Saroj K. Meher

A novel approach to neuro-fuzzy classification

A new model for neuro-fuzzy (NF) classification systems is proposed. The motivation is to utilize the feature-wise degree of belonging of patterns to all classes that are obtained through a fuzzification process. A fuzzification process generates a membership matrix having total number of elements equal to the product of the number of features and classes present in the data set. These matrix elements are the input to neural networks. The effectiveness of the proposed model is established with four benchmark data sets (completely labeled) and two remote sensing images (partially labeled). Different performance measures such as misclassification, classification accuracy and kappa index of agreement for completely labeled data sets, and beta index of homogeneity and Davies-Bouldin (DB) index of compactness for remotely sensed images are used for quantitative analysis of results. All these measures supported the superiority of the proposed NF classification model. The proposed model learns well even with a lower percentage of training data that makes the system fast.

Wavelet-Feature-Based Classifiers for Multispectral Remote-Sensing Images

The objective of this paper is to utilize the extracted features obtained by the wavelet transform (WT) rather than the original multispectral features of remote-sensing images for land-cover classification. WT provides the spatial and spectral characteristics of a pixel along with its neighbors, and hence, this can be utilized for an improved classification. Four classifiers, namely, the fuzzy product aggregation reasoning rule (FPARR), fuzzy explicit, multilayered perceptron, and neuro-fuzzy (NF), are used for this purpose. The performance is tested on multispectral real and synthetic images. The performance of original and wavelet-feature (WF)-based methods is compared. The WF-based methods have consistently yielded better results. Biorthogonal3.3 (Bior3.3) wavelet is found to be superior to other wavelets. FPARR along with the Bior3.3 wavelet outperformed all other methods. Results are evaluated using quantitative indexes like beta and Xie-Beni

Data Compression of Power Quality Events Using the Slantlet Transform

The slantlet transform (SLT) is an orthogonal discrete wavelet transform (DWT) with two zero moments and with improved time localization. It also retains the basic characteristic of the usual filterbank such as octave band characteristic, a scale dilation factor of two and efficient implementation. However, the SLT is based on the principle of designing different filters for different scales unlike iterated filterbank approaches for the DWT. In this paper a novel approach for power quality data compression using the SLT is presented and its performance in terms of compression ratio (CR), percentage of energy retained and mean square error present in the reconstructed signals is assessed. Varieties of power quality events, which include voltage sag, swell, momentary interruption, harmonics, transient oscillation, and voltage flicker are used to test the performance of the new approach. Computer simulation results indicate that the SLT offers superior compression performance compared to the conventional DCT and the DWT based approaches.

A novel fuzzy classifier based on product aggregation operator

The present article proposes a fuzzy set-based classifier with a better learning and generalization capability. The proposed classifier exploits the feature-wise degree of belonging of a pattern to all classes, generalization in the fuzzification process and the combined class-wise contribution of features effectively. The classifier uses a @p-type membership function and product aggregation reasoning rule (operator). Its effectiveness is verified with two conventional (completely labeled) data sets and two remote sensing images (partially labeled data sets). The proposed classifier is compared with similar fuzzy methods. Different performance measures are used for quantitative evaluation of the proposed classifier.

Customer churn time prediction in mobile telecommunication industry using ordinal regression

Customer churn in considered to be a core issue in telecommunication customer relationship management (CRM). Accurate prediction of churn time or customer tenure is important for developing appropriate retention strategies. In this paper, we discuss a method based on ordinal regression to predict churn time or tenure of mobile telecommunication customers. Customer tenure is treated as an ordinal outcome variable and ordinal regression is used for tenure modeling. We compare ordinal regression with the state-of-the-art methods for tenure prediction - survival analysis. We notice from our results that ordinal regression could be an alternative technique for survival analysis for churn time prediction of mobile customers. To the best knowledge of authors, the use of ordinal regression as a potential technique for modeling customer tenure has been attempted for the first time.

Wavelet-fuzzy hybridization: Feature-extraction and land-cover classification of remote sensing images

Abstract: A wavelet feature based supervised scheme for fuzzy classification of land covers in multispectral remote sensing images is proposed. The proposed scheme is developed in the framework of wavelet-fuzzy hybridization, a soft computing approach. The wavelet features obtained from wavelet transform on an image provides spatial and spectral characteristics (i.e., texture information) of pixels and hence can be utilized effectively for improving accuracy in classification, instead of using original spectral features. Four different fuzzy classifiers are considered for this purpose and evaluated using different wavelet features. Wavelet feature based fuzzy classifiers produced consistently better results compared to original spectral feature based methods on various images used in the present investigation. Further, the performance of the Biorthogonal3.3 (Bior3.3) wavelet is observed to be superior to other wavelets. This wavelet in combination with fuzzy product aggregation reasoning-rule outperformed all other methods. Potentiality of the proposed soft computing approach in isolating various land covers are evaluated both visually and quantitatively using indexes like @b measure of homogeneity and Xie-Beni measure of compactness and separability.

Rough-wavelet granular space and classification of multispectral remote sensing image

Abstract: A new rough-wavelet granular space based model for land cover classification of multispectral remote sensing image, is described in the present article. In this model, we propose the formulation of class-dependent (CD) granules in wavelet domain using shift-invariant wavelet transform (WT). Shift-invariant WT is carried out with properly selected wavelet base and decomposition level(s). The transform is used to characterize the feature-wise belonging of granules to different classes, thereby producing wavelet granulation of the feature space. The wavelet granules thus generated possess better class discriminatory information. The granulated feature space not only analyzes the contextual information in time or frequency domain individually, but also looks into the combined time-frequency domain. These characteristics of the generated CD wavelet granules are very useful in the pattern classification with overlapping classes. Neighborhood rough sets (NRS) are employed in the selection of a subset of granulated features that further explore the local/contextual information from neighbor granules. The model thus explores mutually the advantages of shift-invariant wavelet granulation and NRS. The superiority of the proposed model to other similar methods is established both visually and quantitatively for land cover classification of multispectral remote sensing images. With experimental results, it is found that the proposed model is superior with biorthogonal3.3 wavelet, and when integrated with NRS, it performs the best.

Neuro-fuzzy-combiner: an effective multiple classifier system

A neuro-fuzzy-combiner (NFC) is proposed to design an efficient multiple classifier system (MCS) with an aim to have an effective solution scheme for difficult classification problems. Although, a number of combiners exist in the literature, they do not provide consistently good performance on different datasets. In this scenario: 1) we propose an effective multiple classifier system (MCS) based on NFC that fuses the output of a set of fuzzy classifiers; 2) conduct an extensive experimental analysis to justify the effectiveness of the proposed NFC. In the proposed technique, we used a neural network to combine the output of a set of fuzzy classifiers using the principles of neuro-fuzzy hybridisation. The neural combiner adaptively learns its parameters depending on the input data, and thus the output is robust. Superiority of the proposed combiner has been demonstrated experimentally on five standard datasets and two remote sensing images. It performed consistently better than the existing combiners over all the considered datasets.

Explicit rough-fuzzy pattern classification model

An explicit rough-fuzzy model for pattern classification is proposed in the present paper. The model explores and provides the synergistic integration of the merits of both fuzzy and rough sets. It acquires improved learning and generalization capabilities through explicit fuzzification of input features. Likely optimal features are selected from these fuzzified features using neighborhood rough sets, which utilize the neighborhood relative information. The combined class belonging information of features in the designing process of model further enhances its decision-making ability. The resultant features thus provide comprehensive framework for building discriminative pattern classification models for the data sets with highly overlapping class boundaries. The efficacy of the proposed model is verified with four completely labeled data sets including one synthetic remote sensing image, and one partially labeled real remote sensing image. Various performance measurement indexes supported the superiority claim of the model.

Title Natural computing: A problem solving paradigm with granular information processing

Natural computing, inspired by biological course of action, is an interdisciplinary field that formalizes processes observed in living organisms to design computational methods for solving complex problems, or designing artificial systems with more natural behaviour. Based on the tasks abstracted from natural phenomena, such as brain modelling, self-organization, self-repetition, self evaluation, Darwinian survival, granulation and perception, nature serves as a source of inspiration for the development of computational tools or systems that are used for solving complex problems. Nature inspired main computing paradigms used for such development include artificial neural networks, fuzzy logic, rough sets, evolutionary algorithms, fractal geometry, DNA computing, artificial life and granular or perception-based computing. Information granulation in granular computing is an inherent characteristic of human thinking and reasoning process performed in everyday life. The present article provides an overview of the significance of natural computing with respect to the granulation-based information processing models, such as neural networks, fuzzy sets and rough sets, and their hybridization. We emphasize on the biological motivation, design principles, application areas, open research problems and challenging issues of these models.