Quadratic Kernel Research Articles

Pulse shape discrimination of LSO and LuYAP scintillators based on Discrete Fractional Fourier Transform

Parallax error decreases the accuracy of the Positron Emission Tomography (PET) scanner. One of suitable solutions to reduce this error is to gain the depth of interaction information in a PET which uses the phoswich detectors. The pulse shape of the scintillator material can identify the corresponding layer of interaction within the phoswich detector by using Pulse Shape Discrimination (PSD) methods. In this work, we propose the PSD based on a Discrete Fractional Fourier Transform (DFRFT) to extract the features of the scintillation pulses. Then, we use the Support Vector Machine (SVM) to classify these features. A data set consists of 100 000 scintillation pulses for LSO and LuYAP crystals are discriminated using the proposed method. Different fraction factors of the DFRFT are studied to select the optimum one. Also, the SVM is applied using linear, radial basis function, or quadratic kernel. The highest efficiency of the proposed PSD is 96.3% when the applied fraction factor is 0.8 and a quadratic SVM kernel is used.

Non-parametric Estimation of Quadratic Hawkes Processes for Order Book Events

We propose an actionable calibration procedure for general Quadratic Hawkes models of order book events (market orders, limit orders, cancellations). One of the main features of such models is to encode not only the influence of past events on future events but also, crucially, the influence of past price changes on such events. We show that the empirically calibrated quadratic kernel is well described by a diagonal contribution (that captures past realised volatility), plus a rank-one “Zumbach” contribution (that captures the effect of past trends). We find that the Zumbach kernel is a power-law of time, as are all other feedback kernels. As in many previous studies, the rate of truly exogenous events is found to be a small fraction of the total event rate. These two features suggest that the system is close to a critical point – in the sense that stronger feedback kernels would lead to instabilities.

Local feature descriptors based ECG beat classification.

ECG beat type analysis is important in the detection of various heart diseases. The ECG beats give useful information about the status of the monitored heart condition. Up to now, various artificial intelligence-based methods have been proposed for ECG based heart failure detection. These methods were generally based on either time or frequency domain signal processing routines. In this study, we propose a different approach for ECG beat classification. The proposed approach is based on image processing. Thus, the initial step of the proposed work is converting the ECG beat signals to the ECG beat images. To do that, the ECG beat snapshots are initially saved as ECG beat images and then local feature descriptors are considered for feature extraction from ECG beat images. Eight local feature descriptors namely Local Binary Patterns, Frequency Decoded LBP, Quaternionic Local Ranking Binary Pattern, Binary Gabor Pattern, Local Phase Quantization, Binarized Statistical Image Features, CENsus TRansform hISTogram and Pyramid Histogram of Oriented Gradients are considered for feature extraction. The Support Vector Machines (SVM) classifier is used in the classification stage of the study. Linear, Quadratic, Cubic and Gaussian kernel functions are used in the SVM classifier. Five types of ECG beats from the MIT-BIH arrhythmia dataset are considered in experiments and the classification accuracy is used for performance measure. To construct a balanced training and test sets, 5000 and 10,000 ECG beat samples are randomly selected and are used in experiments in tenfold cross-validation fashion. The obtained results show that the proposed method is quite efficient where the calculated accuracy score is 99.9% and the comparisons with the state-of-the-art method show that the proposed method outperforms other methods.

Local fractional Newton\u2019s inequalities involving generalized harmonic convex functions

A new auxiliary result based on a three step quadratic kernel utilizing the concepts of local fractional calculus is obtained. Using this new auxiliary result we have several new Newton type inequalities whose power q of local fractional derivative in modulus is a generalized harmonic convex function.

Deep Learning—Powered Iterative Combinatorial Auctions

In this paper, we study the design of deep learning-powered iterative combinatorial auctions (ICAs). We build on prior work where preference elicitation was done via kernelized support vector regressions (SVRs). However, the SVR-based approach has limitations because it requires solving a machine learning (ML)-based winner determination problem (WDP). With expressive kernels (like gaussians), the ML-based WDP cannot be solved for large domains. While linear or quadratic kernels have better computational scalability, these kernels have limited expressiveness. In this work, we address these shortcomings by using deep neural networks (DNNs) instead of SVRs. We first show how the DNN-based WDP can be reformulated into a mixed integer program (MIP). Second, we experimentally compare the prediction performance of DNNs against SVRs. Third, we present experimental evaluations in two medium-sized domains which show that even ICAs based on relatively small-sized DNNs lead to higher economic efficiency than ICAs based on kernelized SVRs. Finally, we show that our DNN-powered ICA also scales well to very large CA domains.

Constraining teleparallel gravity through Gaussian processes

We apply Gaussian processes (GP) in order to impose constraints on teleparallel gravity and its f(T) extensions. We use available H(z) observations from (i) cosmic chronometers data (CC); (ii) Supernova type Ia (SN) data from the compressed pantheon release together with the CANDELS and CLASH multi-cycle treasury programs; and (iii) baryonic acoustic oscillation (BAO) datasets from the sloan digital sky survey. For the involved covariance functions, we consider four widely used choices, namely the square exponential, Cauchy, Matérn and rational quadratic kernels, which are consistent with one another within 1σ confidence levels. Specifically, we use the GP approach to reconstruct a model-independent determination of the Hubble constant H 0, for each of these kernels and dataset combinations. These analyses are complemented with three recently announced literature values of H 0, namely (i) Riess ; (ii) H0LiCOW collaboration ; and (iii) Carnegie–Chicago Hubble programme . Additionally, we investigate the transition redshift between the decelerating and accelerating cosmological phases through the GP reconstructed deceleration parameter. Furthermore, we reconstruct the model-independent evolution of the dark energy equation of state, and finally reconstruct the allowed f(T) functions. As a result, the ΛCDM model lies inside the allowed region at 1σ in all the examined kernels and datasets, however a negative slope for f(T) versus T is slightly favoured.

Quadratic tuned kernel parameter in Non-linear support vector machine (SVM) for agarwood oil compounds quality classification

<span>This paper presents the analysis of agarwood oil compounds quality classification by tuning quadratic kernel parameter in Support Vector Machine (SVM). The experimental work involved of agarwood oil samples from low and high qualities. The input is abundances (%) of the agarwood oil compounds and the output is the quality of the oil either high or low. The input and output data were processed by following tasks; i) data processing which covers normalization, randomization and data splitting into two parts in which training and testing database (ratio of 80%:20%), and ii) data analysis which covers SVM development by tuning quadratic kernel parameter. The training dataset was used to be train the SVM model and the testing dataset was used to test the developed SVM model. All the analytical works are performed via MATLAB software version R2013a. The result showed that, quadratic tuned kernel parameter in SVM model was successful since it passed all the performance criteria’s in which accuracy, precision, confusion matrix, sensitivity and specificity. The finding obtained in this paper is vital to the agarwood oil and its research area especially to the agarwood oil compounds classification system.</span>

AUTOMATIC ARRHYTHMIA DETECTION ALGORITHM USING STATISTICAL AND AUTOREGRESSIVE MODEL FEATURES

Human heart healthiness is one of the major components to determine a person’s overall healthiness. Automatic arrhythmia detection is important in a remote area where there is a lack of experienced cardiologists. In this work, an automatic arrhythmia detection algorithm is developed using statistical and autoregressive features to assist medical officers in the diagnosis of arrhythmia diseases. Basic statistical components namely mean, energy, standard deviation, mean absolute deviation, fractal, inter-quartile range and min/max value, were calculated. Alongside with statistical features, 10th order auto-regressive model parameters are used as input features to support vector machine (SVM). All features are calculated using an electrocardiogram (ECG) signals windowed into per beat manner. The proposed algorithm is able to classify normal ECG beat and five types of abnormal ECG beat; paced beat, right & left bundle branch block beat, premature ventricular contraction beat and aberrated atrial premature beat. By using SVM with quadratic and cubic kernel function, the proposed algorithm achieved the best accuracy of 95.8%.
 Keywords: ECG, cascade-SVM, AR model, statistical model, heart condition, computer-aided diagnosis

Breast Cancer Lesion Detection From Cranial-Caudal View of Mammogram Images Using Statistical and Texture Features Extraction

Breast cancer is the most common cancer among women in the world today. Mammography screening gives vital information about normal and abnormal regions. The task is to detect the lesion in mammograms using computer-aided diagnosis techniques. The automated detection of cancer decreases the mortality rate and manual error. In this work, the statistical (mean, variance, skewness, kurtosis, energy and entropy) and tamura features (coarseness, contrast and directionality) were extracted from the Cranial-Caudal (CC) view of mammogram images collected from the M.S. Ramaiah Memorial Hospital, Bangalore. The support vector machine was used for classification. Different support vector machine kernels were used and results were tabulated. The highest accuracy was obtained for linear and quadratic kernels with 95.7% with sensitivity of 100% and specificity of 91%.

Quantifying natural organic matter concentration in water from climatological parameters using different machine learning algorithms

Abstract The present understanding of how changes in climate conditions will impact the flux of natural organic matter (NOM) from the terrestrial to aquatic environments and thus aquatic dissolved organic carbon (DOC) concentrations is limited. In this study, three machine learning algorithms were used to predict variations in DOC concentrations in an Australian drinking water catchment as a function of climate, catchment and physical water quality data. Four independent variables including precipitation, temperature, leaf area index and turbidity (n = 5,540) were selected from a large dataset to develop and train each machine learning model. The accuracy of the multivariable linear regression, support vector regression (SVR) and Gaussian process regression algorithms with different kernel functions was determined using adjusted R-squared (adj. R2), root-mean-squared error (RMSE) and mean absolute error (MAE). Model accuracy was very sensitive to the time interval used to average climate observations prior to pairing with DOC observations. The SVR model with a quadratic kernel function and a 12-day time interval between climate and water quality observations outperformed the other machine learning algorithms (adj. R2 = 0.71, RMSE = 1.9, MAE = 1.35). The area under the receiver operating characteristic curve method (AUC) confirmed that the SVR model could predict 92% of the elevated DOC observations; however, it was not possible to estimate DOC values at specific sampling sites in the catchment, probably due to the complex local geological and hydrological changes in the sites that directly surround and feed each sampling point. Further research is required to establish potential relationships between climatological data and NOM concentration in other water catchments – especially in the face of a changing climate.

Tumor cell sensitivity to vemurafenib can be predicted from protein expression in a BRAF-V600E basket trial setting

BackgroundGenetics-based basket trials have emerged to test targeted therapeutics across multiple cancer types. However, while vemurafenib is FDA-approved for BRAF-V600E melanomas, the non-melanoma basket trial was unsuccessful, suggesting mutation status is insufficient to predict response. We hypothesized that proteomic data would complement mutation status to identify vemurafenib-sensitive tumors and effective co-treatments for BRAF-V600E tumors with inherent resistance.MethodsReverse Phase Proteomic Array (RPPA, MD Anderson Cell Lines Project), RNAseq (Cancer Cell Line Encyclopedia) and vemurafenib sensitivity (Cancer Therapeutic Response Portal) data for BRAF-V600E cancer cell lines were curated. Linear and nonlinear regression models using RPPA protein or RNAseq were evaluated and compared based on their ability to predict BRAF-V600E cell line sensitivity (area under the dose response curve). Accuracies of all models were evaluated using hold-out testing. CausalPath software was used to identify protein-protein interaction networks that could explain differential protein expression in resistant cells. Human examination of features employed by the model, the identified protein interaction networks, and model simulation suggested anti-ErbB co-therapy would counter intrinsic resistance to vemurafenib. To validate this potential co-therapy, cell lines were treated with vemurafenib and dacomitinib (a pan-ErbB inhibitor) and the number of viable cells was measured.ResultsOrthogonal partial least squares (O-PLS) predicted vemurafenib sensitivity with greater accuracy in both melanoma and non-melanoma BRAF-V600E cell lines than other leading machine learning methods, specifically Random Forests, Support Vector Regression (linear and quadratic kernels) and LASSO-penalized regression. Additionally, use of transcriptomic in place of proteomic data weakened model performance. Model analysis revealed that resistant lines had elevated expression and activation of ErbB receptors, suggesting ErbB inhibition could improve vemurafenib response. As predicted, experimental evaluation of vemurafenib plus dacomitinb demonstrated improved efficacy relative to monotherapies.Conclusions: Combined, our results support that inclusion of proteomics can predict drug response and identify co-therapies in a basket setting.

ESTIMATING WATER LEVEL IN THE URMIA LAKE USING SATELLITE DATA: A MACHINE LEARNING APPROACH

Abstract. Lakes play a pivotal role in the development of cities and have major impacts on the ecosystem balancing of the area. Remote sensing techniques and advanced modeling methods make it possible to monitor natural phenomena, such as lakes’ water level. The ecosystem of Urmia Lake is one of the most momentous ecosystems in Iran, which is almost close-ended and has become a global environmental issue in recent years. One of the parameters affecting this lake water level is snowfall, which has a key role in the fluctuations of its water level and water resources management. Hence, the purpose of this paper is the Urmia Lake water level estimation during 2000–2006 using observed water level, snow cover, direct precipitation, and evaporation. For this purpose, Support Vector Regression (SVR), which is the most outstanding kernel method (with various kernel types), has been used. Furthermore, four scenarios are considered with different variables as inputs, and the output of all scenarios is the water level of the lake. The results of training and testing data indicate the substantial impact of snow on retrieving the water level of the Urmia Lake at the desired period, and due to the complexity of the data relationships, the Gaussian kernel generally had better results. On the other hand, Quadratic and Cubic kernels did not work well. The fourth scenario, with RBF kernel has the best results [Training: R2 = 97% and RMSE = 0.09 m, Testing: R2 = 96.97% and RMSE = 0.08 m].

Magnetic and Radar Sensing for Multimodal Remote Health Monitoring

With the increased life expectancy and rise in health conditions related to aging, there is a need for new technologies that can routinely monitor vulnerable people, identify their daily pattern of activities and any anomaly or critical events such as falls. This paper aims to evaluate magnetic and radar sensors as suitable technologies for remote health monitoring purpose, both individually and fusing their information. After experiments and collecting data from 20 volunteers, numerical features has been extracted in both time and frequency domains. In order to analyze and verify the validation of fusion method for different classifiers, a support vector machine with a quadratic kernel, and an artificial neural network with one and multiple hidden layers have been implemented. Furthermore, for both classifiers, feature selection has been performed to obtain salient features. Using this technique along with fusion, both classifiers can detect 10 different activities with an accuracy rate of approximately 96%. In cases where the user is unknown to the classifier, an accuracy of approximately 92% is maintained.

Wind power prediction using bootstrap aggregating trees approach to enabling sustainable wind power integration in a smart grid

Precise prediction of wind power is important in sustainably integrating the wind power in a smart grid. The need for short-term predictions is increased with the increasing installed capacity. The main contribution of this work is adopting bagging ensembles of decision trees approach for wind power prediction. The choice of this regression approach is motivated by its ability to take advantage of many relatively weak single trees to reach a high prediction performance compared to single regressors. Moreover, it reduces the overall error and has the capacity to merge numerous models. The performance of bagged trees for predicting wind power has been compared to four commonly know prediction methods namely multivariate linear regression, support vector regression, principal component regression, and partial least squares regression. Real measurements recorded every ten minutes from an actual wind turbine are used to illustrate the prediction quality of the studied methods. Results showed that the bagged trees regression approach reached the highest prediction performance with a coefficient of determination of 0.982. The result showed that the bagged trees approach is followed by support vector regression with Gaussian kernel, the same model when using a quadratic kernel, and the multivariate linear regression, partial least squares, and principal component regression gave the lowest prediction. The investigated models in this study can represent a helpful tool for model-based anomaly detection in wind turbines.

3D Facial Action Units Recognition for Emotional Expression

The muscular activities caused the activation of facial action units (AUs) when a facial expression is shown by a human face. This paper presents the methods to recognize AU using a distance feature between facial points which activates the muscles. The seven AU involved are AU1, AU4, AU6, AU12, AU15, AU17 and AU25 that characterizes a happy and sad expression. The recognition is performed on each AU according to the rules defined based on the distance of each facial point. The facial distances chosen are computed from twelve salient facial points. Then the facial distances are trained using Support Vector Machine (SVM) and Neural Network (NN). Classification result using SVM is presented with several different SVM kernels while result using NN is presented for each training, validation and testing phase. By using any SVM kernels, it is consistent that AUs that are corresponded to sad expression has a high recognition compared to happy expression. The highest average kernel performance across AUs is 93%, scored by quadratic kernel. Best results for NN across AUs is for AU25 (Lips parted) with lowest CE (0.38%) and 0% incorrect classification.

Quadratic Vertex Kernel for Rainbow Matching

In this paper, we study the NP-complete colorful variant of the classic matching problem, namely, the Rainbow Matching problem. Given an edge-colored graph G and a positive integer k, the goal is to decide whether there exists a matching of size at least k such that the edges in the matching have distinct colors. Previously, in [MFCS’17], we studied this problem from the view point of Parameterized Complexity and gave efficient FPT algorithms as well as a quadratic kernel on paths. In this paper we design a quadratic vertex kernel for Rainbow Matching on general graphs; generalizing the earlier quadratic kernel on paths to general graphs. For our kernelization algorithm we combine a graph decomposition method with an application of expansion lemma.

Pattern Classification of Hand Movement Tremor in MS Patients with DBS ON and OFF.

Background: Hand tremor is one of the consequences of MS disease degrading quality of patient’s life. Recently DBS is used as a prominent treatment to reduce this effect. Evaluation of this approach has significant importance because of the prevalence rate of disease. Objective: The purpose of this study was the nonlinear analysis of tremor signal in order to evaluate the quantitative effect of DBS on reducing MS tremor and differentiating between them using pattern recognition algorithms.Material and Methods: In this analytical study, nine features were extracted from the tremor signal. Through statistical analysis, the significance level of each feature was examined. Finally, tremor signals were categorized by SVM, weighted KNN and NN classifiers. The performance of methods was compared with an ROC graph.Results: The results have demonstrated that dominant frequency, maximum amplitude and energy of the first IMF, deviation of the direct path, sample entropy and fuzzy entropy have the potential to create a significant difference between the tremor signals. The classification accuracy rate of tremor signals in three groups for Weighted KNN, NN and SVM with Gaussian and Quadratic kernels resulted in 95.1%, 93.2%, 91.3% and 88.3%, respectively.Conclusion: Generally, nonlinear and nonstationary analyses have a high potential for a quantitative and objective measure of MS tremor. Weighted KNN has shown the best performance of classification with the accuracy of more than 95%. It has been indicated that DBS has a positive influence on reducing the MS tremor. Therefore, DBS can be used in the objective improvement of tremor in MS patients.

Are Affine Volatility Restrictions Still Costly when the Pricing Kernel is Quadratic?

I compare the quadratic kernel used by Christoffersen, Heston and Jacobs (2013) with the commonly used Rubinstein's (1976) power pricing kernel in terms of option valuation performance. I do so in both affine and nonaffine GARCH(1,1) models. I find that, in both cases, the quadratic kernel outperforms its linear counterpart. I find no evidence that the performance gap between the affine and the nonaffine models shrinks with the quadratic kernel.

Interpolation Kernels for Synthetic Aperture Sonar Along-Track Motion Estimation

The formation of high-resolution synthetic aperture sonar (SAS) imagery requires accurate estimates of the sensor's trajectory. This is frequently accomplished using the displaced phase center antenna technique, which utilizes cross correlation of the signals received on successive pings. Accurate estimates of the sensor's ping-to-ping advance are then made by measuring the along-track spatial coherence of the scattered field. Unbiased advance-per-ping estimates require an accurate model for the spatial coherence of the scattered field. This model may be found by the application of the van Cittert-Zernike theorem to the problem of pulsed active sonar systems. In this paper, it is shown that the spatial coherence for a typical high-frequency SAS collection geometry is well approximated by a Gaussian whose width is proportional to the sensor's element size. Gaussian and quadratic along-track interpolation kernel performances are compared for a pair of at sea data collections. A relative image quality metric, based on image contrast, is defined to quantitatively assess the performance of the pair of interpolation kernels. In both tests, the use of an along-track estimator is shown to provide improved image quality. Also in both tests, the performance of the Gaussian kernel exceeds that of the quadratic kernel.

Machine Learning Approaches for Automated Lesion Detection in Microwave Breast Imaging Clinical Data

Breast lesion detection employing state of the art microwave systems provide a safe, non-ionizing technique that can differentiate healthy and non-healthy tissues by exploiting their dielectric properties. In this paper, a microwave apparatus for breast lesion detection is used to accumulate clinical data from subjects undergoing breast examinations at the Department of Diagnostic Imaging, Perugia Hospital, Perugia, Italy. This paper presents the first ever clinical demonstration and comparison of a microwave ultra-wideband (UWB) device augmented by machine learning with subjects who are simultaneously undergoing conventional breast examinations. Non-ionizing microwave signals are transmitted through the breast tissue and the scattering parameters (S-parameter) are received via a dedicated moving transmitting and receiving antenna set-up. The output of a parallel radiologist study for the same subjects, performed using conventional techniques, is taken to pre-process microwave data and create suitable data for the machine intelligence system. These data are used to train and investigate several suitable supervised machine learning algorithms nearest neighbour (NN), multi-layer perceptron (MLP) neural network, and support vector machine (SVM) to create an intelligent classification system towards supporting clinicians to recognise breasts with lesions. The results are rigorously analysed, validated through statistical measurements, and found the quadratic kernel of SVM can classify the breast data with 98% accuracy.