In this paper, a model-independent sensitivity analysisfor (deep) neural network, Bilateral Sensi... more In this paper, a model-independent sensitivity analysisfor (deep) neural network, Bilateral Sensitivity Analysis (BiSA), is proposed to measure the relationship between neurons and layers. Both the BiSA between pair of layers and the BiSA between any pair neurons in different layers are defined for (deep) neural networks. This sensitivity can measure the influence or contribution from any layer to another layer behind this layer in the (deep) neural networks. It provides a helpful tool to interpret the learned model. The BiSA can also measure the influence or contribution from any neuron to another neuron in a subsequent layer and is critical to analyze the relationship between neurons in different layers. Then the BiSA from any input to any output of a network is easily defined to assess the connections between the inputs and outputs. The proposed BiSA of (deep) neural networks is then applied to characterize the well connectivity in reservoir engineering. Given a network trained b...
2016 International Conference on Fuzzy Theory and Its Applications (iFuzzy), 2016
This paper presents a novel oversampling technique that addresses highly imbalanced data distribu... more This paper presents a novel oversampling technique that addresses highly imbalanced data distribution. At present, the imbalanced data that have anomalous class distribution and underrepresented data are difficult to deal with through a variety of conventional machine learning technologies. In order to balance class distributions, an adaptive subspace self-organizing map (ASSOM) that combines the local mapping scheme and globally competitive rule is proposed to artificially generate synthetic samples focusing on minority class samples. The ASSOM is conformed with feature-invariant characteristics, including translation, scaling and rotation, and it retains the independence of basis vectors in each module. Specifically, basis vectors generated via each ASSOM module can avoid generating repeated representative features that offer nothing but heavy computational load. Several experimental results demonstrate that the proposed ASSOM method with supervised learning manner is superior to ...
2020 International Conference on Fuzzy Theory and Its Applications (iFUZZY), 2020
Kohonen’s Adaptive Subspace Self-Organizing Map (ASSOM) learns several subspaces of the data wher... more Kohonen’s Adaptive Subspace Self-Organizing Map (ASSOM) learns several subspaces of the data where each subspace represents some invariant characteristics of the data. To deal with the imbalance classification problem, earlier we have proposed a method for oversampling the minority class using Kohonen’s ASSOM. This investigation extends that study, clarifies some issues related to our earlier work, provides the algorithm for generation of the oversamples, applies the method on several benchmark data sets, and makes an application to a Brain Computer Interface (BCI) problem. First we compare the performance of our method using some benchmark data sets with several state-of-the-art methods. Finally, we apply the ASSOM-based technique to analyze a BCI based application using electroencephalogram (EEG) datasets. Our results demonstrate the effectiveness of the ASSOM-based method in dealing with imbalance classification problem.
We present a three stage hierarchical self-organized genetic-algorithm based rule generation (SOG... more We present a three stage hierarchical self-organized genetic-algorithm based rule generation (SOGARG) method for fuzzy controllers. The first stage selects rules required to control the system in the vicinity of the set point. The second stage extends the rule base to span the entire input space. The third stage then refines the rule-base. The first two stages use the same
Neural, Parallel & Scientific Computations archive, Mar 1, 2007
Hydrophobicity is probably the most important property of amino acid that is being often exploite... more Hydrophobicity is probably the most important property of amino acid that is being often exploited to predict protein folds. There are many amino acid hydrophobicity scales available in the literature. Here we propose two computational approaches based on self-organizing map ...
Abstract. Most methods of classification either ignore feature analysis or do it in a separate ph... more Abstract. Most methods of classification either ignore feature analysis or do it in a separate phase, offline prior to the main classification task. This paper proposes a novel neuro-fuzzy scheme for classification with online feature selection. It is a four-layered feed-forward network for ...
International Journal of Intelligent Systems, 2006
Many attempts have been made to analyze gene expression data. Typical goals of such analysis incl... more Many attempts have been made to analyze gene expression data. Typical goals of such analysis include discovery of subclasses, designing predictors/classifiers for diseases, identifying marker genes, and trying to get a deeper understanding of underlying biological process. ...
Abstract Many attempts have been made to analyze gene expression data. Typical goals of such anal... more Abstract Many attempts have been made to analyze gene expression data. Typical goals of such analysis include discovery of subclasses, designing predictors/classifiers for diseases, identifying marker genes, and trying to get a deeper understanding of underlying ...
In this article we present a self-tuning PD-type fuzzy logic controller (FLC). The output scaling... more In this article we present a self-tuning PD-type fuzzy logic controller (FLC). The output scaling factor (SF) of the proposed controller is modified on-line by an updating factor (α) according to the process trend. The value of a is determined from a rule-base defined on error (e) and change of error (Δe) of the controlled variable. The proposed self-tuning FLC is designed using a very simple control rule-base and the most natural and unbiased membership functions (MFs) (symmetric triangles with equal base and 50% overlap with neighbouring MFs). The effectiveness of the scheme is established through simulation experiments on various types of second order processes such as (i) marginally stable, (ii) non-linear and (iii) non-minimum phase-pole (unstable) systems with different values of dead time. Performance of the proposed self-tuning FLC is compared with that of its conventional counterpart with respect to both step set-point change and load disturbance using several performance indices. In each case th...
In this paper, a model-independent sensitivity analysisfor (deep) neural network, Bilateral Sensi... more In this paper, a model-independent sensitivity analysisfor (deep) neural network, Bilateral Sensitivity Analysis (BiSA), is proposed to measure the relationship between neurons and layers. Both the BiSA between pair of layers and the BiSA between any pair neurons in different layers are defined for (deep) neural networks. This sensitivity can measure the influence or contribution from any layer to another layer behind this layer in the (deep) neural networks. It provides a helpful tool to interpret the learned model. The BiSA can also measure the influence or contribution from any neuron to another neuron in a subsequent layer and is critical to analyze the relationship between neurons in different layers. Then the BiSA from any input to any output of a network is easily defined to assess the connections between the inputs and outputs. The proposed BiSA of (deep) neural networks is then applied to characterize the well connectivity in reservoir engineering. Given a network trained b...
2016 International Conference on Fuzzy Theory and Its Applications (iFuzzy), 2016
This paper presents a novel oversampling technique that addresses highly imbalanced data distribu... more This paper presents a novel oversampling technique that addresses highly imbalanced data distribution. At present, the imbalanced data that have anomalous class distribution and underrepresented data are difficult to deal with through a variety of conventional machine learning technologies. In order to balance class distributions, an adaptive subspace self-organizing map (ASSOM) that combines the local mapping scheme and globally competitive rule is proposed to artificially generate synthetic samples focusing on minority class samples. The ASSOM is conformed with feature-invariant characteristics, including translation, scaling and rotation, and it retains the independence of basis vectors in each module. Specifically, basis vectors generated via each ASSOM module can avoid generating repeated representative features that offer nothing but heavy computational load. Several experimental results demonstrate that the proposed ASSOM method with supervised learning manner is superior to ...
2020 International Conference on Fuzzy Theory and Its Applications (iFUZZY), 2020
Kohonen’s Adaptive Subspace Self-Organizing Map (ASSOM) learns several subspaces of the data wher... more Kohonen’s Adaptive Subspace Self-Organizing Map (ASSOM) learns several subspaces of the data where each subspace represents some invariant characteristics of the data. To deal with the imbalance classification problem, earlier we have proposed a method for oversampling the minority class using Kohonen’s ASSOM. This investigation extends that study, clarifies some issues related to our earlier work, provides the algorithm for generation of the oversamples, applies the method on several benchmark data sets, and makes an application to a Brain Computer Interface (BCI) problem. First we compare the performance of our method using some benchmark data sets with several state-of-the-art methods. Finally, we apply the ASSOM-based technique to analyze a BCI based application using electroencephalogram (EEG) datasets. Our results demonstrate the effectiveness of the ASSOM-based method in dealing with imbalance classification problem.
We present a three stage hierarchical self-organized genetic-algorithm based rule generation (SOG... more We present a three stage hierarchical self-organized genetic-algorithm based rule generation (SOGARG) method for fuzzy controllers. The first stage selects rules required to control the system in the vicinity of the set point. The second stage extends the rule base to span the entire input space. The third stage then refines the rule-base. The first two stages use the same
Neural, Parallel & Scientific Computations archive, Mar 1, 2007
Hydrophobicity is probably the most important property of amino acid that is being often exploite... more Hydrophobicity is probably the most important property of amino acid that is being often exploited to predict protein folds. There are many amino acid hydrophobicity scales available in the literature. Here we propose two computational approaches based on self-organizing map ...
Abstract. Most methods of classification either ignore feature analysis or do it in a separate ph... more Abstract. Most methods of classification either ignore feature analysis or do it in a separate phase, offline prior to the main classification task. This paper proposes a novel neuro-fuzzy scheme for classification with online feature selection. It is a four-layered feed-forward network for ...
International Journal of Intelligent Systems, 2006
Many attempts have been made to analyze gene expression data. Typical goals of such analysis incl... more Many attempts have been made to analyze gene expression data. Typical goals of such analysis include discovery of subclasses, designing predictors/classifiers for diseases, identifying marker genes, and trying to get a deeper understanding of underlying biological process. ...
Abstract Many attempts have been made to analyze gene expression data. Typical goals of such anal... more Abstract Many attempts have been made to analyze gene expression data. Typical goals of such analysis include discovery of subclasses, designing predictors/classifiers for diseases, identifying marker genes, and trying to get a deeper understanding of underlying ...
In this article we present a self-tuning PD-type fuzzy logic controller (FLC). The output scaling... more In this article we present a self-tuning PD-type fuzzy logic controller (FLC). The output scaling factor (SF) of the proposed controller is modified on-line by an updating factor (α) according to the process trend. The value of a is determined from a rule-base defined on error (e) and change of error (Δe) of the controlled variable. The proposed self-tuning FLC is designed using a very simple control rule-base and the most natural and unbiased membership functions (MFs) (symmetric triangles with equal base and 50% overlap with neighbouring MFs). The effectiveness of the scheme is established through simulation experiments on various types of second order processes such as (i) marginally stable, (ii) non-linear and (iii) non-minimum phase-pole (unstable) systems with different values of dead time. Performance of the proposed self-tuning FLC is compared with that of its conventional counterpart with respect to both step set-point change and load disturbance using several performance indices. In each case th...
Uploads
Papers by Nikhil R. Pal