Search

Smog in China has been a major issue in recent years. As the main component of smog, PM 2.5 has received a lot of attention from the public, as the pollution caused by PM 2.5 is believed to have negative impacts in many aspects. Accordingly, modeling PM 2.5 concentrations has become research of interest for the sake of future prediction and control. Based on the data collected from monitoring stations nationwide, we find that PM 2.5 concentrations show both spatial and temporal dependency. In this paper, we adopt a grouped spatial-temporal model to depict the distribution of PM 2.5 , capturing the spatial heterogeneity in the patterns of PM 2.5 concentrations. An estimation method based on the EM algorithm is used since the group label assignment is unknown, and the results are further explained. Finally, an outlier detection procedure based on the grouped spatial-temporal model is presented. [ABSTRACT FROM AUTHOR]

The minimum covariance determinant (MCD) estimate is an important subsample method in robust statistics. It is commonly used to estimate parameters of multivariate normal distributions when outliers exist because of its good robustness properties. However, MCD has some disadvantages including low efficiency when there is no outliers and poor performance when there are clustered outliers. This paper first introduces an adaptive subsample method, and shows that the adaptive MCD estimator can possess both full asymptotic efficiency and maximum breakdown value. We then propose a minimum distance subsample estimate to handle the situations where there are clustered outliers. Simulation results indicate that the adaptive version of the minimum distance subsample estimate has satisfactory performance for various types of outliers. [ABSTRACT FROM AUTHOR]

Wavelets analysis has become a powerful mathematical tool to decompose a series and to provide its frequent and temporal features. Despite the fact that many researchers have begun testing this approach on financial and economic data to detect and treat outliers, it still not widely applied for longitudinal data. Therefore, this article proposes two algorithms for improving the accuracy of outlier detection in longitudinal data using Wavelets Decomposition, namely, Wavelets Decomposition for Outliers Detection and Handling across Subjects (WDODHAS), and Wavelets Decomposition for Outliers Detection and Handling within Subjects (WDODHWIS). The results show that Wavelets Decomposition is capable of detecting and handling outliers without erasing them, as well as highlighting hypothetical scenarios when these observations cannot be handled. [ABSTRACT FROM AUTHOR]

A two-stage joint survival model is used to analyze time to event outcomes that could be associated with biomakers that are repeatedly collected over time. A Two-stage joint survival model has limited model checking tools and is usually assessed using standard diagnostic tools for survival models. The diagnostic tools can be improved and implemented. Time-varying covariates in a two-stage joint survival model might contain outlying observations or subjects. In this study we used the variance shift outlier model (VSOM) to detect and down-weight outliers in the first stage of the two-stage joint survival model. This entails fitting a VSOM at the observation level and a VSOM at the subject level, and then fitting a combined VSOM for the identified outliers. The fitted values were then extracted from the combined VSOM which were then used as time-varying covariate in the extended Cox model. We illustrate this methodology on a dataset from a multi-center randomized clinical trial. A multi-center trial showed that a combined VSOM fits the data better than an extended Cox model. We noted that implementing a combined VSOM, when desired, has a better fit based on the fact that outliers are down-weighted. [ABSTRACT FROM AUTHOR]

Nonlinear models are commonly used for analyzing real-life data such as in medicine, engineering, and economics. To make efficient inferences about model parameter estimations and statistical results in nonlinear regression, assumptions related to error term are needed to be satisfied. Ordinary least squares and some modified least squares methods fail to give efficient parameter estimates when there are the problems of autocorrelation and outlier together in nonlinear regression. In this study, a novel hybrid robust tapering approach called as robust modified two-stage least squares is proposed to overcome the problems for obtaining more efficient parameter estimates in nonlinear regression. Two numerical examples and a comprehensive Monte–Carlo simulation study are given in order to examine the performance of robust modified two-stage least squares. [ABSTRACT FROM AUTHOR]

In recent years, clustering analysis for interval data has attracted the attention of many researchers. Nevertheless, an algorithm that can automatically determine the number of clusters, and can effectively detect the outlier intervals at the same time has not been studied so far. Therefore, in this paper, we propose a robust automatic clustering algorithm that only can automatically determine the number of clusters but also can assign the outlier intervals into separated clusters. The proposed algorithm is then applied in detecting the abnormal images consisting of the new image categories, and the images contaminated with noise. [ABSTRACT FROM AUTHOR]

The ridge regression estimator (RRE) is a widely used estimation method for the multiple linear regression model when the explanatory variables are correlated. The situation becomes problematic for the RRE when the data set contains outliers in the y-direction. The use of the RRE in the presence of outliers may have some adverse effects on parameter estimates. To address this issue, the robust ridge estimators based on M-estimator are available in the literature which are less sensitive to the presence of outliers. It is a well-known fact that the selection of ridge parameter k is very crucial while using the RRE and the same phenomenon may happen in the case of robust ridge estimators. This study proposes some robust ridge estimators for the ridge parameter k. The performance of proposed estimators is evaluated with the help of the Monte Carlo simulations and a real application where the mean squared error (MSE) is considered as a performance evaluation criterion. Results show a better performance of the proposed robust ridge estimators as compared to the RRE, least square and M-estimation methods. While for modified ridge M-estimator, different ridge parameters found to be better for different conditions. [ABSTRACT FROM AUTHOR]

The assumption of normality has a crucial importance in many statistical procedures. Therefore, a number of normality tests are proposed. Also, many investigations are conducted on the performance of these normality tests under a set of alternative distribution. However, there are few studies to compare the performance of the commonly used normality tests in the presence of outliers, but they are not comprehensive. It is important, since the outliers may increase the variability in the data set, they cause the decrease in the statistical power. In this study we show the performance of the commonly used normality tests in the presence of outlier in the data set. An outlier generation method is implied to generate uniform tails and the effects of various magnitude of outliers on the normality tests are obtained in terms of penalized power and Type I error probability. As a result, the most powerful tests are suggested to the researchers for different magnitude of outliers, sample sizes and the contamination ratios. Two real data applications are given to illustrate the importance of choosing the appropriate test. [ABSTRACT FROM AUTHOR]

The outlier detection task is of great importance for identifying spoiling observations in regression analysis. In this paper a novel strategy is proposed to determine outliers for regression models by using multivariate adaptive regression splines (MARS). Within this approach, it is seen that MARS is capable of capturing the clear observation set even when the data set contains numerous outliers. The performance of the proposed strategy is validated with simulation studies and three real data set applications. All the weak and strong aspects of the proposed strategy are evaluated with extensive discussions. [ABSTRACT FROM AUTHOR]

The Almon technique is a widely used estimation procedure for the distributed lag model (DLM) to encounter the problems associated with direct application of the ordinary least squares method to this model. The Almon estimator may be sensitive to outliers in the y-direction. This study is aimed to propose a robust estimator for parameter vector of the DLM when data set is contaminated with outliers. Moreover, this study proposes the robust t-tests and confidence intervals for the lag coefficients. Performance of our proposed estimator is evaluated through the Monte Carlo simulations by comparing the estimated mean squared error while the performance of t-tests and confidence intervals is evaluated using the null rejection rates and coverage, respectively. The simulation results reveal an attractive performance of the proposed methods in the presence of outliers in the y-direction. [ABSTRACT FROM AUTHOR]

In this article, we have concentrated on the influence diagnostics based on the mean shift outlier models (MSOM) and case deletion methods in the linear mixed measurement error models with Ridge estimation using the corrected likelihood of Nakamura (1990). We have derived the corrected score test statistic for outliers detection according to MSOM. Then, several case deletion measures are constructed as a tool for influence diagnostics. A parametric bootstrap procedure is used to obtain empirical distributions of the test statistics, and a simulation study has been used to show the performance of the proposed score test statistic and cook distance. Finally, a numerical example is given to illustrate the theoretical results. [ABSTRACT FROM AUTHOR]

In agricultural experiments, if outliers are present in a data set, inference of experiment may be reversed. The purpose of this article is to develop a method for detection of subset of outlier vectors in block designs for incomplete multiresponse experiments in presence of masking. We defined an influence matrix comprising of Cook-statistics in its diagonal and product of two Cook-statistics in its off-diagonal positions. We then obtained eigenvectors corresponding to large eigenvalues of this matrix which can be used for identification of influential subsets. The proposed procedure has been illustrated with a real life data set. [ABSTRACT FROM AUTHOR]

This paper presents a robust method for robust estimation of quadratic discriminant analysis. The mean and covariance matrix for estimating quadratic discriminant rule is computed using a robust estimation method called Sn method established from a robust covariance estimator S n C o v. The performance of the proposed method is evaluated using the results of simulated samples. This outlier detection method is compared with some well-known methods available in the current literature. The application of the proposed method in real-life data is also executed in this paper. [ABSTRACT FROM AUTHOR]

Confidence intervals for the population variance in the presence of outliers or spikes in the distribution of a variable of interest are topics that have not been investigated in depth previously. Results derived from a first Monte Carlo simulation study reveal the limitations of the customary confidence interval for the population variance when the underlying assumptions are violated, and the use of alternative confidence intervals is thus justified. We suggest confidence intervals based on the rescaled bootstrap method for many reasons. First, this is a simple technique that can be easily applied in practice. Second, it is free of probabilistic distributions. Finally, it can be easily applied to the cases of finite populations and samples selected from complex sampling designs. Results derived from a second Monte Carlo simulation study indicate that the suggested confidence intervals have desirable coverage rates with smaller average widths. Accordingly, an advantage of the suggested confidence intervals is that they offer a good compromise between simplicity and desirable properties. The various simulation studies are based on different scenarios that may arise in practice, such as the presence of outliers or spikes, and the fact that the underlying assumptions of the customary confidence interval are violated. [ABSTRACT FROM AUTHOR]

The histogram and boxplot are effective and simple graphical tools, which are broadly used to explore the characteristics of the distribution of univariate data. In this proposed work, two statistical plots called sample variance plots (Sv-plots), are defined which illustrate squared deviations from the sample variance. Sv-plots exhibit the contribution of each data value toward the sample variability with profound insight. Besides capturing symmetry and skewness of the distribution, these plots can detect outliers in the data based on two novel bounds introduced on each plot analogous to a boxplot. In comparison with histogram and boxplot, for actual and simulated data, Sv-plots firmly lead in revealing more information. Displaying the exact data values, Sv-plots improve visual sensation for detecting masked atypical values and asymmetry of the data. Remarkably, one version of the Sv-plots can be employed to display the testing hypotheses for a single and two population means and discovers a graphical method to make the decision. It is found that the Sv-plots outperform both histogram and boxplot, and have additional benefits beyond identifying characteristics of the distribution. [ABSTRACT FROM AUTHOR]

Varying-coefficient models (VCMs) are widely used in a variety of statistical applications. However, the classical VCMs based on least squares are prone to the presence of even a few severe outliers. In this article, a mean shift parameter is added for each observation to reflect outliers, and different penalties are then applied to the shift parameters to get sparse estimates. The jointly penalized optimization problem is solved through an efficient algorithm, and the tuning parameters are chosen by the Bayesian information criteria (BIC). The efficiency of the new approach is demonstrated via simulation studies as well as a real application on the Hong Kong environmental data. [ABSTRACT FROM AUTHOR]

This article presents a robust method for detecting multiple outliers from multidimensional data using robust Mahalanobis distance. Initial scatter matrix for robust Mahalanobis distance is constructed using a robust estimator of covariance ( S n C o v ) established from a robust scale estimator Sn and casewise median are chosen to be the location vector. The performance of the proposed method is evaluated using the results of simulated samples. This outlier detection method is compared with some well-known methods available in the current literature. The application of the proposed method in real-life data is also executed in this article. [ABSTRACT FROM AUTHOR]

Many clustering methods perform well with spherical clusters but poorly with elongated clusters. The Single-linkage method is suitable for finding such type of long clusters, but it can be sensitive to outliers and noise in the data, causing the so-called chain-effect. This work proposes a modification of Cross-Clustering algorithm, the SingleCross-Clustering (SCC), a partial clustering algorithm that estimates the number of clusters, recognizes outliers and that is useful for the identification of elongated clusters. SCC has been validated by comparing it with a number of existing clustering methods, showing on both simulated and real datasets that SCC is a reliable solution for the identification of the correct number of clusters and of the number of clusters memberships. The algorithm has been implemented in the R package CrossClustering, which can be downloaded for free from the CRAN contributed package repository. [ABSTRACT FROM AUTHOR]

Because the weighted total least squares (WTLS) method lacks robustness and is sensitive to gross errors, it cannot eliminate the influence of outliers effectively. A small number of gross errors may produce a devastating effect on estimates. Focusing on the limitation of the WTLS method regarding the influence of gross errors, this work combines Jackknife resampling theory with the WTLS algorithm for the identification and detection of outliers. The gross errors located in the WTLS method by the Jackknife method are identified to further improve the qualities of the estimated values if the observation data are falsified by outliers. This paper focuses on the following two aspects: only one gross error and multiple gross errors. Detailed calculation steps and the whole procedure for outlier detection using the new method are given. This algorithm is applied to the straight-line fitting model and the plane coordinate transformation model. From the experimental estimation results, we can see that the method proposed in this paper can identify gross errors that are greater than or equal to three times the standard error, and obtain more accurate estimation values when compared with the WTLS method and the classic robust weighted total least squares (RWTLS) method. The numerical case studies verify the effectiveness and practicality of the proposed procedure. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of robust parameter estimation of simple linear profiles in multistage processes in the presence of outliers in Phase I. In this regard, two robust approaches, namely the Huber's M-estimator and the MM estimator, are proposed to estimate the parameters of the process in Phase I in the presence of outliers in historical data. In addition, the U statistic is applied to the robust parameter estimates to remove the effect of the cascade property in multistage processes and as a result, to obtain adjusted robust estimates of the parameters of simple linear profiles. The performance of the proposed methods is evaluated under weak and strong autocorrelations involved in some numerical experimentation. The results show that the proposed robust methods perform efficiently to eliminate the effects of the outliers and the cascade property and demonstrate a better performance compared to the one of the classical approach under both weak and strong autocorrelations. Moreover, based on the results, the proposed MM estimator provides much better estimates of the parameters compared to the M estimator. [ABSTRACT FROM AUTHOR]

The R-package REPPlab is designed to explore multivariate data sets using one-dimensional unsupervised projection pursuit. It is useful as a preprocessing step to find clusters or as an outlier detection tool for multivariate data. Except from the packages tourr and rggobi, there is no implementation of exploratory projection pursuit tools available in R. REPPlab is an R interface for the Java program EPP-lab that implements four projection indices and three biologically inspired optimization algorithms. It also proposes new tools for plotting and combining the results and specific tools for outlier detection. The functionality of the package is illustrated through some simulations and using some real data. [ABSTRACT FROM AUTHOR]

A challenging problem in a linear regression model is to select a parsimonious model which is robust to the presence of contamination in the data. In this paper, we present a sparse linear approach which detects outliers by using a highly robust regression method. The model with optimal predictive ability as measured by the median absolute deviation of the prediction errors on JackKnife subsets is used to detect outliers. The performance of the proposed method is evaluated on a simulation study with a different type of outliers and high leverage points and also on a real data set. [ABSTRACT FROM AUTHOR]

Outliers presence may lead to misdetection on out-of-control observations in Phase II, therefore, they should be cleaned in Phase I. This paper proposes PCA Mix based T2 chart with Kernel Density control limit for mixed continuous and categorical data. Simulation studies are conducted to evaluate the performance of proposed chart in detecting outliers from clean and contaminated data. The proposed chart has better performance than the benchmark in monitoring clean data. For contaminated data, proposed chart has optimal performance in situation when categorical data are generated from multinomial distribution with balanced parameters. This is confirmed by simulated and real dataset. Compared to the conventional and other robust charts, the proposed chart demonstrated a great performance by success to detect more outlier correctly for the higher percentage of outlier added. [ABSTRACT FROM AUTHOR]

Partial Robust M (PRM) is a partial robust regression estimator using robust M-estimators. The aim of this study is to see the effects of using alternative Iteratively Reweighted Least Squares (IRLS) weight functions instead of Fair weight function used in original robust Partial Least Squares Regression (PLSR) method PRM, furthermore, is to examine the effects of soft, semi-hard and hard weightings on this algorithm in terms of efficiency, goodness of fit and prediction. Hence, classical SIMPLS, original PRM algorithm using Fair weight function and four alternative PRM algorithms named as PRMBSQR, PRMCHY, PRMHBR, PRMTLWRTH (obtained using Bisquare, Cauchy, Huber and Talworth weight functions) are compared. The simulation results and a real data application show the original PRM and PRMCHY, both of using soft weighting functions, are the leading algorithms in terms of efficiency and prediction for both low and high dimensional data sets in case of moderate outliers existence. However, when the proportion of outliers is getting higher, semi-hard weighting PRMBSQR or a hard weighting PRMTLWRTH could be good alternatives. Moreover, real data application showed that generally original PRM and PRMCHY, PRMBSQR and PRMTLWRTH algorithms have better performances in terms of outlier detection than both PRMHBR and classical SIMPLS algorithms. [ABSTRACT FROM AUTHOR]

Determining outliers is more complicated in multivariate data sets than it is in univariate cases. The aim of this study is to evaluate the blocked adaptive computationally efficient outlier nominators (BACON) algorithm, the fast minimum covariance determinant (FAST-MCD) method, and the robust Mahalanobis distance (RM) method in multivariate data sets. For this purpose, outlier detection methods were compared for multivariate normal, Laplace, and Cauchy distributions with different sample sizes and numbers of variables. False-negative and false-positive ratios were used to evaluate the methods' performance. The results of this work indicate that the performance of these methods varies according to the distribution type. [ABSTRACT FROM AUTHOR]

Cylindrical data are bivariate data from the combination of circular and linear variables. However, up to now no work has been done on the detection of outlier in cylindrical data. We introduce a definition of outlier for cylindrical data and present a new test of discordancy to detect outlier in this type of data, based on the k-nearest neighbor's distance. Cut-off points of the new test statistic based on the Johnson-Wehrly distribution are calculated and its performance is examined using simulation. A practical example is presented using wind speed and wind direction data obtained from the Malaysian Meteorological Department. [ABSTRACT FROM AUTHOR]

This article proposes a development of detecting patches of additive outliers in autoregressive time series models. The procedure improves the existing detection methods via Gibbs sampling. We combine the Bayesian method and the Kalman smoother to present some candidate models of outlier patches and the best model with the minimum Bayesian information criterion (BIC) is selected among them. We propose that this combined Bayesian and Kalman method (CBK) can reduce the masking and swamping effects about detecting patches of additive outliers. The correctness of the method is illustrated by simulated data and then by analyzing a real set of observations. [ABSTRACT FROM AUTHOR]

Outlier detection plays an important role in the pre-treatment of sequential datasets to obtain pure valuable data. This paper proposes an outlier detection scheme for dynamical sequential datasets. First, the conception of forward outlier factor(FOF) and backward outlier factor(BOF) are employed to measure an object's similarity shared with its sequentially adjacent objects. The object that shows no similarity with its sequential neighbors is labeled as suspicious outliers, which will be treated subsequently to judge whether it is really an outlier in the dataset. Second, the sequentially adjacent suspicious outliers are defined as suspicious outlier series(SOS), then the expected path representing the ideal transition path through the suspicious outliers in the SOS and the measured path representing the real path through all the objects in the SOS are employed, and the ratio of the length of the expected path to that of the measured path indicates whether there exist outliers in the SOS. Third, in the case that there exist outliers in the SOS, if there are N suspicious outliers in the SOS, then 2N − 2 remaining path will be generated by removing k(0 < k < N) suspicious outliers and sequentially connecting the remaining ones. The dynamical sequential outlier factor(DSOF) is employed to represent the ratio of the length of measured path of the considered remaining path to the that of the the expected path of the corresponding SOS, and the degree of the objects removed in a remaining path being outliers is indicated by the DSOF. The proposed outlier detection scheme is conducted from a dynamical perspective, and breaks the tight relation between being an outlier and being not similar with adjacent objects. Experiments are conducted to evaluate the effectiveness of the proposed scheme, and the experimental results verify that the proposed scheme has higher detection quality for sequential dataset. In addition, the proposed outlier detection scheme is not dependent on the size of dataset and needs no prior information about the distribution of the data. [ABSTRACT FROM AUTHOR]

Nonparametric charts are useful in statistical process controlwhen there is a lack of or limited knowledge about the underlying process distribution. Most existing approaches in the literature of Phase I monitoring assume that outliers have the same distributions as the in-control sample but only differ in location or scale parameters, they may not be effective with distributional changes. This article develops a new procedure based on the integration of the classical Anderson-Darling goodnessof- fit test and the stepwise isolation method. Our proposed procedure is efficient in detecting potential shifts in location, scale, or shape, and thus it offers robust protection against variation in various underlying distributions. The finite sample performance of our method is evaluated through simulations and is compared with that of available outlier detection methods for Phase I monitoring. [ABSTRACT FROM AUTHOR]

Several methods have been proposed to solve the one-class classification problem for vectors. Three methods are mainly used: density estimation, boundary methods, and reconstruction methods. The focus here is on boundary methods. These methods include the k-center method, the nearest neighbor method, one-class support vector machine (OCSVM), and the support vector data description (SVDD). In industrial applications, like statistical process control (SPC), practitioners successfully used SVDD to detect anomalies or outliers in the process. However, when a multivariate process shifts, it occurs in either location or scale. This far, most of the research effort, like the OCSVM or SVDD, has focused on location (vectors or mean vectors). Several methods have been proposed recently to monitor the scale, i.e., the covariance matrix. Most of these methods deal with a full rank covariance matrix, i.e., a situation where the number of rational subgroups is larger than the number of variables. When the number of variables is nearly as large as, or larger than, the number of observations, most Shewhart-type charts are unable to solve this problem as the estimated covariance matrix is not full rank. This work will extend the one-class classification method using kernels to detect changes in the covariance matrix when the number of observations available is less than the number of variables. [ABSTRACT FROM AUTHOR]

Distributions of a response y (height, for example) differ with values of a factor t (such as age). Given a response y* for a subject of unknown t*, the objective of inverse prediction is to infer the value of t* and to provide a defensible confidence set for it. Training data provide values of y observed on subjects at known values of t. Models relating the mean and variance of y to t can be formulated as mixed (fixed and random) models in terms of sets of functions of t, such as polynomial spline functions. A confidence set on t* can then be had as those hypothetical values of t for which y* is not detected as an outlier when compared to the model fit to the training data. With nonconstant variance, the p-values for these tests are approximate. This article describes how versatile models for this problem can be formulated in such a way that the computations can be accomplished with widely available software for mixed models, such as SAS PROC MIXED. Coverage probabilities of confidence sets on t* are illustrated in an example. [ABSTRACT FROM AUTHOR]

We present influence diagnostics for linear measurement error models with stochastic linear restrictions using the corrected likelihood of Nakamura in 1990. The case deletion and mean shift outlier models are developed to identify outlying and influential observations. We derive a corrected score test statistic for outlier detection based on mean shift outlier models. The analogs of Cook's distance and likelihood distance are proposed to determine influential observations based on case deletion models. A parametric bootstrap procedure is used to obtain empirical distributions of the test statistics and a simulation study has been used to evaluate the performance of the proposed estimators based on the mean squares error criterion and the score test statistic. Finally, a numerical example is given to illustrate the theoretical results. [ABSTRACT FROM AUTHOR]

Outlier detection has always been of interest for researchers and data miners. It has been well researched in different knowledge and application domains. This study aims at exploring the correctly identifying outliers using most commonly applied statistics. We evaluate the performance of AO, IO, LS, and TC as vulnerability to spurious outliers by means of empirical level of significance (ELS), power of the test indicating the sensitivity of the statistical tests in detecting changes and the vulnerability to masking of outliers in terms of misspecification frequencies are determined. We have observed that the sampling distribution of test statistic ηtp;tp=AO, IO, LS, TC  in case of AR(1) model is connected with the values of n and φ. The sampling distribution of  ηTC  is less concentrated than the sampling distribution of ηAO, ηIO, and ηLS. In AR(1) process, empirical critical values for 1%, 5%, and 10% upper percentiles are found to be higher than those generally used. We have also found the evidence that the test statistics for transient change (TC) needs to be revisited as the test statistics ηTCis found to be eclipsed by ηAO, ηLSand ηIOat different δ values. TC keeps on confusing with IO and AO, and at extreme δ values it just gets equal to AO and LS. [ABSTRACT FROM PUBLISHER]

Asymmetric models have been extensively studied in recent years, in situations where the normality assumption is not satisfied due to lack of symmetry of the data. Techniques for assessing the quality of fit and diagnostic analysis are important for model validation. This paper presents a study of the mean-shift method for detecting outliers in asymmetric normal regression models. Analytical solutions for the estimators of the parameters are obtained using the algorithm. Simulation studies and application to real data are presented, showing the efficiency of the method in detecting outliers. [ABSTRACT FROM AUTHOR]

The problem of outliers in statistical data has attracted many researchers for a long time. Consequently, numerous outlier detection methods have been proposed in the statistical literature. However, no consensus has emerged as to which method is uniformly better than the others or which one is recommended for use in practical situations. In this article, we perform an extensive comparative Monte Carlo simulation study to assess the performance of the multiple outlier detection methods that are either recently proposed or frequently cited in the outlier detection literature. Our simulation experiments include a wide variety of realistic and challenging regression scenarios. We give recommendations on which method is superior to others under what conditions. [ABSTRACT FROM AUTHOR]