Resampling Methods

Box filter is simple and well known technique for image resizing. This technical paper describes an algorithm for image downscaling (shrinking) using box filter with subpixel accuracy, based on color space rescaling, which avoids oating point arithmetics. All operations in accompaniment C++ code are completely integer based, which signicantly improves image shrinking speed and quality.

Centrographic measures of spatial dispersion, such as the standard distance, provide a numerical value to summarize the radial scattering of a set of points around their centre of gravity or centroid. This paper develops a procedure to test for the statistical significance of differences in dispersion between two sets of phenomena intertwined in space. The significance test is implemented using a resampling randomization procedure based on the pooled locations from both sets to estimate the sampling distribution of their differences. Repeated thousands of times, that yields empirical frequency thresholds of the sampling distributions to assess the statistical significance of the observed differences. Case studies based on residential locations of lone-parent families and retired couples in the Quebec City Metropolitan Area illustrate the procedure. This paper shows how randomization procedures can be used to adapt classical tests to assess the statistical significance of differences between indices of spatial dispersion.

Les mesures centrographiques de dispersion spatiale, telles que la distance-type, fournissent une valeur numérique pour synthétiser la diffusion radiale d’un ensemble de points autour de leur centre de gravité ou barycentre. Cet article développe une procédure pour tester la signification statistique des différences de dispersion entre deux ensembles de phénomènes qui s’entrecroisent dans l’espace. Le test de signification utilise un procédé de rééchantillonnage par randomisation qui combine les localisations des deux ensembles pour estimer les distributions d’échantillonnage des différences. En répétant des milliers de fois, on obtient des seuils empiriques de fréquence dans les distributions d’échantillonnage afin d’établir la signification statistique des différences observées. Des études de cas fondées sur la localisation résidentielle des familles monoparentales et des couples retraités de la région métropolitaine de Québec illustrent la procédure. L’article montre comment des procédés de randomisation permettent d’adapter des tests classiques pour évaluer la signification statistique des différences entre les indices de dispersion spatiale.

'El Diario de Juárez' is a local newspaper in a city of 1.5 million Spanish-speaking inhabitants that publishes texts of which citizens read them on both a website and an RSS (Really Simple Syndication) service. This research applies natural-language-processing and machine-learning algorithms to the news provided by the RSS service in order to classify them based on whether they are about a traffic incident or not, with the final intention of notifying citizens where such accidents occur. The classification process explores the bag-of-words technique with five learners (Classification and Regression Tree (CART), Naïve Bayes, kNN, Random Forest, and Support Vector Machine (SVM)) on a class-imbalanced benchmark; this challenging issue is dealt with via five sampling algorithms: synthetic minority oversampling technique (SMOTE), borderline SMOTE, adaptive synthetic sampling, random oversampling, and random undersampling. Consequently, our final classifier reaches a sensitivity of 0.86 and an area under the precision-recall curve of 0.86, which is an acceptable performance when considering the complexity of analyzing unstructured texts in Spanish.

In today's world, due to the advancement of technology, predicting the students' performance is among the most beneficial and essential research topics. Data Mining is extremely helpful in the field of education, especially for analyzing students' performance. It is a fact that predicting the students' performance has become a severe challenge because of the imbalanced datasets in this field, and there is not any comparison among different resampling methods. This paper attempts to compare various resampling techniques such as Borderline SMOTE, Random Over Sampler, SMOTE, SMOTE-ENN, SVM-SMOTE, and SMOTE-Tomek to handle the imbalanced data problem while predicting students' performance using two different datasets. Moreover, the difference between multiclass and binary classification, and structures of the features are examined. To be able to check the performance of the resampling methods better in solving the imbalanced problem, this paper uses various machine learning classifiers including Random Forest, K-Nearest-Neighbor, Artificial Neural Network, XG-boost, Support Vector Machine (Radial Basis Function), Decision Tree, Logistic Regression, and Naïve Bayes. Furthermore, the Random hold-out and Shuffle 5-fold cross-validation methods are used as model validation techniques. The achieved results using different evaluation metrics indicate that fewer numbers of classes and nominal features will lead models to better performance. Also, classifiers do not perform well with imbalanced data, so solving this problem is necessary. The performance of classifiers is improved using balanced datasets. Additionally, the results of the Friedman test, which is a statistical significance test, confirm that the SVM-SMOTE is more efficient than the other resampling methods. Moreover, The Random Forest classifier has achieved the best result among all other models while using SVM-SMOTE as a resampling method. INDEX TERMS Classification, data mining, educational data mining, imbalanced data problem, machine learning, resampling methods, statistical analysis.

We aim at developing and improving the imbalanced business risk modeling via jointly using proper
evaluation criteria, resampling, cross-validation, classifier regularization, and ensembling techniques.
Area Under the Receiver Operating Characteristic Curve (AUC of ROC) is used for model comparison
based on 10-fold cross validation. Two undersampling strategies including random undersampling (RUS)
and cluster centroid undersampling (CCUS), as well as two oversampling methods including random
oversampling (ROS) and Synthetic Minority Oversampling Technique (SMOTE), are applied. Three highly
interpretable classifiers, including logistic regression without regularization (LR), L1-regularized LR
(L1LR), and decision tree (DT) are implemented. Two ensembling techniques, including Bagging and
Boosting, are applied on the DT classifier for further model improvement. The results show that, Boosting
on DT by using the oversampled data containing 50% positives via SMOTE is the optimal model and it can
achieve AUC, recall, and F1 score valued 0.8633, 0.9260, and 0.8907, respectively.

Con el avance tecnológico y de las comunicaciones, los cambios de hábitos en la forma de consumir, el crecimiento de la comercialización de productos en forma virtual y la aparición de distintas formas de pago, el fraude de tarjetas de crédito se ha vuelto una práctica en crecimiento. Para el caso de detección de fraudes en transacciones financieras se han propuesto varios algoritmos de Machine Learning supervisados y sin supervisión. A continuación presentamos una descripción de distintos algoritmos aplicables en WEKA para la detección de fraudes con tarjetas de crédito.

Proses pengolahan citra mulai dari pembesaran (enlargement), pengecilan (shrinking), perputaran (rotating), koreksi geometri, rektifikasi, registrasi citra, pergeseran (translasi) dan lainnya. Semua proses pengolahan citra digital tersebut menggunakan resampling yang berarti merubah nilai piksel raster asal (domain) menjadi suatu nilai piksel raster hasil (kodomain). Terdapat tiga teknik resampling yang umum digunakan dalam proses resampling yaitu tetangga terdekat, interpolasi bilinear dan konvolusi kubik. Ketiga teknik ini sering digunakan perangkat lunak pengolahan citra secara digital. Untuk resampling dengan metode tetangga terdekat tidak mengubah nilai piksel asli dari input raster tetapi hasilnya menjadi jaggy atau checkerboard (kotak-kotak). Sedangkan untuk teknik resampling bilinear dan konvolusi kubik menghasilkan citra lebih halus tetapi membutuhkan waktu lebih lama dibandingkan dengan tetangga terdekat. Waktu tersebut digunakan untuk komputasi lebih rumit dibanding teknik tetangga terdekat.

Refreshing and crunchy salad of cucumber and seawed marinated with dashi and sesame oil 15 POTATO & EGG SALAD Japanese-style creamy mashed potatoes mixed with egg, smoked beef, onions, cucumber and carrots 19 MARINATED TOMATOES AND ONIONS Healthy tomatoes and onions with Japanese vinaigrette dressing 19 TORI KAWA SALAD Mix of fresh lettuce, tomato, cucumber, carrot, and onion topped with crispy tori kawa Add the Midori dressing for an excellent flavor 28 CHICKEN KARAAGE SALAD Mix of fresh lettuce, tomato, cucumber, carrot, and onion topped with chicken karaage bites Add the Midori dressing for an excellent flavor 32 VEGETABLE SALAD Mix of fresh lettuce, tomato, cucumber, carrot, and onion All prices are subjected to prevailing government tax and service charge. Pictures are for illustration purpose only. All prices are subjected to prevailing government tax and service charge. Pictures are for illustration purpose only. *One refillable drink is good for 1 person. Additional fee will be charged if shared.

In today's world, due to the advancement of technology, predicting the students' performance is among the most beneficial and essential research topics. Data Mining is extremely helpful in the field of education, especially for analyzing students' performance. It is a fact that predicting the students' performance has become a severe challenge because of the imbalanced datasets in this field, and there is not any comparison among different resampling methods. This paper attempts to compare various resampling techniques such as Borderline SMOTE, Random Over Sampler, SMOTE, SMOTE-ENN, SVM-SMOTE, and SMOTE-Tomek to handle the imbalanced data problem while predicting students' performance using two different datasets. Moreover, the difference between multiclass and binary classification, and structures of the features are examined. To be able to check the performance of the resampling methods better in solving the imbalanced problem, this paper uses various machine learning classifiers including Random Forest, K-Nearest-Neighbor, Artificial Neural Network, XG-boost, Support Vector Machine (Radial Basis Function), Decision Tree, Logistic Regression, and Naïve Bayes. Furthermore, the Random hold-out and Shuffle 5-fold cross-validation methods are used as model validation techniques. The achieved results using different evaluation metrics indicate that fewer numbers of classes and nominal features will lead models to better performance. Also, classifiers do not perform well with imbalanced data, so solving this problem is necessary. The performance of classifiers is improved using balanced datasets. Additionally, the results of the Friedman test, which is a statistical significance test, confirm that the SVM-SMOTE is more efficient than the other resampling methods. Moreover, The Random Forest classifier has achieved the best result among all other models while using SVM-SMOTE as a resampling method.

Özet Parametrik testler evren dağılımına ilişkin bir takım varsyaımların karşılanmasını gerektirir. Bu varsayımların karşılanmadığı durumlarda araştırmacılar parametrik olmayan istatistiksel yöntemler kullanılır. Geleneksel para-metrik olmayan yöntemlerde sıra puanları ve sıra ortalamalarına dayalı işlemler gerçekleştirirken yeniden örnek-leme yöntemleri bu sürece farklı bir bakış açısı getirmiştir. En yaygın kullanılan yeniden örnekleme yöntemlerinin başında randomizasyon testleri gelir. Randomizasyon testlerinin temel mantığı orijinal örneklemden hesaplanan test istatistiğinin rastgele olarak oluşturulan örneklemlerdeki test istatistiği ile karşılaştırılmasına dayalıdır. Bu an-lamda kullanımı dünyada gitgide yaygınlaşan randomizasyon testlerine ilişkin yeterli sayıda kaynak olmaması, özel-likle Türkiye'de kullanımının çok sınırlı olması bu makalenin yazımına temel oluşturmuştur. Ayrıca randomizasyon testlerinin son yıllarda gündemde önemli bir yer tutan R progrlamla dili üzerinden örneklendirilerek açıklanması çalışmanın diğer bir önemli unsuru olarak düşünülmektedir. Bu çalışmada randomizasyon testlerinin R proglam-la dili ile örneklendirilerek açıklanması amaçlanmıştır. Bu bağlamda randomizasyon testlerinin temel kavramları açıklanmış akabinde sosyal bilimler alanında yaygun kullanıldığı düşünülen bağımsız ve tekrarlı örneklemler için t testi, bağınmsız gruplar için tek yönlü varyans analizi ve korelasyon analizi özelinde R kodlarından faydalanılarak örneklendirmelere gidilmiştir. Bu çalışma ile özellikle Türkiye'de randomizasyon testlerinin ve R progrlamlama dili-nin kullanımının yaygınlaştırılması beklenmektedir. Anahtar Kelimeler: yeniden örnekleme yöntemleri, randomizasyon restleri, permütasyon testleri, r proglamla-ma dili, parametrik olmayan testler, bootsrap Abstract Parametric tests such as t-test, ANOVA, etc. requires some assumptions about the distribution of the scores in the universe. If those assumptions are not met it is a good idea to compute non-parametric tests instead of para-metric tests. Traditional non-parametric tests such as Wilcoxon Sum of Ranks and Kruskal Wallis tests etc. focus on the sum of ranks and mean ranks to compare the group scores. On the other hand, resampling methods present a different point of view on this process. One of the mostly used resampling methods is the randomization test. The basic principles of randomization tests are comparing the original test statistic (t values, F values, r coefficient, etc.) to the test statistics derived from randomly generated samples. Although usage of randomization tests in the world is pervading day by day in Turkey it is very rarely used. This may be because of insufficient written source published in Turkey. Moreover, the R programming language has become very popular recently. So in this study, it is aimed to explain the computation process of randomization tests using R codes. In this study, at first, some basic concepts about randomization tests were presented. Then randomization tests were exemplified for independent samples t-test, repeated sample t-test, one-way analysis of variance (one way ANOVA) using R codes. It is hoped that this study guide and motivate researchers to use randomizations tests and r programming language in their research.

In order to estimate model parameters in multiple regression models, resampling methods of bootstrap and jackknife are used. Resampling methods are used as an alternative readjustment method to the least squares method (OLS) especially when assumptions belonging to error term in regression analysis are not met. Data used in the study are taken from 25 advertisements in Sahibinden.com website and the price of beetle car brand is accepted as dependent variable for multiple linear regression models. It is aimed that price variable taken is tried to be explained with the help of variables of fuel, case type, salesman, sunroof, wind shield, upholstery, age and engine size. When we examined the variables, it is seen that categorical variables are in question and dummy variable must be used. Firstly, model parameters of this obtained data are estimated using OLS and significances of parameters are tested, then, model parameters, significances of estimated parameters, coefficient of determination, and standard error of the model and % 90 confidence intervals are estimated using one of the resampling methods, bootstrap and jackknife method and results belonging to these three methods are compared. Also, generalization condition, which is to the population, of parameter estimation results belonging to explanatory variables used in this study are reviewed with the help of jackknife resampling method ve It has been seen that the salesman and upholstery independent variables have a considerable effect at a significance level of .10 on the dependent variable of price dependence of decision making (p < .10) and Jackknife have confirmed these generalization.

PMUs are the edifice of wide area monitoring and measurement system. The operation of PMU is essential to monitor and control the dynamic nature of the grid where voltage and current have constantly changing fundamental frequency due to change in load, generation imbalance, transients due to faults etc. Hence grid frequency is continuously varying about its nominal frequency therefore modeling of frequency adaptive PMU is significant. This paper aims to simulate a model of phasor measurement unit which helps to calculate phasors at off-nominal condition by adjusting sample values automatically using re-sampling filter, which is a more elegant than conventional anti-aliasing filter. The algorithm behind phasor estimation is DFT based and it is snuffed out in off-nominal condition, thus requiring the re-sampler. The performance of PMU is tested underwide variation of frequency to officiate the result.

The advance of DNA sequencing technology presents a significant bioinformatic challenges in a downstream analysis such as identification of single nucleotide polymorphism (SNP). SNP is the most abundant form of genetic marker and have been one of the most crucial researches in bioinformatics. SNP has been applied in wide area, but analysis of SNP in plants is very limited, as in cultivated soybean (Glycine max L.). This paper discusses the identification of SNP in cultivated soybean using Support Vector Machine (SVM). SVM is trained using positive and negative SNP. Previously, we performed a balancing positive and negative SNP with undersampling and oversampling to obtain training data. As a result, the model which is trained with balanced data has better performance than that with imbalanced data.

In many real-world binary classification tasks (e.g. detection of certain objects from images), an available dataset is imbalanced, i.e., it has much less representatives of a one class (a minor class), than of another. Generally, accurate prediction of the minor class is crucial but it's hard to achieve since there is not much information about the minor class. One approach to deal with this problem is to preliminarily resample the dataset, i.e., add new elements to the dataset or remove existing ones. Resampling can be done in various ways which raises the problem of choosing the most appropriate one. In this paper we experimentally investigate impact of resampling on classification accuracy, compare resampling methods and highlight key points and difficulties of resampling.

Data plays a key role in the design of expert and intelligent systems and therefore, data preprocessing appears to be a critical step to produce high-quality data and build accurate machine learning models. Over the past decades, increasing attention has been paid towards the issue of class imbalance and this is now a research hotspot in a variety of fields. Although the resampling methods, either by under-sampling the majority class or by over-sampling the minority class, stand among the most powerful techniques to face this problem, their strengths and weaknesses have typically been discussed based only on the class imbalance ratio. However, several questions remain open and need further exploration. For instance, the subtle differences in performance between the over-and under-sampling algorithms are still undercomprehended, and we hypothesize that they could be better explained by analyzing the inner structure of the data sets. Consequently, this paper attempts to investigate and illustrate the effects of the resampling methods on the inner structure of a data set by exploiting local neighborhood information, identifying the sample types in both classes and analyzing their distribution in each resampled set. Experimental results indicate that the resampling methods that produce the highest proportion of safe samples and the lowest proportion of unsafe samples correspond to those with the highest overall performance. The significance of this paper lies in the fact that our findings may contribute to gain a better understanding of how these techniques perform on class-imbalanced data and why over-sampling has been reported to be usually more efficient than under-sampling. The outcomes in this study may have impact on both research and practice in the design of expert and intelligent systems since a priori knowledge about the internal structure of the imbalanced data sets could be incorporated to the learning algorithms.

This work presents a system to perform autonomous landing of a small size fixed-wing Unmanned Aerial Vehicle (UAV) on a Fast Patrol Boat (FPB). We propose a ground-based vision system with the camera, image capture and processing equipment installed in the ship, thus reducing the UAV size, weight and power requirements. The system observes the UAV and computes the control commands to send to the UAV via radio. This approach makes it also possible to use standard UAVs equipped with commercial autopilots. The developed system uses the captured image as input and a Particle Filter (PF) structure to estimate the UAV trajectory. It is also used an Unscented Kalman Filter (UKF) for the translational motion filtering and an Unscented Bingham Filter (UBiF) for the rotational motion filtering. This filtering structure is reminiscent of the Unscented Particle Filter (UPF). The obtained tracking error is compatible with automatic landing requirements.

Detrended correlation analysis in the sample period 1958-2015 shows the expected logarithmic relationship between atmospheric CO2 concentration and surface temperature at an annual time scale. However, the correlation is found to be sensitive to small changes in endpoints that define the sample period. The extent of this instability is described and a resampling procedure is demonstrated for a robust estimation of the statistical significance of correlation under these conditions 1 .

Early Bronze Age society in Central Europe is often depicted as a society dominated by males. It is suggested that archaeological evidence is not always in agreement with this view. This study reports the analysis of mortuary variability at the Early Bronze cemetery Rebesˇovice in Mora- via (Czech Republic) that uses computer-intensive resampling techniques. The results indicate that there are only minor differences in the mor- tuary treatment of females and males. There is neither a qualitative restriction of body treatment and objects buried with males, nor a quantitative emphasis on burials of males. In fact, burials of females are equipped with a higher number of bronze artifacts. Funerals of females were at least as significant social events as the funerals of males. We argue that the model of a male dominated community that used this cemetery is un- warranted. These findings do not refute the existence of gender-specific activities and roles that may be valued differently, but instead suggest that both females and males had access to resources from which they could draw power.

Objectives: Neandertals and humans are closely related but differ noticeably in adult morphology. Previous work has been equivocal as to the contribution of postnatal growth and development to these differences. Due to disparate preservation, most analyses focus on specific anatomies, reconstructed fossils, or limited sample sizes. The objective of this research is to highlight the importance of postnatal growth in expressing Neandertal-human distinctions in the craniofacial skeleton, using a large and unreconstructed Neandertal sample.
Materials/Methods: A resampling approach is utilized to compare relative size change in 20 craniofacial dimen-sions between Neandertals (n = 42) and humans (n = 262). The large number of immature Neandertal sampleswithin and between dental stages provides the necessary variation to test for growth differences. Nested resampling using human-human comparisons assesses the likelihood of observing human-Neandertal growth differences under the null hypothesis of similar ontogenetic variation.
Results: Humans and Neandertals undergo comparable levels of overall size change. However, we identify growth differences for a number of traits, helping explain some of the unique features of this fossil taxon. Nested resampling shows it is unlikely that a Neandertal-like maturation would be observed in a random ontogenetic sample of humans.
Discussion: Growth during adolescence appears to be fundamental in the expression of some Neandertal anatomies. Neandertal upper facial and nasal breadths appear to have expanded rapidly after puberty to account for differences between preadolescents and adults, and Neandertals and humans. Mandibular growth differences may relate to anterior tooth use to process foods and paramastication during Neandertal maturation.

'El Diario de Juárez' is a local newspaper in a city of 1.5 million Spanish-speaking inhabitants that publishes texts of which citizens read them on both a website and an RSS (Really Simple Syndication) service. This research applies natural-language-processing and machine-learning algorithms to the news provided by the RSS service in order to classify them based on whether they are about a traffic incident or not, with the final intention of notifying citizens where such accidents occur. The classification process explores the bag-of-words technique with five learners (Classification and Regression Tree (CART), Naïve Bayes, kNN, Random Forest, and Support Vector Machine (SVM)) on a class-imbalanced benchmark; this challenging issue is dealt with via five sampling algorithms: synthetic minority oversampling technique (SMOTE), borderline SMOTE, adaptive synthetic sampling, random oversampling, and random undersampling. Consequently, our final classifier reaches a sensitivity of 0.86 and an area under the precision-recall curve of 0.86, which is an acceptable performance when considering the complexity of analyzing unstructured texts in Spanish.

Network intrusion problem poses serious threat to all users of computer systems due to the increase in size of modern networks and the convolution of large network traffic data. The problem exceeds the weighty limits of conventional technique of intrusion prevention system. Therefore, new solutions for efficient and effective intrusion detection with low false alarm rate are required. This research proposes a new methodology for intrusion detection by combining wrapper feature selection approach based on a genetic algorithm with Synthetic Minority Over Sampling (SMOTE) and Resample techniques for the balancing of the class distribution. The two selected traffic datasets (KDDCUP99 and NSL-KDD) were subjected to hybrid preprocessing of filtering technique, where SMOTE and Resample were used to recognize the oversampling of the minority samples in a bid to constructively increase the prediction accuracy of the minority class under the assumption that the overall distribution is unchang...

In many real-world binary classification tasks (e.g. detection of certain objects from images), an available dataset is imbalanced, i.e., it has much less representatives of a one class (a minor class), than of another. Generally, accurate prediction of the minor class is crucial but it&#39;s hard to achieve since there is not much information about the minor class. One approach to deal with this problem is to preliminarily resample the dataset, i.e., add new elements to the dataset or remove existing ones. Resampling can be done in various ways which raises the problem of choosing the most appropriate one. In this paper we experimentally investigate impact of resampling on classification accuracy, compare resampling methods and highlight key points and difficulties of resampling.

The effectiveness of a learning task depends on data complexity (class overlap, class imbalance, irrelevant features, etc.). When more than one complexity factor appears, two or more preprocessing techniques should be applied. Nevertheless, no much effort has been devoted to investigate the importance of the order in which they can be used. This paper focuses on the joint use of feature reduction and balancing techniques, and studies which could be the application order that leads to the best classification results. This analysis was made on a specific problem whose aim was to identify the melodic track given a MIDI file. Several experiments were performed from different imbalanced 38-dimensional training sets with many more accompaniment tracks than melodic tracks, and where features were aggregated without any correlation study. Results showed that the most effective combination was the ordered use of resampling and feature reduction techniques.

With the advancement in computer vision and image processing field and due to the high computing power of graphic devices, it is high time to think of an alternate sampling grid called hexagonal grid which exactly resembles our rods and cones in the photoreceptors of human retina(eyes). Hexagonal grid structure framework provides higher angular resolution, equidistant property and consistent connectivity. As the angular resolution is more in hexagonal grid, curved structures can be well represented in hexagonal grid. As there is no hexagonal capturing and display device, software simulation methods are used for hexagonal image processing before proceeding the edge detection in hexagonal framework. Edge detection algorithm is the cornerstone for many pre-processing task in image processing field This paper covers different edge detection algorithms like robert, sobel, prewitt, log and canny in hexagonal and rectangular grid and the performance is assessed based on the design metrics like PSNR (peak signal to noise ratio) and MSE (mean square error). For all edge detection algorithms, results shows that hexagonal structural framework outperforms rectangular grid framework

These days, telecommunications is very much needed in all areas of life. This condition has made the competition among the company is extremely tense. One strategic way to protect the company is to retain existing customers. The retention program as a scheme to retain customers must be implemented precisely and efficiently so that the company can maintain as many customers as possible. In this case, customer churn prediction holds an essential role. However, the existence of imbalanced data can increase prediction errors and create problems. Hence, in order to overcome the issue, this study combined the Synthetic Minority Oversampling Technique-Nominal Continuous (SMOTE-NC) with Classification and Regression Trees (CART). SMOTE-NC was applied to balance classes on training data, while CART formed a classification tree from those balanced data. Then, this classification tree created by CART algorithm had become the basis for predicting customer churn. The data used in this study are from https://community.ibm.com/, where the variables are related to customer demographics, customer contracts, usage history, and customer status of one of the telecom companies. Based on the analysis of these data, SMOTE-NC and CART combination succeeded in reducing errors in predicting customer churn, which also led recall value to increase by approximately 19%. Moreover, the accuracy generated from this combination method was still in a pretty good range of over 75%. Therefore, this study proposes an excellent way to improve the performance of churn prediction, especially in the telecommunications industry.

In many real-world binary classification tasks (e.g. detection of certain objects from images), an available dataset is imbalanced, i.e., it has much less representatives of a one class (a minor class), than of another. Generally, accurate prediction of the minor class is crucial but it's hard to achieve since there is not much information about the minor class. One approach to deal with this problem is to preliminarily resample the dataset, i.e., add new elements to the dataset or remove existing ones. Resampling can be done in various ways which raises the problem of choosing the most appropriate one. In this paper we experimentally investigate impact of resampling on classification accuracy, compare resampling methods and highlight key points and difficulties of resampling.

With the advancement in computer vision and image processing field and due to the high computing power of graphic devices, it is high time to think of an alternate sampling grid called hexagonal grid which exactly resembles our rods and cones in the photoreceptors of human retina(eyes). Hexagonal grid structure framework provides higher angular resolution, equidistant property and consistent connectivity. As the angular resolution is more in hexagonal grid, curved structures can be well represented in hexagonal grid. As there is no hexagonal capturing and display device, software simulation methods are used for hexagonal image processing before proceeding the edge detection in hexagonal framework. Edge detection algorithm is the cornerstone for many pre-processing task in image processing field This paper covers dif erent edge detection algorithms like robert, sobel, prewitt, log and canny in hexagonal and rectangular grid and the performance is assessed based on the design metrics like PSNR (peak signal to noise ratio) and MSE (mean square error). For all edge detection algorithms, results shows that hexagonal structural framework outperforms rectangular grid framework.

With the advancement in computer vision and image processing field and due to the high computing power of graphic devices, it is high time to think of an alternate sampling grid called hexagonal grid which exactly resembles our rods and cones in the photoreceptors of human retina(eyes). Hexagonal grid structure framework provides higher angular resolution, equidistant property and consistent connectivity. As the angular resolution is more in hexagonal grid, curved structures can be well represented in hexagonal grid. As there is no hexagonal capturing and display device, software simulation methods are used for hexagonal image processing before proceeding the edge detection in hexagonal framework. Edge detection algorithm is the cornerstone for many pre-processing task in image processing field This paper covers dif erent edge detection algorithms like robert, sobel, prewitt, log and canny in hexagonal and rectangular grid and the performance is assessed based on the design metrics like PSNR (peak signal to noise ratio) and MSE (mean square error). For all edge detection algorithms, results shows that hexagonal structural framework outperforms rectangular grid framework.

Resampling techniques, of both bootstrap and surrogate data type, are used in the evaluation of linear trend tests. Monte Carlo simulations were done for several distributions of correlated and independent residuals. In particular for AR(1) residuals, the discrimination of strong autocorrelation from linear trend is investigated with respect to the sample size. The overall results show that resampling reduces the type I and II errors of the trend tests. Following the guidelines suggested by the simulation results, we could find significant linear trend in the data of land air temperature and sea surface temperature.

The aim of this paper is to introduce a priori sensitive
resampling (PSR) based particle filter (PF) for
prospective use in dynamic state estimation towards
navigation of unmanned underwater vehicles (UUVs).
The proposed method targets a common pitfall of
conventional resampling based PFs, in the sense that
classical resampling is likelihood biased operation which
progressively leads to particle impoverishment and
ultimately degrades the estimation quality. The presented
method however generates a resampled population
balanced between the significant regions of both
likelihood and state transition prior. The algorithm is
tested with a simulated navigation scenario for UUVs
using simplified motion model. Results reveal that by
using only a small population size, PSR provides a lower
root mean square error (RMSE) of estimation in
comparison to that obtained with Extended Kalman Filter
(EKF) and classical resampling particle filter as well as
an Exquisite Resampling algorithm. The method is also
shown to be insensitive to significant simulated
measurement outliers.

The search for similarities in large data sets has a very important role in many scientific fields. It permits to classify several types of data without an explicit information about it. In many cases researchers use analysis methodologies such as clustering to classify data with respect to the patterns and conditions together. But in the last few years new analysis tool such as a biclustering were proposed and applied to the many specific problems. Biclustering algorithms permit not only to classify data with respect to selected conditions, ...