APPLYING M-ESTIMATES TO DEFINE START VALUE OF EXPONENTIAL AVERAGE IN BRAUN’S FORECAST MODEL OF ZERO LEVEL

В экономическом прогнозировании коротких временных рядов часто применяется модель Брауна нулевого порядка. К одной из проблем использования этой модели на первых шагах прогнозирования относится оценка начального значения экспоненциальной средней. Как правило, в качестве такой оценки используется простое среднее арифметическое значение первых уровней ряда, которое является неустойчивой статистической оценкой. Поэтому в данном исследовании предложено для оценки начального значения экспоненциальной средней использовать робастные М-оценки Тьюки, Хампеля, Хьюбера и Эндрюса. Цель исследования заключается в определении целесообразности применения М-оценок для определения начального значения экспоненциальной средней в модели Брауна при прогнозировании коротких временных рядов экономических показателей. В результате проведенного экспериментального исследования установлено: а) к наиболее значимым факторам, влияющим на точность прогноза с использованием модели Брауна, относятся вид временного ряда, значение постоянной сглаживания, отбраковка аномальных уровней и вид весов; б) вид оценки начального значения экспоненциальной средней и число итераций при вычислении М-оценки являются менее значимыми факторами (в связи с этим обоснована целесообразность применения одношаговых М-оценок); в) на начальных шагах прогнозирования при ограниченном количестве уровней временного ряда, когда невозможно достоверно определить вид ряда и когда отсутствуют основания для отбраковки аномальных уровней, предпочтительнее использовать модель Брауна с весами Вейда и определять начальное значение экспоненциальной средней на основе одношаговых робастных М-оценок (в остальных случаях целесообразно применять простое среднее арифметическое значение). In economic forecasting of short-term time series Braun’s model of zero level is often applied. One of issues of usage of this model from the very beginning of forecasting is estimation of start value of exponential average. As usual, simple arithmetic mean of first levels of series, used as such estimate, is volatile statistical estimate. That’s why in this investigation it’s suggested to use Tukey’s, Hampel’s, Huber’s and Andrews’ robust M-estimates for estimation of start value of exponential average. Purpose of research is definition of reasonability of M-estimates application to define start value of exponential average in Braun’s model during forecasting of short-term time series of economic indicators. The results of conducted experimental research are as follows: a) the most important factors, that have significant impact on forecast accuracy with usage of Braun’s model, are type of time series, value of smoothing constant, removal of abnormal levels and type of weights; b) type of estimate of start value of exponential average and quantity of iterations in process of calculation of M-estimate are less significant factors; c) consequently, reasonability of usage of one-step M-estimates is justified; d) on the first steps of forecasting with limited quantity of levels of time series, when it’s impossible to define with certainty type of series and when there is no reasons for removal of abnormal levels, it’s preferable to use Braun’s model with Wade’s weights and define start value of exponential average based on one-step robust M-estimates (in other cases it’s better to use simple arithmetical mean).

Asymptotic Distributions of M-Estimates for Parameters of Multivariate Time Series with Strong Mixing Property

The publication is devoted to studying asymptotic properties of statistical estimates of the distribution parameters u∈Rq of a multidimensional random stationary time series zt∈Rm, t∈ℤ satisfying the strong mixing conditions. We consider estimates u^nδ(z¯n), z¯n=(z1T,…,znT)T∈Rmn that provide in asymptotic n→∞ the maximum values for some objective functions Qn(z¯n;u), which have properties similar to the well-known property of local asymptotic normality. These estimates are constructed by solving the equations δn(z¯n;u)=0, where δn(z¯n;u) are arbitrary functions for which δn(z¯n;u)−gradhQn(z¯n;u+n−1/2h)→0(n→∞) in Pn,u(z¯n)-probability uniformly on u∈U, were U is compact in Rq. In many cases, the estimates u^nδ(z¯n) have the same asymptotic properties as well-known M-estimates defined by equations u^nQ(z¯n)=arg maxu∈UQn(z¯n;u) but often can be much simpler computationally. We consider an algorithmic method for constructing estimates u^nδ(z¯n), which is similar to the accumulation method first proposed by R. Fischer and rigorously developed by L. Le Cam. The main theoretical result of the article is the proof of the theorem, in which conditions of the asymptotic normality of estimates u^nδ(z¯n) are formulated, and the expression is proposed for their matrix of asymptotic mean-square deviations limn→∞nEn,u{(u^δ(z¯n)−u)(u^δ(z¯n)−u)T}.

Correlation of Metabolic Profiles of Plasma and Cerebrospinal Fluid of High-Grade Glioma Patients

This work compares the metabolic profiles of plasma and the cerebrospinal fluid (CSF) of the patients with high-grade (III and IV) gliomas and the conditionally healthy controls using the wide-range targeted screening of low molecular metabolites by HPLC-MS/MS. The obtained data were analyzed using robust linear regression with Huber’s M-estimates, and a number of metabolites with correlated content in plasma and CSF was identified. The statistical analysis shows a significant correlation of metabolite content in plasma and CSF samples for the majority of metabolites. Several metabolites were shown to have high correlation in the control samples, but not in the glioma patients. This can be due to the specific metabolic processes in the glioma patients or to the damaged integrity of blood-brain barrier. The results of our study may be useful for the understanding of molecular mechanisms underlying the development of gliomas, as well as for the search of potential biomarkers for the minimally invasive diagnostic procedures of gliomas.

Highly Efficient Robust and Stable M-Estimates of Location

This article is partially a review and partially a contribution. The classical two approaches to robustness, Huber’s minimax and Hampel’s based on influence functions, are reviewed with the accent on distribution classes of a non-neighborhood nature. Mainly, attention is paid to the minimax Huber’s M-estimates of location designed for the classes with bounded quantiles and Meshalkin-Shurygin’s stable M-estimates. The contribution is focused on the comparative performance evaluation study of these estimates, together with the classical robust M-estimates under the normal, double-exponential (Laplace), Cauchy, and contaminated normal (Tukey gross error) distributions. The obtained results are as follows: (i) under the normal, double-exponential, Cauchy, and heavily-contaminated normal distributions, the proposed robust minimax M-estimates outperform the classical Huber’s and Hampel’s M-estimates in asymptotic efficiency; (ii) in the case of heavy-tailed double-exponential and Cauchy distributions, the Meshalkin-Shurygin’s radical stable M-estimate also outperforms the classical robust M-estimates; (iii) for moderately contaminated normal, the classical robust estimates slightly outperform the proposed minimax M-estimates. Several directions of future works are enlisted.

Addressing the statistical analysis dilemma that exists when analyzing clinical trial results with full efficacy using the Kaplan Meier survival analysis method

The use of a Kaplan–Meier (K–M) survival time approach is generally considered appropriate to report antimalarial efficacy trials. However, when a treatment arm has 100% efficacy, confidence intervals may not be computed. Furthermore, methods that use probability rules to handle missing data for instance by multiple imputation, encounter perfect prediction problem when a treatment arm has full efficacy, in which case all imputed values are either treatment success or all imputed values are failures. The use of a survival K–M method addresses this imputation problem in estimating the efficacy estimates also referred to as cure rates. We discuss the statistical challenges and propose a potential way forward. The proposed approach includes the use of K–M estimates as the main measure of efficacy. Confidence intervals could be computed using the binomial exact method. p-Values for comparison of difference in efficacy between treatments can be estimated using Fisher’s exact test. We emphasize that when efficacy rates are not 100% in both groups, the K–M approach remains the main strategy of analysis considering its statistical robustness in handling missing data and confidence intervals can be computed under such scenarios.

Bootstrapping M-estimators in generalized autoregressive conditional heteroscedastic models

Summary We consider the weighted bootstrap approximation to the distribution of a class of M-estimators for the parameters of the generalized autoregressive conditional heteroscedastic model. We prove that the bootstrap distribution, given the data, is a consistent estimate in probability of the distribution of the M-estimator, which is asymptotically normal. We propose an algorithm for the computation of M-estimates which at the same time is useful for computing bootstrap replicates from the given data. Our simulation study indicates superior coverage rates for various weighted bootstrap schemes compared with the rates based on the normal approximation and existing bootstrap methods for the generalized autoregressive conditional heteroscedastic model, such as percentile $t$-subsampling schemes. Since some familiar bootstrap schemes are special cases of the weighted bootstrap, this paper thus provides a unified theory and algorithm for bootstrapping in generalized autoregressive conditional heteroscedastic models.