Comorbid disease association refers to the simultaneous occurrence of a disease with the coexistence of another primary disease. Due to the complex traits of these co-occurring multi-diseases, it is crucial to know the underlying genetic molecular basis of the prevalent diseases. The inference of common genetic association based on gene co-expression data helps to unveil the pathogenesis of comorbid diseases. There exist a few disease-specific gene co-expression-based analyses to predict the hub genes causing these diseases. However, works lack multi-relational biological data integration. In addition, there still does not exist any unified method to predict the common genetic associations from the co-expression graph across comorbid diseases. Hence, we introduce a generalized and novel approach to predict overlapping genetic associations from disease-specific gene co-expression networks with a self-supervised edge-masking technique catapult with a hypergraph-based pre-embedding learning approach. The advantage of hypergraph learning is that it induces higher-order rich biological information of candidate genes. In addition, we use the self-supervised-based edge masking strategy to attain model training over only a few numbers of edge labels. Our proposed approach outperforms the six baseline models for our case-study datasets and also predicts novel genetic associations across comorbid disease pairs.

Deepfakes are a major security risk for biometric authentication. This technology creates realistic fake videos that can impersonate real people, fooling systems that rely on facial features and voice patterns for identification. Existing multimodal deepfake detectors rely on conventional fusion methods, such as majority rule and ensemble voting, which often struggle to adapt to changing data characteristics and complex patterns. In this paper, we introduce the Straight-through Gumbel-Softmax (STGS) framework, offering a comprehensive approach to search multimodal fusion model architectures. Using a two-level search approach, the framework optimizes the network architecture, parameters, and performance. Initially, crucial features were efficiently identified from backbone networks, whereas within the cell structure, a weighted fusion operation integrated information from various sources. An architecture that maximizes the classification performance is derived by varying parameters such as temperature and sampling time. The experimental results on the FakeAVCeleb and SWAN-DF datasets demonstrated an impressive AUC value 94.4% achieved with minimal model parameters. https://github.com/Aravinda27/STGS-BMNAS

Neurological voice disorders are caused by problems in the nervous system as it interacts with the larynx. In this paper, we propose to use wavelet scattering transform (WST)-based features in automatic classification of neurological voice disorders. As a part of WST, a speech signal is processed in stages with each stage consisting of three operations–convolution, modulus and averaging–to generate low-variance data representations that preserve discriminability across classes while minimizing differences within a class. The proposed WST-based features were extracted from speech signals of patients suffering from either spasmodic dysphonia (SD) or recurrent laryngeal nerve palsy (RLNP) and from speech signals of healthy speakers of the Saarbruecken voice disorder (SVD) database. Two machine learning algorithms (support vector machine (SVM) and feed forward neural network (NN)) were trained separately using the WST-based features, to perform two binary classification tasks (healthy vs. SD and healthy vs. RLNP) and one multi-class classification task (healthy vs. SD vs. RLNP). The results show that WST-based features outperformed state-of-the-art features in all three tasks. Furthermore, the best overall classification performance was achieved by the NN classifier trained using WST-based features.

The query-document term matching plays an important role in information retrieval. However, the retrieval performance degrades when the documents get matched with the extraneous terms of the query which frequently arises in verbose queries. To address this problem, we generate the dense vector of the entire query and individual query terms using the pre-trained BERT (Bidirectional Encoder Representations from Transformers) model and subsequently analyze their relation to focus on the central terms. We then propose a context-aware attentive extension of unsupervised Markov Random Field-based sequential term dependence model that explicitly pays more attention to those contextually central terms. The proposed model utilizes the strengths of the pre-trained large language model for estimating the attention weight of terms and rank the documents in a single pass without any supervision.

We propose a set of four novel fault-based regression test case prioritization (TCP) techniques for object-oriented programs. We seed bugs into a program to create large number of mutants. We execute each mutant with the originally designed test suite. From this, we record the number of mutants for which a test case fails. Based on this, we prioritize the test cases using four base fault-based prioritization techniques that we have proposed. Finally, we combine the results of our four base prioritizers using three ensemble methods. We have conducted experimental studies to determine the effectiveness of our proposed approaches. Our experimental results show that our proposed TCP techniques exhibit superior performance over related techniques.

Winning lottery tickets refer to sparse subgraphs of deep neural networks which have classification accuracy close to the original dense networks. Resilient connectivity properties of such sparse networks play an important role in their performance. The attempt is to identify a sparse and yet well-connected network to guarantee unhindered information flow. Connectivity in a graph is best characterized by its spectral expansion property. Ramanujan graphs are robust expanders which lead to sparse but highly-connected networks, and thus aid in studying the winning tickets. A feedforward neural network consists of a sequence of bipartite graphs representing its layers. We analyze the Ramanujan graph property of such bipartite layers in terms of their spectral characteristics using the Cheeger’s inequality for irregular graphs. It is empirically observed that the winning ticket networks preserve the Ramanujan graph property and achieve a high accuracy even when the layers are sparse. Accuracy and robustness to noise start declining as many of the layers lose the property. Next we find a robust winning lottery ticket by pruning individual layers while retaining their respective Ramanujan graph property. This strategy is observed to improve the performance of existing network pruning algorithms.

Existing traffic video summarization algorithms are capable of detecting one-class (i.e., collision) anomaly and cannot handle uncertainty issues arising between two-class anomalies, such as collision and near-miss. To address the issues, a new video summarization algorithm, namely Z-number s-based spatio-temporal rough fuzzy granulation (Z-STRFG) is developed. In Z-STRFG, various spatio-temporal features are computed over the video frames and used for obtaining the approximate anomaly-prone regions in terms of granules. In these regions, uncertainty (i.e., fuzziness) may arise among three scenarios, namely collision, near-miss, and normal traffic. Therefore, two types of rough fuzzy granules (RFGs) along with their roughness scores are computed to distinguish the aforesaid three scenarios. For each RFG, Z-number is computed based on the membership value of its roughness score to ensure a higher degree of reliability in the detection of anomaly class. Aforesaid characteristics of Z-STRFG improve its speed and accuracy for traffic anomaly detection. The efficacy of Z-STRFG has been demonstrated over 130 real-time traffic videos containing collisions, near-misses, and normal traffics. The superiority of Z-STRFG over some state-of-the-art is also proved through extensive experiments.

Quantization of the parameters of machine learning models, such as deep neural networks, requires solving constrained optimization problems, where the constraint set is formed by the Cartesian product of many simple discrete sets. For such optimization problems, we study the performance of the Alternating Direction Method of Multipliers for Quantization (ADMM-Q) algorithm, which is a variant of the widely-used ADMM method applied to our discrete optimization problem. We establish the convergence of the iterates of ADMM-Q to certain stationary points. To the best of our knowledge, this is the first analysis of an ADMM-type method for problems with discrete variables/constraints. Based on our theoretical insights, we develop a few variants of ADMM-Q that can handle inexact update rules, and have improved performance via the use of "soft projection" and "injecting randomness to the algorithm". We empirically evaluate the efficacy of our proposed approaches.

In this paper, we propose an augmented Graph Convolutional Network (GCN) mechanism wherein additional information of local interaction patterns between a node with its neighbors (specifically, in the form of distribution of cosine similarity values of a pre-trained node vector with its neighbors) is used to enrich a node's representation prior to training a GCN. This provides additional information about the structural properties of a node, which the standard convolution operation in a GCN can then leverage for obtaining potentially improved effectiveness in a down-stream task. Our experiments demonstrate that adding these node interaction patterns (NIPs) along with an additional noise-contrastive pairwise document similarity objective within a GCN improves the linked document classification task.

Co-morbid disease condition refers to the simultaneous presence of one or more diseases along with the primary disease. A patient suffering from co-morbid diseases possess more mortality risk than with a disease alone. So, it is necessary to predict co-morbid disease pairs. In past years, though several methods have been proposed by researchers for predicting the co-morbid diseases, not much work is done in prediction using knowledge graph embedding using tensor factorization. Moreover, the complex-valued vector-based tensor factorization is not being used in any knowledge graph with biological and biomedical entities. We propose a tensor factorization based approach on biological knowledge graphs. Our method introduces the concept of complex-valued embedding in knowledge graphs with biological entities. Here, we build a knowledge graph with disease-gene associations and their corresponding background information. To predict the association between prevalent diseases, we use ComplEx embedding based tensor decomposition method. Besides, we obtain new prevalent disease pairs using the MCL algorithm in a disease-gene-gene network and check their corresponding inter-relations using edge prediction task.

Contemporary deep learning based semantic inpainting can be approached from two directions. First, and the more explored, approach is to train an offline deep regression network over the masked pixels with an additional refinement by adversarial training. This approach requires a single feed-forward pass for inpainting at inference. Another promising, yet unexplored approach is to first train a generative model to map a latent prior distribution to natural image manifold and during inference time search for the best-matching prior to reconstruct the signal. The primary aversion towards the latter genre is due to its inference time iterative optimization and difficulty to scale to higher resolution. In this paper, going against the general trend, we focus on the second paradigm of inpainting and address both of its mentioned problems. Most importantly, we learn a data driven parametric network to directly predict a matching prior for a given masked image. This converts an iterative paradigm to a single feed forward inference pipeline with around 800X speedup. We also regularize our network with structural prior (computed from the masked image itself) which helps in better preservation of pose and size of the object to be inpainted. Moreover, to extend our model for sequence reconstruction, we propose a recurrent net based grouped latent prior learning. Finally, we leverage recent advancements in high resolution GAN training to scale our inpainting network to 256X256. Experiments (spanning across resolutions from 64X64 to 256X256) conducted on SVHN, Standford Cars, CelebA, CelebA-HQ and ImageNet image datasets, and FaceForensics video datasets reveal that we consistently improve upon contemporary benchmarks from both schools of approaches.

In this paper, we present a texture aware lightweight deep learning framework for iris recognition. Our contributions are primarily three fold. Firstly, to address the dearth of labelled iris data, we propose a reconstruction loss guided unsupervised pre-training stage followed by supervised refinement. This drives the network weights to focus on discriminative iris texture patterns. Next, we propose several texture aware improvisations inside a Convolution Neural Net to better leverage iris textures. Finally, we show that our systematic training and architectural choices enable us to design an efficient framework with upto 100× fewer parameters than contemporary deep learning baselines yet achieve better recognition performance for within and cross dataset evaluations.

To design an efficient fusion scheme for the generation of a highly informative fused image by combining multiple images is still a challenging task in computer vision. A fast and effective image fusion scheme based on multi-resolution singular value decomposition (MR-SVD) with guided filter (GF) has been introduced in this paper. The proposed scheme decomposes an image of two-scale by MR-SVD into a lower approximate layer and a detailed layer containing the lower and higher variations of pixel intensity. It generates lower and details of left focused (LF) and right focused (RF) layers by applying the MR-SVD on each series of multi-focus images. GF is utilized to create a refined and smooth-textured weight fusion map by the weighted average approach on spatial features of the lower and detail layers of each image. A fused image of LF and RF has been achieved by the inverse MR-SVD. Finally, a deep convolutional autoencoder (CAE) has been applied to segment the fused results by generating the trained-patches mechanism. Comparing the results by state-of-the-art fusion and segmentation methods, we have illustrated that the proposed schemes provide superior fused and its segment results in terms of both qualitatively and quantitatively.

In this article, we propose a hierarchical fault localization technique using a deep neural network (DNN). First, we prioritize the functions of a program based on their suspiciousness score. Subsequently, the fault is localized to specific statements within the top k suspected functions, where the value of k is determined heuristically. We use two function-level features to train a DNN for fault localization at the function level. Subsequently, the invocation information of the statements of the top-k functions is used to train another neural network to localize the faulty statement. We also report an extension to our approach for localizing multiple faults. This involves partitioning the failed test cases into clusters such that they target different faults. Our empirical evaluation indicates that our proposed approach requires examining 30.05 to 50.74% less code on an average, as compared to related fault localization techniques.

With the knowledge of adverse effects of harmonics in power systems, parameters (amplitude, frequency, phase, and harmonic contents) estimation of a signal have attracted a rising interest. In recent years sparse representation of a signal has shown a growing interest over the existing techniques that follow Shannon/Nyquist sampling theorem, which requires large storage space, huge computational time, and moreover, the overall process is cost-effective. Keeping in mind the flaws of the conventional methods, this paper presents a novel sparse power system domain based cubature Kalman filter (SPSD-CKF) algorithm, which is the completely new framework that exploits the orthogonal matching pursuit (OMP), a sparse coding algorithm, and K-SVD, a learned dictionary, for sparse representation of a signal. The K-SVD contains the prototype signal-atoms where the signals are described by sparse linear combinations of these atoms, is flexible one and can work with any pursuit method. On the other hand, the OMP is easy to implement and takes less time. The CKF is utilized to estimate the amplitude, phase, frequency, and harmonics using fewer measurements in the sparse domain. To examine the efficacy of learning-based dictionary obtained using the K-SVD, some well-known static dictionaries such as Gabor dictionary (GD) and an overcomplete hybrid dictionary (OHD) have been adopted and their results are compared. Various simulation results suggest that the proposed algorithm provides an efficient mechanism to estimate the parameters and also robust against the noise.

Differential Synthetic Aperture Radar Interferometry (DIn-SAR) can be used for observing and monitoring land surface change over a large area using multi-temporal SAR images. To remove signal decorrelation issues of DInSAR, Persistent Scatterer InSAR (PSInSAR) has come into practice over the last decade that can measure deformation at small scales and fine accuracy (mm-level). In this paper, we estimate ground deformation of Kolkata, India during 2007-2011 using a spatial correlation based PSInSAR method. Twenty L-band (ALOS PALSAR) SAR images along ascending orbit over Kolkata city were utilised to extract deformation time series, and a validation exercise was carried out using groundwater level data of two wells. PSInSAR extracted a large number of measurement points (363 per km2for a total 568014 points), and we found a mean annual deformation rate of -16 to +16 mm/year in Kolkata area along Line-Of-Sight (LOS) of ALOS PALSAR. Comparison between PSInSAR derived deformation and groundwater level fluctuation reveals a good correlation using time series analysis with some discrepancies.

Fault localization (FL) is a time consuming and tedious task during program debugging. Most of the existing FL methods use statement coverage information to prioritize the statements based upon a computed suspiciousness score. We use predicate level execution trace to train a multilayer perceptron neural network model for effective fault localization. After prioritizing the fault at predicate level, we search the statements bounded by the predicates. Also, dynamic slicing is used to reduce the search space. We have experimentally studied the performance of our approach over Siemens suite and Space program and found that it is performing on an average 39.12% more effectively than DStar, a state-of-the-art bug localization technique.

Deep learning based object detectors require thousands of diversified bounding box and class annotated examples. Though image object detectors have shown rapid progress in recent years with release of multiple large scale static image datasets, object detection on videos still remains an open problem due to unavailability of annotated video frames. Having a robust video object detector is an essential component for video understanding and curating large scale automated annotations in videos. Domain difference between images and videos makes the transferability of image object detectors to videos sub-optimal. The most common solution is to use weakly supervised annotations where a video frame has to be tagged for presence/absence of object categories. This still takes up manual effort. In this paper we take a step forward to attain zero supervision on video domain by adapting the concept of unsupervised adversarial image-to-image translation to perturb static high quality images to be visually indistinguishable from set of video frmes. We assume the presence of a fully annotated static image dataset and an unannotated video frames. Object detector is trained on adversarially transformed image dataset using the annotations of original dataset. Experiments on Youtube-Objects and Youtube-Objects-Subset datasets with two contemporary baseline object detectors reveal that such unsupervised pixel level domain adaptation boosts the generalization performance on video frames compared to direct application of image object detector. Also we achieve competitive performance compared to recent baselines of weakly supervised methods. This paper can be seen as an application of image translation for cross domain object detection.

Ultrasound (US) is widely used as a low-cost alternative to computed tomography or magnetic resonance and primarily for preliminary imaging. Since speckle intensity in US images is inherently stochastic, readers are often challenged in their ability to identify the pathological regions in a volume of a large number of images. This paper introduces a generalized approach for volumetric segmentation of structures in US images and volumes. We employ an iterative random walks (IRW) solver, a random forest learning model, and a gradient vector flow (GVF) based interframe belief propagation technique for achieving cross-frame volumetric segmentation. At the start, a weak estimate of the tissue structure is obtained using estimates of parameters of a statistical mechanics model of US tissue interaction. Ensemble learning of these parameters further using a random forest is used to initialize the segmentation pipeline. IRW is used for correcting the contour in various steps of the algorithm. Subsequently, a GVF-based interframe belief propagation is applied to adjacent frames based on the initialization of contour using information in the current frame to segment the complete volume by frame-wise processing. We have experimentally evaluated our approach using two different datasets. Intravascular ultrasound (IVUS) segmentation was evaluated using 10 pullbacks acquired at 20 MHz and thyroid US segmentation is evaluated on 16 volumes acquired at [Formula: see text] MHz. Our approach obtains a Jaccard score of [Formula: see text] for IVUS segmentation and [Formula: see text] for thyroid segmentation while processing each frame in [Formula: see text] for the IVUS and in [Formula: see text] for thyroid segmentation without the need of any computing accelerators such as GPUs.

The goal of our work is to use visual attention to enhance autonomous driving performance. We present two methods of predicting visual attention maps. The first method is a supervised learning approach in which we collect eye-gaze data for the task of driving and use this to train a model for predicting the attention map. The second method is a novel unsupervised approach where we train a model to learn to predict attention as it learns to drive a car. Finally, we present a comparative study of our results and show that the supervised approach for predicting attention when incorporated performs better than other approaches.

Fast and accurate estimation of harmonics of a typical power system signal is very much desirable for power quality assessment. This paper proposes an application of the cubature Kalman filter (CKF)for estimating the parameters of harmonics, sub-harmonics, and inter-harmonics of a signal in presence of noise. CKF utilizes a third-degree spherical radial cubature rule to estimate the probability density functions of the states as well as the measurements. At the same time, this technique does not require any linearization and saves the computational time. The effectiveness of CKF has been compared with UKF by performing various test cases. It is observed from the simulation results that CKF exhibits superior performance in estimating the parameters of harmonics, inter-harmonics, and sub-harmonics of a distorted power system static as well as the dynamic signal by virtue of execution time and accuracy.

Differential Synthetic Aperture Radar interferometry (DIn-SAR) is an efficient technique for observing and monitoring ground deformation over a large area at a millimetric level using multi-temporal SAR images. However, the phase decorrelation phenomena along with other bottlenecks, such as atmospheric nuisance, orbital, and DEM inaccuracy limit the accuracy of the traditional DInSAR technique. Persistent Scatterer InSAR (PSInSAR), based on DInSAR, circumvents these limitations using temporally stable reflectors or permanent scatterers (PS) of earth surface using a long temporal stack of SAR images. In this paper, a spatial correlation based PSInSAR technique is applied to detect the ground deformation of Las Vegas, Nevada, US between 2002 to 2010. This aim of this study is focused on the precise estimation and validation of ground deformation using the PSInSAR analysis of descending pass Envisat ASAR data, and further cross-validation of displacement time-series is carried out with GPS time-series observations. Archived C-band Envisat ASAR SAR data stack was used to estimate ground deformation map of the concerned area. A good correlation is observed between PSInSAR and GPS time-series. We have observed a mean annual deformation rates of -5 to 5.1 mm/ year during 2002-2010 along satellite line of sight (LOS).

The existence of mixed pixels is common in hyperspectral data. Although, the proportion of each spectral signature in a given mixed pixels scenario may be determined through the spectral unmixing operations. In this work, Particle Swarm Optimization (PSO) based approach is proposed in order to estimate the abundances fractions for spectral unmixing. It calculates the position in order to estimate the fractions. In this, the concept of particles per solution is omitted in order to do unmixing. Hence, each pixel of data is our particle, and the solution is its abundance fractions. This approach is less computationally complex as compared to other proposed PSO based approaches for abundance estimation. Herein, supervised linear mixing model and spatially correlated data are the assumptions considered for unmixing operation. The proposed method is tested on simulated data and it has been observed to be performing well.

In this paper, we introduce a system named Portable Personality Recognizer (PPR), which classifies the personality of an individual using his/her transitions of affective states. This work attempts to reveal the latent relationship between emotions and personality of a person. Here, we train a hidden Markov model (HMM) with observable emotional states viz. Happiness (H), Anger (A), Surprise (S) and Disgust (D) and the hidden traits viz. Psychoticism (P), Extraversion (E) and Neuroticism (N). Based on the model, the system estimates the personality as Psychotic, Extravert or Neurotic. It does so by capturing the facial images of an individual using a visible and a thermal camera to decide the present affective state of the person. The emotion classification is carried out using fused eigenfeatures from the visible and blood perfused thermal images. The emotional state changes are observed using the trained HMM to estimate the personality. The proposed hardware prototype consists of a Banana Pi board with a seven inch LCD screen having a thermal and a visible camera add-ons. The system achieves an emotion classification accuracy of 87.145 percent, while an accuracy of 87.87 percent is achieved for personality recognition.

Neural activities recorded using electroencephalography (EEG) are mostly contaminated with eye blink (EB) artifact. This results in undesired activation of brain-computer interface (BCI) systems. Hence, removal of EB artifact is an important issue in EEG signal analysis. Of late, several artifact removal methods have been reported in the literature and they are based on independent component analysis (ICA), thresholding, wavelet transformation, etc. These methods are computationally expensive and result in information loss which makes them unsuitable for online BCI system development. To address the above problems, we have investigated sparsity-based EB artifact removal methods. Two sparsity-based techniques namely morphological component analysis (MCA) and K-SVD-based artifact removal method have been evaluated in our work. MCA-based algorithm exploits the morphological characteristics of EEG and EB using predefined Dirac and discrete cosine transform (DCT) dictionaries. Next, in K-SVD-based algorithm an overcomplete dictionary is learned from the EEG data itself and is designed to model EB characteristics. To substantiate the efficacy of the two algorithms, we have carried out our experiments with both synthetic and real EEG data. We observe that the K-SVD algorithm, which uses a learned dictionary, delivers superior performance for suppressing EB artifacts when compared to MCA technique. Finally, the results of both the techniques are compared with the recent state-of-the-art FORCe method. We demonstrate that the proposed sparsity-based algorithms perform equal to the state-of-the-art technique. It is shown that without using any computationally expensive algorithms, only with the use of over-complete dictionaries the proposed sparsity-based algorithms eliminate EB artifacts accurately from the EEG signals.

Precipitation nowcasting is an important component for accurate weather modeling and Doppler radar data acts as an important input for nowcasting models. In this work, we propose a deep learning based approach for radar echo states prediction. Our approach uses a hybrid structure of convolutions within Long Short Term Memory recurrent network structure and a discriminator network is added in the loss objective to refine the predictions acting as a regularizer. This models the spatio-temporal nature of the problem explicitly in the neural network. We show that this model can be applied for fine grained short term precipitation prediction with improvement in evaluation metrics as compared to strong baselines. The proposed model improves recall score by 11% compared to without adversarial regularization. Results are presented using usual train, test strategy for the task of echo state prediction and derived precipitation based skill scores on the data from Seattle, WA, USA.

Deep learning based land cover classification algorithms have recently been proposed in the literature. In hyperspectral images (HSIs), they face the challenges of large dimensionality, spatial variability of spectral signatures, and scarcity of labeled data. In this paper, we propose an end-to-end deep learning architecture that extracts band specific spectral-spatial features and performs land cover classification. The architecture has fewer independent connection weights and thus requires fewer training samples. The method is found to outperform the highest reported accuracies on popular HSI data sets.

Statistical models built on historical data are often found to be effective in forecasting Indian summer monsoon. However, linear models are found to be inadequate, and non-linear models like neural networks provide better performance. In this article, we study the use of recurrent neural network for long range forecast of Indian monsoon at lead of one season. Recurrent network model the sequential structure of the historical data yielding higher importance to near years than distant ones. A detailed study of the effectiveness of a set of fourteen global climatic predictors is carried out and Indian summer monsoon is predicted. The proposed recurrent network model gives mean absolute error in prediction as 3.3%, which is appreciable for forecasting complex monsoon phenomenon. Prediction skill of our model outperforms traditional multilayer neural network and it is also found to be superior than existing India Meteorology Department models.

Repeat Purchases have become increasingly important in measuring customer's satisfaction and loyalty to e-commerce websites in regard to online shopping. In this paper, we first propose a model for estimating repeat purchase frequency in a given time period from a given product category using Poisson/Gamma model. Second, we estimate the purchase probabilities of different product types in a product category for each customer using Dirichlet model. Experimental results on data collected by a real-world e-commerce website show that it can predict a user's average repeat purchase frequency along with their product types with decent accuracy. We also argue that the output of our models can be used as prior information to enhance the performance of time-sensitive recommendation.

The causes of the many communicable diseases have some link with the geographical locations. In the developing countries, the patient information are not often collected systemically due to lack of technological support and scarcity of health professionals. As a result, the actual causes of different community diseases are not properly investigated and it delays to take preventive actions, especially in the rural areas. In this work, a service oriented framework has been proposed for collecting the health information of the patients along with their actual geo-locations to generate the health map of the region which may help in adopting mitigation measures. A mobile application has been developed for collecting these information by the health workers or the health volunteers. The application can be tuned to provide suggestive primary treatment of the disease. The collected volunteered disease information can be analyzed with respect to that geospatial information according to guidelines of the health experts and the disease map can produced on-the-fly. The disease map helps to predict the spreading trend of a disease in the context of geospatial attributes. This may facilitate, the public health decision makers to take preventive actions to mitigate the spread of the diseases.

Various statistical and machine learning based algorithms have been proposed in literature for selecting an informative subset of genes from micro array data sets. The recent trend is to use functional knowledge to aid the gene selection process. In this paper we propose a clustering algorithm which generates multiple views (clusters) from the micro array expression profiles, each representing a particular facet of the data. Such multiple clusters are found to represent strongly connected regions of the known protein -- protein interaction (PPI) networks, perhaps corresponding to those responsible for certain biological processes. Thus we integrate micro array data clustering with PPI knowledge to obtain enriched gene sets. Results on benchmark micro array data sets demonstrate the competitiveness of our method compared to gene selection techniques.

Prediction of spatial attributes has attracted significant research interest in recent years. It is challenging especially when spatial data contain errors and missing values. Geostatistical estimators are used to predict the missing attribute values from the observed values of known surrounding data points, a general form of which is referred as kriging in the field of geographic information system and remote sensing. The proposed semantic kriging ( SemK) tries to blend the semantics of spatial features (of surrounding data points) with ordinary kriging (OK) method for prediction of the attribute. Experimentation has been carried out with land surface temperature data of four major metropolitan cities in India. It shows that SemK outperforms the OK and most of the existing spatial interpolation methods.

In the Recommendation Problem, it is often important to find a set of items similar to a particular item or a group of items. This problem of finding similar items for the recommendation task may also be viewed as a link prediction problem in a network, where the items can be treated as the nodes. The strength of the edge connecting two items represents the similarity between the items. In this context, a central challenge is to suitably define an appropriate dissimilarity function between the items. For content based recommender systems, the dissimilarity function should take into account the individual attributes of the items. The same attribute may have different importances in different parts of the underlying network. We focus on the problem of learning a suitable dissimilarity function between items and address it by formulating it as a constrained optimization problem which captures the local weightages of the attributes in different regions of the graph. The constraints are imposed in such a way that the non-connected nodes show higher value of dissimilarity than the connected nodes. The local tuning of the weights learns the optimal value of weights in various parts of the network: from the portions having rich graph information to the portions having only content information. Detailed experimentation shows the superiority of the proposed algorithm over the Adamic Adar metric as well as logistic regression methodology.

Bag of visual words approach for image retrieval does not exploit the spatial distribution of visual words in an image. Previous attempts to incorporate the spatial distribution include modification of visual vocabulary using visual phrases along with visual words and use of spatial pyramid matching (SPM) techniques for comparing two images. This paper proposes a novel extended vector space based image retrieval technique which takes into account the spatial occurrence (context) of a visual word in an image along with the co-occurrence of other visual words in a pre-defined region (block) of the image obtained by quadtree decomposition of the image up to a fixed level of resolution. Experiments show a 19.22% increase in Mean Average Precision (MAP) over the BoW approach for the Caltech 101 database.

Collaborative filtering is a widely used technique for rating prediction in recommender systems. Memory based collaborative filtering algorithms assign weights to the users to capture similarities between them. The weighted average of similar users' ratings for the test item is output as prediction. We propose a memory based algorithm that is markedly different from the existing approaches. We use preference relations instead of absolute ratings for similarity calculations, as preference relations between items are generally more consistent than ratings across like-minded users. Each user's ratings are viewed as a preference graph. Similarity weights are learned using an iterative method motivated by online learning. These weights are used to create an aggregate preference graph. Ratings are inferred to maximally agree with this aggregate graph. The use of preference relations allows the rating of an item to be influenced by other items, which is not the case in the weighted-average approaches of the existing techniques. This is very effective when the data is sparse, specially for the items rated by few users. Our experiments show that the our method outperforms other methods in the sparse regions. However, for dense regions, sometimes our results are comparable to the competing approaches, and sometimes worse.

An emotional text may be judged to belong to multiple emotion categories because it may evoke different emotions with varying degrees of intensity. For emotion analysis of text in a supervised manner, it is required to annotate text corpus with emotion categories. Because emotion is a very subjective entity, producing reliable annotation is of prime requirement for developing a robust emotion analysis model, so it is wise to have the data set annotated by multiple human judges and generate an aggregated data set provided that the emotional responses provided by different annotators over the data set exhibit substantial agreement. In reality, multiple emotional responses for an emotional text are common. So, the data set is a multilabel one where a single data item may belong to more than one category simultaneously. This article presents a new agreement measure to compute interannotator reliability in multilabel annotation. The new reliability coefficient has been applied to measure the quality of an emotion text corpus. The procedure for generating aggregated data and some corpus cleaning techniques are also discussed.

The performance of the acoustic models is highly reflective on the overall performance of any continuous speech recognition system. Hence generation of an accurate and robust acoustic model holds the key to satisfactory recognition performance. As phones are found to vary according to the position of occurrence within a particular word, context information is of prime importance in acoustic modeling of phonetic signals. In this paper we look at the effect of triphone-based acoustic modeling over monophone based acoustic models in the context of continuous speech recognition in Bengali. Keeping in mind the lack of training resources for triphone-based acoustic modeling in Bengali, we have also described herein, the method of generating triphone clusters using decision tree based techniques. These triphone clusters have then been used to generate tied-state triphone based acoustic models to be used in a continuous speech recognizer.

Combination of structure and content features is necessary for effective retrieval and classification of XML documents. Composite kernels provide a way for fusion of content and structure information. In this paper, we demonstrate that a linear combination of simple and low cost kernels such as cosine similarity on terms and selective paths provide a good classification performance. We also propose a corpus-driven entropy-based heuristic for determining the optimal combination weights. Classification experiments performed on the INEX 1.3 XML corpus, demonstrate that the composite kernel classifier achieves significantly better performance as compared to complex and time consuming approaches.

This paper proposes a parallel associative memory (PAM) based face recognition system. An image is split into blocks which are processed in parallel by conventional associative memory. The combined output of these are used to discriminate between faces. Further, a parameter termed as run length count has been used to discriminate between similar faces. This approach helps to scale up associative memory to higher resolution images with more detailed features thus improving recognition performance. To analyze the efficiency and goodness of the proposed system the experiments have been performed on ORL face database. The proposed method out performed most of the existing methods.

We propose a hybridization of collaborative filtering and content based recommendation system. Attributes used for content based recommendations are assigned weights depending on their importance to users. The weight values are estimated from a set of linear regression equations obtained from a social network graph which captures human judgment about similarity of items.

Clustering sensors into groups, so that sensors communicate information only to cluster-heads and then the cluster-heads communicate the aggregated information to the base station, saves energy and thus prolongs network lifetime. Adapting this approach, we propose a Distributive Energy Efficient Adaptive Clustering (DEEAC) protocol. This protocol is adaptive in terms of data reporting rates and residual energy of each node within the network. Motivated by the LEACH protocol [I], we extend its stochastic cluster selection algorithm for networks having spatio-temporal variations in data reporting rates across different regions. Simulation results demonstrate that DEEAC is able to distribute energy consumption more effectively among the sensors, thereby prolonging the network lifetime by as much as 50% compared to LEACH.

We present an efficient dynamic algorithm for clustering undirected graphs, whose edge property is changing continuously. The algorithm maintains clusters of high quality in presence of insertion and deletion (update) of edges. The algorithm is motivated by the minimum-cut tree based partitioning algorithm of [3] and [4]. It takes O(k3) time for each update processing, where k is the maximum size of any cluster. This is the worst case time complexity, and in general update time taken is much less. The clusters satisfy the bicriteria for quality guarantee proposed in [3].

Stemmers attempt to reduce a word to its stem or root form and are used widely in information retrieval tasks to increase the recall rate. Most popular stemmers encode a large number of language-specific rules built over a length of time. Such stemmers with comprehensive rules are available only for a few languages. In the absence of extensive linguistic resources for certain languages, statistical language processing tools have been successfully used to improve the performance of IR systems. In this article, we describe a clustering-based approach to discover equivalence classes of root words and their morphological variants. A set of string distance measures are defined, and the lexicon for a given text collection is clustered using the distance measures to identify these equivalence classes. The proposed approach is compared with Porter's and Lovin's stemmers on the AP and WSJ subcollections of the Tipster dataset using 200 queries. Its performance is comparable to that of Porter's and Lovin's stemmers, both in terms of average precision and the total number of relevant documents retrieved. The proposed stemming algorithm also provides consistent improvements in retrieval performance for French and Bengali, which are currently resource-poor.

We address the problem of answering broad-topic queries on the World Wide Web. We present a link based analysis algorithm SelHITS, which is an improvement over Kleinberg's HITS [2] algorithm. We introduce the concept of virtual links to exploit the latent information in the hyperlinked environment. We propose a novel approach to calculate hub and authority values. We also present a selective expansion method which avoids topic drift and provides results consistent with only one interpretation of the query, even if the query is ambiguous. Initial experimental evaluation and user feedback show that our algorithm indeed distills the most important and relevant pages for broad-topic queries. We also infer that there exists a uniform notion of quality of search results within users.

The paper describes a probabilistic active learning strategy for support vector machine (SVM) design in large data applications. The learning strategy is motivated by the statistical query model. While most existing methods of active SVM learning query for points based on their proximity to the current separating hyperplane, the proposed method queries for a set of points according to a distribution as determined by the current separating hyperplane and a newly defined concept of an adaptive confidence factor. This enables the algorithm to have more robust and efficient learning capabilities. The confidence factor is estimated from local information using the k nearest neighbor principle. The effectiveness of the method is demonstrated on real-life data sets both in terms of generalization performance, query complexity, and training time.

We propose a rough-fuzzy hybridization scheme for case generation. Fuzzy set theory is used for linguistic representation of patterns, thereby producing a fuzzy granulation of the feature space. Rough set theory is used to obtain dependency rules which model informative regions in the granulated feature space. The fuzzy membership functions corresponding to the informative regions are stored as cases along with the strength values. Case retrieval is made using a similarity measure based on these membership functions. Unlike the existing case selection methods, the cases here are cluster granules and not sample points. Also, each case involves a reduced number of relevant features. These makes the algorithm suitable for mining data sets, large both in dimension and size, due to its low-time requirement in case generation as well as retrieval. Superiority of the algorithm in terms of classification accuracy and case generation and retrieval times is demonstrated on some real-life data sets.

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.

In this article, we describe an unsupervised feature selection algorithm suitable for data sets, large in both dimension and size. The method is based on measuring similarity between features whereby redundancy therein is removed. This does not need any search and, therefore, is fast. A new feature similarity measure, called maximum information compression index, is introduced. The algorithm is generic in nature and has the capability of multiscale representation of data sets. The superiority of the algorithm, in terms of speed and performance, is established extensively over various real-life data sets of different sizes and dimensions. It is also demonstrated how redundancy and information loss in feature selection can be quantified with an entropy measure.

A problem gaining interest in pattern recognition applied to data mining is that of selecting a small representative subset from a very large data set. In this article, a nonparametric data reduction scheme is suggested. It attempts to represent the density underlying the data. The algorithm selects representative points in a multiscale fashion which is novel from existing density-based approaches. The accuracy of representation by the condensed set is measured in terms of the error in density estimates of the original and reduced sets. Experimental studies on several real life data sets show that the multiscale approach is superior to several related condensation methods both in terms of condensation ratio and estimation error. The condensed set obtained was also experimentally shown to be effective for some important data mining tasks like classification, clustering, and rule generation on large data sets. Moreover, it is empirically found that the algorithm is efficient in terms of sample complexity.

The problem of segmentation of multispectral satellite images is addressed. An integration of rough-set-theoretic knowledge extraction, the Expectation Maximization (EM) algorithm, and minimal spanning tree (MST) clustering is described. EM provides the statistical model of the data and handles the associated measurement and representation uncertainties. Rough-set theory helps in faster convergence and in avoiding the local minima problem, thereby enhancing the performance of EM. For rough-set-theoretic rule generation, each band is discretized using fuzzy-correlation-based gray-level thresholding. MST enables determination of nonconvex clusters. Since this is applied on Gaussians, determined by granules, rather than on the original data points, time required is very low. These features are demonstrated on two IRS-1A four-band images. Comparison with related methods is made in terms of computation time and a cluster quality measure.

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.

The paper summarizes the different characteristics of Web data, the basic components of Web mining and its different types, and the current state of the art. The reason for considering Web mining, a separate field from data mining, is explained. The limitations of some of the existing Web mining methods and tools are enunciated, and the significance of soft computing (comprising fuzzy logic (FL), artificial neural networks (ANNs), genetic algorithms (GAs), and rough sets (RSs) are highlighted. A survey of the existing literature on "soft Web mining" is provided along with the commercially available systems. The prospective areas of Web mining where the application of soft computing needs immediate attention are outlined with justification. Scope for future research in developing "soft Web mining" systems is explained. An extensive bibliography is also provided.

An algorithm for data condensation using support vector machines (SVM) is presented. The algorithm extracts data points lying close to the class boundaries, which form a much reduced but critical set for classification. The problem of large memory requirements for training SVM in batch mode is circumvented by adopting an active incremental learning algorithm. The learning strategy is motivated from the condensed nearest neighbor classification technique. Experimental results presented show that such active incremental learning enjoy superiority in terms of computation time and condensation ratio, over related methods.