Questions tagged [robust]

Robustness in general refers to a statistic's insensitivity to deviations from its underlying assumptions (Huber and Ronchetti, 2009).

1
vote
1answer
39 views

Inference using robust statistics

I'm simulating the overall usage of a cluster using historic deployment data. Due to the nature of the simulation, there are some heavy points (i.e. very low overall usage). As a result, the variance ...
0
votes
0answers
17 views

Looking for a Regression Method with way to enfornce a reflection/ flip consistency for input [closed]

I have a set of $N$ dimensional 1D features using which I build a linear a regression model to predict a single scalar value. Say, $\hat{y}(w, x) = w_0 + w_1 x_1 + ... + w_p x_p$ with the regression ...
2
votes
1answer
16 views

Centering input data for Robust PCA (RPCA)?

I know that before running Principal component analys, the input data needs to be centered around its mean (subtract the mean from each keypoint) before running the algorithm. Do I need to center my ...
0
votes
0answers
15 views

Median point vs. number of points

I am trying to assign a binary label to a group of 2D points. Essentially I want to know whether or not the group lies mainly within a certain region in 2D space, such as above a line. I've been ...
0
votes
0answers
17 views

Is there a standard approach for estimating robust multinomial logit models?

I have been reading the "Robust Statistics" book by Morona, Martin and Yohai. To estimate a robust version of logistic regression, they recommend using redescending weighted $M$-estimator. For more ...
1
vote
1answer
15 views

Robust two-sample test with triplicate measurements?

When testing for a difference in mean between two conditions, biologists typically use a $t$-test, and wring their hands endlessly about how to justify removing outliers. Whereas I typically use a ...
4
votes
0answers
89 views

“Forward search” methods for outlier detection etc. in regression

I am reading Robust Diagnostic Regression Analysis by Anthony Atkinson and Marco Riani. They propose a robust "forward search" method for detecting outliers and other problematic data in regression (...
3
votes
0answers
26 views

Using robust regression to detect outliers

Rousseeuw and van Zomeren (1990) propose using robust regression to detect multivariate outliers, particularly in OLS regression. This approach seems to make sense (although I have not studied it in ...
0
votes
0answers
8 views

Placebo Thresholds in a Fuzzy Regression Discontinuity Design

Matsudaira (2008) used test scores to evaluate the impact of summer school attendance on students' performance. In this system, a student that scored less than an arbitrary grade was more likely to be ...
3
votes
0answers
37 views

Asymptotic exactness of Welch's t-test under arbitrary distributions

A common recommendation when using the Welch t-test for comparing two unpaired sample means is that the assumption of normality is not a problem when the sample size $n$ is bigger than some constant $...
0
votes
0answers
32 views

Doing bootstrapping to test the distribution means

Whenever I have two samples (A and B), I've been doing conventional t-test to compare the significant difference in means if I find that these two sample come from normal population (using normality ...
0
votes
2answers
41 views

Can one use parametric tests (e.g. ANOVA) if part of the data/ variables meet normality criteria but others not?

I want to compare differences between two groups(n1=26, n2=18) regarding their performance on different tests but also explore how each group performed on different times of these tests. I am ...
2
votes
0answers
13 views

F-test equivalent for residuals from robust regression

I'd like to compare fits for two datasets that have clearly non-normal errors. Is there a way to compare nested models fit with robust regression?
0
votes
0answers
22 views

Using t-dist (Excel) to compute p values for robust (rlm in R) regression coefficients

A few questions have been asked in the past in relation to the fact that the rlm robust regression function in the MASS package for R does not provide p values for regression coefficients. This answer ...
3
votes
1answer
78 views

The use of the quantile regression

I would like your help regarding the quantile regression. I was wondering if it makes sense to use the quantile regression when the relation of the number of data between variable x and y is 1 to 1, ...
4
votes
0answers
35 views

Can we estimate the mean of an asymmetric distribution in an unbiased and robust manner?

Suppose I have i.i.d. samples $X_1, \cdots, X_n$ from some unknown distribution $F$ and I wish to estimate the mean $\mu=\mu(F)$ of that distribution and I insist that the estimator be unbiased - i.e.,...
0
votes
1answer
45 views

What are the units of a medcouple?

Let's say I have a distribution of some weights in kg. I want to estimate the skewness of this distribution using a robust statistic, the medcouple [ G. Brys; M. Hubert; A. Struyf (2004). A robust ...
0
votes
1answer
23 views

How to report robust anovas (WRS2::t2way and t3way): no df

I have done experiments with a parasitic wasp species, comparing its life table parameters (response variables: longevity (in days), number of offspring and development time of offspring (in days)) ...
3
votes
2answers
49 views

Bayesian updating with multiple priors and multiple likelihood functions

$\Omega$ (finite) state space, $S$ (finite) signal space. Suppose we have a closed and convex set of priors $\mathcal{M}\subseteq \Delta(\Omega)$ such that each of them has full support. Let $\...
0
votes
0answers
50 views

SPSS ANCOVA with robust standard errors estimates

I am using SPSS version 25 to run an ANCOVA with heteroskedastic-consistent standard errors estimators (HC3 procedure) and I am puzzled by some of the output I get. I am fine with the robust standard ...
1
vote
1answer
177 views

Why do we use squared deviations to compute the SD, given that it amplifies the effect of outliers? [duplicate]

Suppose I have the following hypothetical data: One thousand times value 15 (i.e., 15 occurs 1000 times) and a single outlier value - 115 (i.e., 115 occurs just once - an outlier) Thus the mean is: $...
0
votes
0answers
29 views

Robust meta-analysis with robumeta package in R

I used the samples' mean ages as a moderating variable for a meta-analysis on the relationship between two variables. I used the robumeta package and I found the mean age in 20 of the 21 studies (the ...
3
votes
1answer
176 views

use and misuse of Winsorization

I am doing research on Winsorization (and trimming), which has been broadly applied in many fields, but I think many researchers didn't do it in a "rigorous" way. Or maybe even worse, they misuse it. ...
1
vote
1answer
65 views

What are some good robust loss functions for binary classification using LDA?

I am doing a project where I use LDA for binary classification. I want to know how it performs when there are outliers. What are some good robust loss functions for binary classification?
0
votes
0answers
29 views

Meaning of the low rank matrix in robust pca (rpca) particularly within a time series

I understand that RPCA tries to split a matrix into a low rank component and a sparse component. My understanding is that when we do normal PCA we reconstruct the data using fewer basis vectors than ...
0
votes
0answers
45 views

Neural Network Robust to Random Error in training set

I have read that deep learning algorithms are robust to random errors in the training set - is anyone able to explain why this is?
1
vote
0answers
24 views

Open implementation of Xu, Caramis and Mannor's outlier-robust PCA?

The answer linked below discusses an outlier-tolerant PCA method. Is there a publicly available implementation? https://stats.stackexchange.com/a/71928/86176 Here's the paper: Xu, H., Caramanis, C....
0
votes
0answers
14 views

Reference line for QQPlots: robust regression vs quartiles method

Using the car package, I drew two qqPlots of 144 p values from correlated tests on one dataset. The top plot uses a reference line drawn by "quartiles" and the bottom plot uses a line drawn by "robust"...
1
vote
0answers
43 views

What do they mean by Robust Cross-Validation?

I was reading a Kaggler Interview article and they kept specifying the importance of a stable and good cross-validation in order to win their competitions. What do they mean by that? I usually just ...
1
vote
0answers
75 views

R code for robust ridge regression

I am having trouble in searching for the MSE value in using robust ridge regression. The robust estimators that i used is LTS and MM. However, when both robust estimators were applied to ridge, the ...
0
votes
0answers
18 views

Robustness checks vs. BMA

I have a simple question. Let's say I have a model with 13 IVs and I am using a BMA for analysis. At the end, I would like to add few other variables for robustness checks. However, isn't BMA by ...
0
votes
0answers
29 views

M-estimator (S- and MM-) for regression models

Let's assume simple regression model $y_{i}=\beta_{0}+\beta_{1}x_{i}+\varepsilon_{i}$. It is obvious that the interpretation of the $\beta_{1}$ parameter is given by $\frac{\partial E(y_{i}|x_{i})}{\...
1
vote
1answer
66 views

Numerically Distinguish Between Real Correlation and Artifact

I'm looking at correlation for a large number of vectors, and many (about 3000) of these pairwise comparisons appear to have a significant correlation even after Bonferroni correction. Plotting these ...
1
vote
0answers
22 views

Discussion of Robustness

I have to include a discussion of robustness in my report. What are ways to discuss the robustness of my models? I run several logit models. But I don't know how to include something about the ...
0
votes
0answers
25 views

How do I weigh values by their relative frequency and then get an average?

Suppose I have three values that were measured repeatedly in a study. The conditions imply that the values should be close to each other, i.e. ideally three times the same value. The measured ...
1
vote
1answer
58 views

How to calculate the standard average of a set excluding outliers? [closed]

I have a set of numbers, and I need to calculate their average excluding outlier values (which I don't know a priori). It came to mind that many years ago I studied Standard Deviation. Could I apply ...
1
vote
1answer
117 views

Robust analogues of Mean, CV and Skewness

I need to characterize the mean, CV and skewness of my observed data (it is gamma-like distributed). This data is an artifact (outliers) enriched, so I decide to use robust statistics: median, ...
4
votes
1answer
45 views

If we change only one value of a data set, will the mean absolute deviation behave in the same way as standard deviation?

I took the new data as b and the data removed as a and calculated the new mean and used that to find the new mean and deviation in terms of the old. But it gets too complicated and there is no way to ...
3
votes
2answers
179 views

Why robust PCA results change with each run?

According to Filzmoser et al. 2009, the best way to conduct a principal component analysis for compositional data with outliers is: using a robust PCA method and using the isometric log ratio ...
1
vote
1answer
29 views

How to model which fixed effect is most responsible for variation in the DV?

Let's say I have data on firm revenue among multinational firms. I want to test the question of what explains more variation in revenue: firm culture or country-culture. A bit confused as to the ...
0
votes
0answers
33 views

Understanding MCD

I have recently stumpled upon the robust MCD (Minimal Covariance Determinant) Estimator. If I have $n$ datapoints of dimension $p$. Let`s say we want to obtain a robust estimate for the Covariance ...
2
votes
1answer
71 views

Tradeoffs of robust mean measures (trimmed, Huber, cosh, etc)

After recently having delved into the world of robust measures (for location, mean being the classical case), I have had difficulty understanding robust measures' core dynamic. Basically, what are ...
2
votes
0answers
127 views

Robust Regression in MATLAB's robustfit: what is the optimal weight function to tackle heteroskedasticity?

I'm currently performing a linear regression analysis and encountered a fair amount of heteroskedasticity. Increases in predicted values go along with decreases in residual variance. Otherwise, the ...
0
votes
0answers
19 views

Robust distance measure for correlated data

I read a paper in which the authors want to compare the overall predictive accuracy of various predictors on a set of variables by using the Mahalanobis-Distance. However the data is not even ...
1
vote
0answers
15 views

Deconvoluting an ECDF via mixed modeling

I have data with measurement error, $W_i$, with the following structure: $$W_i = \mu + \gamma_i + U_i$$ where $U_i \sim N(0, \sigma^2_i)$, with known $\sigma^2_i$, and $U_i \; \amalg \; \gamma_i$. I ...
4
votes
1answer
451 views

Robust covariance and OGK outlier detection

I'm calculating the robust covariance of a data set in order to use mahalanobis distance for outlier detection. There are few methods to calculate the covariance in the equation. Using the Fast-MCD ...
0
votes
1answer
42 views

Multiple comparison tests after using Robust method (lmrob)

I have some violation of assumptions (normality and equality of variances) is my analysis and I decided to a use robust technique (lmrob function in R). I have a continuous response, one categorical ...
1
vote
0answers
275 views

Robust regression - differences in approach (rlm and lmrob)?

I am looking to implement robust regression in R for large data (n=~500,000). The two options that come up are lmrob and rlm. ...
1
vote
0answers
105 views

Robust regression with Sandwich estimator

I understand that rlm (robust regression) addresses issues of outliers and influential observations, but does not address heteroskedasticity. I have come to learn ...
2
votes
0answers
169 views

Point Biserial Correlation

In terms of testing the assumption of normality for a Point Biserial Correlation; I have a dichotomous variable (1 = microsleep, 0 = no microsleep) and a continuous variable (number of driving events ...