Resistant Dimension Reduction
نویسندگان
چکیده
Existing dimension reduction (DR) methods such as ordinary least squares (OLS) and sliced inverse regression (SIR) often perform poorly in the presence of outliers. Ellipsoidal trimming can be used to create outlier resistant DR methods that can also give useful results when the assumption of linearly related predictors is violated. Theory for SIR and OLS is reviewed, and it is shown that several important hypothesis tests for an important class of regression models can be done using OLS output originally meant for multiple linear regression.
منابع مشابه
Nanomedicine for tuberculosis: Insights from animal models
Patient noncompliance to current tuberculosis (TB) therapy owing to multidrug administration daily leads to treatment failure and emergence of multidrug resistant and extensively drug resistant TB. To avoid the daily dosing, application of nanotechnology is the only viable solution by virtue of sustained release of drugs. Other potential advantages of the system include the possibility of selec...
متن کاملNanomedicine for tuberculosis: Insights from animal models
Patient noncompliance to current tuberculosis (TB) therapy owing to multidrug administration daily leads to treatment failure and emergence of multidrug resistant and extensively drug resistant TB. To avoid the daily dosing, application of nanotechnology is the only viable solution by virtue of sustained release of drugs. Other potential advantages of the system include the possibility of selec...
متن کاملOn Multivariate Methods in Robust Econometrics
This work studies implicitly weighted robust statistical methods suitable for econometric problems. We study robust estimation mainly for the context of heteroscedasticity or high dimension, which are up-to-date topics of current econometrics. We describe a modifi cation of linear regression resistant to heteroscedasticity and study its computational aspects. For a robust version of the instrum...
متن کاملApproximation of Functions over Manifolds: A Moving Least-Squares Approach
We present an algorithm for approximating a function defined over a d-dimensional manifold utilizing only noisy function values at locations sampled from the manifold with noise. To produce the approximation we do not require any knowledge regarding the manifold other than its dimension d. The approximation scheme is based upon the Manifold Moving Least-Squares (MMLS) presented in [25]. The pro...
متن کاملAnalysis of Censored Survival Data with Dimension Reduction Methods: Tehran Lipid and Glucose Study
Cardiovascular diseases (CVDs) are the leading cause of death worldwide. To specify an appropriate model to determine the risk of CVD and predict survival rate, users are required to specify a functional form which relates the outcome variables to the input ones. In this paper, we proposed a dimension reduction method using a general model, which includes many widely used survival m...
متن کامل