps in the automated selection of biologically meaningful partitions. The use of Gene Ontology (GO) study helps in detecting biologically enriched and statistically significant clusters. Fold-change is measured to select the differentially expressed genes as the representatives of these clusters. Tools like Eisen plot and cluster profiles of these clusters help establish their coherence. Important representative features (or genes) are extracted from each enriched gene partitions to form the reduced gene space. While the reduced gene set forms a biologically meaningful gene space it simultaneously leads to a decrease in computational burden. External validation of the reduced subspace, using various well-known classifiers, establishes the effectiveness of the proposed methodology.

In this paper, a novel feature selection algorithm, which is governed by biological knowledge, is developed. Gene expression data being high dimensional and redundant, dimensionality reduction is of prime concern. We employ the algorithm clustering large applications based on RAN-domized search (CLARANS) for attribute clustering and dimensionality reduction based on gene ontology (GO) study. Feature selection with unsupervised learning is a difficult problem, with neither class labels present nor any guidance available to the search. Determination of the optimal number of clusters is another major issue, and has an impact on the resulting output. The use of GO analysis helps in the automated selection of biologically meaningful partitions. Tools such as Eisen plot and cluster profiles of these clusters help establish their coherence. Important representative features (or genes) are extracted from each correlated set of genes in such partitions. The algorithm is implemented on high-dimensional Yeast cell-cycle, Human Multiple Tissues, and Leukemia microarray data. In the second pass, clustering on the reduced gene space validates preservation of the inherent behavior of the original high-dimensional expression profiles. While the reduced gene set forms a biologically meaningful gene space, it simultaneously leads to a decrease in computational burden. External validation of the reduced subspace, using various well-known classifiers, establishes the effectiveness of the proposed methodology.

The analysis of gene regulatory networks provides enormous information on various fundamental cellular processes involving growth, development, hormone secretion, and cellular communication. Their extraction from available gene expression profiles is a challenging problem. Such reverse engineering of genetic networks offers insight into cellular activity toward prediction of adverse effects of new drugs or possible identification of new drug targets. Tasks such as classification, clustering, and feature selection enable efficient mining of knowledge about gene interactions in the form of networks. It is known that biological data is prone to different kinds of noise and ambiguity. Soft computing tools, such as fuzzy sets, evolutionary strategies, and neurocomputing, have been found to be helpful in providing low-cost, acceptable solutions in the presence of various types of uncertainties. In this paper, we survey the role of these soft methodologies and their hybridizations, for the purpose of generating genetic networks.

An evolutionary rough feature selection algorithm is used for classifying microarray gene expression patterns. Since the data typically consist of a large number of redundant features, an initial redundancy reduction of the attributes is done to enable faster convergence. Rough set theory is employed to generate reducts, which represent the minimal sets of nonredundant features capable of discerning between all objects, in a multiobjective framework. The effectiveness of the algorithm is demonstrated on three cancer datasets.

Gesture recognition pertains to recognizing meaningful expressions of motion by a human, involving the hands, arms, face, head, and/or body. It is of utmost importance in designing an intelligent and efficient human-computer interface. The applications of gesture recognition are manifold, ranging from sign language through medical rehabilitation to virtual reality. In this paper, we provide a survey on gesture recognition with particular emphasis on hand gestures and facial expressions. Applications involving hidden Markov models, particle filtering and condensation, finite-state machines, optical flow, skin color, and connectionist models are discussed in detail. Existing challenges and future research possibilities are also highlighted

Sonic detection and ranging (SODAR) systems are efficient and economical tool to probe the lower planetary boundary layer on a continuous basis. The lower atmospheric patterns (each depicting a different atmospheric condition) recorded by this system can prove to be extremely useful if classified and interpreted correctly. The manual identification of these SODAR patterns is a laborious task and requires considerable expertise. A connectionist system has already been developed by the authors to automate the process to some extent. In this letter, we enhance its generalization of performance, by incorporating feature extraction using the fast Fourier transform. The results are compared with that in earlier work to demonstrate its effectiveness.

Biclustering or simultaneous clustering of both genes and conditions have generated considerable interest over the past few decades, particularly related to the analysis of high-dimensional gene expression data in information retrieval, knowledge discovery, and data mining. The objective is to find sub-matrices, i.e., maximal subgroups of genes and subgroups of conditions where the genes exhibit highly correlated activities over a range of conditions. Since these two objectives are mutually conflicting, they become suitable candidates for multi-objective modeling. In this study, a novel multi-objective evolutionary biclustering framework is introduced by incorporating local search strategies. A new quantitative measure to evaluate the goodness of the biclusters is developed. The experimental results on benchmark datasets demonstrate better performance as compared to existing algorithms available in literature.

In this study, we introduce a novel clustering architecture, in which several subsets of patterns can be processed together with an objective of finding a common structure. The structure revealed at the global level is determined by exchanging prototypes of the subsets of data and by moving prototypes of the corresponding clusters toward each other. Thereby, the required communication links are established at the level of cluster prototypes and partition matrices, without hampering the security concerns. A detailed clustering algorithm is developed by integrating the advantages of both fuzzy sets and rough sets, and a measure of quantitative analysis of the experimental results is provided for synthetic and real-world data.

Soft computing is gradually opening up several possibilities in bioinformatics, especially by generating low-cost, low-precision (approximate), good solutions. In this paper, we survey the role of different soft computing paradigms, like fuzzy sets (FSs), artificial neural networks (ANNs), evolutionary computation, rough sets (RSes), and support vector machines (SVMs), in this direction. The major pattern-recognition and data-mining tasks considered here are clustering, classification, feature selection, and rule generation. Genomic sequence, protein structure, gene expression microarrays, and gene regulatory networks are some of the application areas described. Since the work entails processing huge amounts of incomplete or ambiguous biological data, we can utilize the learning ability of neural networks for adapting, uncertainty handling capacity of FSs and RSes for modeling ambiguity, searching potential of genetic algorithms for efficiently traversing large search spaces, and the generalization capability of SVMs for minimizing errors

SODAR (or acoustic radar) systems are a useful tool to efficiently probe the lower planetary boundary layer (LPBL). The observations obtained by these systems can prove to be extremely useful if classified and interpreted correctly. The manual identification of different types of SODAR-recorded lower atmospheric microstructures is a laborious task and can be performed only by an expert having wide experience with the system and the variety of observations recorded by it. In this letter, we have developed a connectionist system to classify or identify SODAR patterns. The results demonstrate that the multilayer perceptron-based model is capable of successfully identifying the different SODAR patterns.

A methodology is described for evolving a Rough-fuzzy multi layer perceptron with modular concept using a genetic algorithm to obtain a structured network suitable for both classification and rule extraction. The modular concept, based on "divide and conquer" strategy, provides accelerated training and a compact network suitable for generating a minimum number of rules with high certainty values. The concept of variable mutation operator is introduced for preserving the localized structure of the constituting knowledge-based subnetworks, while they are integrated and evolved. Rough set dependency rules are generated directly from the real valued attribute table containing fuzzy membership values. Two new indices viz., "certainty" and "confusion" in a decision are defined for evaluating quantitatively the quality of rules. The effectiveness of the model and the rule extraction algorithm is extensively demonstrated through experiments alongwith comparisons.

The present article provides a survey of the available literature on data mining using soft computing. A categorization has been provided based on the different soft computing tools and their hybridizations used, the data mining function implemented, and the preference criterion selected by the model. The utility of the different soft computing methodologies is highlighted. Generally fuzzy sets are suitable for handling the issues related to understandability of patterns, incomplete/noisy data, mixed media information and human interaction, and can provide approximate solutions faster. Neural networks are nonparametric, robust, and exhibit good learning and generalization capabilities in data-rich environments. Genetic algorithms provide efficient search algorithms to select a model, from mixed media data, based on some preference criterion/objective function. Rough sets are suitable for handling different types of uncertainty in data. Some challenges to data mining and the application of soft computing methodologies are indicated. An extensive bibliography is also included.

A fuzzy knowledge-based network is developed based on the linguistic rules extracted from a fuzzy decision tree. A scheme for automatic linguistic discretization of continuous attributes, based on quantiles, is formulated. A novel concept for measuring the goodness of a decision tree, in terms of its compactness (size) and efficient performance, is introduced. Linguistic rules are quantitatively evaluated using new indices. The rules are mapped to a fuzzy knowledge-based network, incorporating the frequency of samples and depth of the attributes in the decision tree. New fuzziness measures, in terms of class memberships, are used at the node level of the tree to take care of overlapping classes. The effectiveness of the system, in terms of recognition scores, structure of decision tree, performance of rules, and network size, is extensively demonstrated on three sets of real-life data.

The present article is a novel attempt in providing an exhaustive survey of neuro-fuzzy rule generation algorithms. Rule generation from artificial neural networks is gaining in popularity in recent times due to its capability of providing some insight to the user about the symbolic knowledge embedded within the network. Fuzzy sets are an aid in providing this information in a more human comprehensible or natural form, and can handle uncertainties at various levels. The neuro-fuzzy approach, symbiotically combining the merits of connectionist and fuzzy approaches, constitutes a key component of soft computing at this stage. To date, there has been no detailed and integrated categorization of the various neuro-fuzzy models used for rule generation. We propose to bring these together under a unified soft computing framework. Moreover, we include both rule extraction and rule refinement in the broader perspective of rule generation. Rules learned and generated for fuzzy reasoning and fuzzy control are also considered from this wider viewpoint. Models are grouped on the basis of their level of neuro-fuzzy synthesis. Use of other soft computing tools like genetic algorithms and rough sets are emphasized. Rule generation from fuzzy knowledge-based networks, which initially encode some crude domain knowledge, are found to result in more refined rules. Finally, real-life application to medical diagnosis is provided.

A new scheme of knowledge encoding in a fuzzy multilayer perceptron (MLP) using rough set-theoretic concepts is described. Crude domain knowledge is extracted from the data set in the form of rules. The syntax of these rules automatically determines the appropriate number of hidden nodes while the dependency factors are used in the initial weight encoding. The network is then refined during training. Results on classification of speech and synthetic data demonstrate the superiority of the system over the fuzzy and conventional versions of the MLP (involving no initial knowledge).

A new scheme of knowledge-based classification and rule generation using a fuzzy multilayer perceptron (MLP) is proposed. Knowledge collected from a data set is initially encoded among the connection weights in terms of class a priori probabilities. This encoding also includes incorporation of hidden nodes corresponding to both the pattern classes and their complementary regions. The network architecture, in terms of both links and nodes, is then refined during training. Node growing and link pruning are also resorted to. Rules are generated from the trained network using the input, output, and connection weights in order to justify any decision(s) reached. Negative rules corresponding to a pattern not belonging to a class can also be obtained. These are useful for inferencing in ambiguous cases. Results on real life and synthetic data demonstrate that the speed of learning and classification performance of the proposed scheme are better than that obtained with the fuzzy and conventional versions of the MLP (involving no initial knowledge encoding). Both convex and concave decision regions are considered in the process.

A connectionist inferencing network, based on the fuzzy version of Kohonen's model already developed by the authors, is proposed. It is capable of handling uncertainty and/or impreciseness in the input representation provided in quantitative, linguistic and/or set forms. The output class membership value of an input pattern is inferred by the trained network. A measure of certainty expressing confidence m the decision is also defined. The model is capable of querying the user for the more important input feature information, if required, in case of partial inputs. Justification for an inferred decision may be produced in rule form, when so desired by the user. The connection weight magnitude of the trained neural network are utilized in every stage of the proposed inferencing procedure. The antecedent and consequent parts of the justificatory rules are provided in natural forms. The effectiveness of the algorithm is tested on the vowel recognition problem and on two sets of artificially generated nonconvex pattern classes.

A connectionist expert system model, based on a fuzzy version of the multilayer perceptron developed by the authors, is proposed. It infers the output class membership value(s) of an input pattern and also generates a measure of certainty expressing confidence in the decision. The model is capable of querying the user for the more important input feature information, if and when required, in case of partial inputs. Justification for an inferred decision may be produced in rule form, when so desired by the user. The magnitudes of the connection weights of the trained neural network are utilized in every stage of the proposed inferencing procedure. The antecedent and consequent parts of the justificatory rules are provided in natural forms. The effectiveness of the algorithm is tested on the speech recognition problem, on some medical data and on artificially generated intractable (linearly nonseparable) pattern classes.

A two-level pattern classification scheme is considered. At its first level the scheme labels the input object as "doubtful" or "certain" and at the second one applies the respective classification rule. Complicating the classifier in such a way the authors aim at a more accurate result than that obtained through either of the classification rules itself. A fuzzy neural network with linguistically interpretable inputs has been applied to detect the boundaries of the "doubtful" region(s) in the feature space. A fuzzy k-nearest neighbors rule with k=1 and k=5 has been used at the second level for the "doubtful" and "certain" regions, respectively. The idea has been demonstrated on a generated data set (two separable classes with uniform distribution). The results show the tendency of improvement of the classification accuracy.

This paper describes a self-organizing artificial neural network, based on Kohonen's model of self-organization, which is capable of handling fuzzy input and of providing fuzzy classification. Unlike conventional neural net models, this algorithm incorporates fuzzy set-theoretic concepts at various stages. The input vector consists of membership values for linguistic properties along with some contextual class membership information which is used during self-organization to permit efficient modeling of fuzzy (ambiguous) patterns. A new definition of gain factor for weight updating is proposed. An index of disorder involving mean square distance between the input and weight vectors is used to determine a measure of the ordering of the output space. This controls the number of sweeps required in the process. Incorporation of the concept of fuzzy partitioning allows natural self-organization of the input data, especially when they have ill-defined boundaries. The output of unknown test patterns is generated in terms of class membership values. Incorporation of fuzziness in input and output is seen to provide better performance as compared to the original Kohonen model and the hard version. The effectiveness of this algorithm is demonstrated on the speech recognition problem for various network array sizes, training sets and gain factors.>

A fuzzy neural network model based on the multilayer perceptron, using the backpropagation algorithm, and capable of fuzzy classification of patterns is described. The input vector consists of membership values to linguistic properties while the output vector is defined in terms of fuzzy class membership values. This allows efficient modeling of fuzzy uncertain patterns with appropriate weights being assigned to the backpropagated errors depending upon the membership values at the corresponding outputs. During training, the learning rate is gradually decreased in discrete steps until the network converges to a minimum error solution. The effectiveness of the algorithm is demonstrated on a speech recognition problem. The results are compared with those of the conventional MLP, the Bayes classifier, and other related models.