Improving the detection of unusual observations in high‐dimensional settings

Multivariate control charts are used to monitor stochastic processes for changes and unusual observations. Hotelling's T² statistic is calculated for each new observation and an out‐of‐control signal is issued if it goes beyond the control limits. However, this classical approach becomes unreliable as the number of variables p approaches the number of observations n, and impossible when p exceeds n. In this paper, we devise an improvement to the monitoring procedure in high‐dimensional settings. We regularise the covariance matrix to estimate the baseline parameter and incorporate a leave‐one‐out re‐sampling approach to estimate the empirical distribution of future observations. An extensive simulation study demonstrates that the new method outperforms the classical Hotelling T² approach in power, and maintains appropriate false positive rates. We demonstrate the utility of the method using a set of quality control samples collected to monitor a gas chromatography–mass spectrometry apparatus over a period of 67 days.     

SPECIFICATION AND ESTIMATION OF LOGICALLY CONSISTENT LINEAR MODELS

The implications of constrained dependent and independent variables for model parameters are examined. In the context of linear model systems, it is shown that polyhedral constraints on the dependent variables will hold over the domain of the independent variables when a set of polyhedral constraints is satisfied by the model parameters. This result may be used in parameter estimation, in which case all predicted values of the dependent variables are consistent with constraints on the actual values. Also, the implicit constraints that define the set of parameters for many commonly used linear stochastic models with an error term yield values of the dependent variables consistent with the explicit constraints. Models possessing these properties are termed “logically consistent”.     

Ausgelernt? Professor/innen im Prozess der Professionalisierung

Do university teachers have finished to learn, when they are hired for a position as full professor? That seems to be false, if you look at knowledge enrichment by research as learning. For Teaching and management it seems more justified. Factually full professors very seldom participate at corresponding continuing education or consulting. Learning takes place mostly on the side as learning by doing. Professionalization in a full sense however is only realized, if competences are developed also in these fields. As corresponding learning can be supported by continuous education and consulting is a very important task of academic personnel development.     

Evaluation von Netzwerkentwicklungen in einer lernenden Region — ein Praxisbericht

The following contribution gives some suggestions on how a formative evaluation of a regional network of educational institutions and actors can be carried out. The formation of the network will be continuously supported by an iterative proceeding composed of the steps “data survey — feedback of results — consultation — transformation”. A central aim was to derive maximum benefit from linking together network evaluation and counselling. At the time of the first of three planned interview surveys the network was in a rather early stage of development. Central themes were role expectations, communicative and organizational structures, actors’ resources management and how a network is in general estimated as a learning opportunity.     

Esther Thelen

This study explored whether the reported inability of newborns to perceive object unity could result from the limited abilities of newborns to recognize the correspondence between 2 stimuli that were identical except for the presence or absence of an occluder. Five experiments were carried out using a visual habituation technique. The results of Experiment 1 indicated that newborns were capable of recognizing the perceptual correspondence between a nonoccluded and an occluded form. More intriguing, the outcomes of Experiments 2, 3, and 4 suggested that newborns find a partly occluded form to be more similar to a completely unoccluded form than to an unoccluded form containing a gap. Finally, even if newborns are able to perceive the correspondence between a partly occluded object and a complete form, the presence of this skill does not appear sufficient to imply the ability to manifest veridical object unity perception at birth (Experiment 5).     

Time-discrete nonparametric hazard model using panel data

In this contribution a nonparametric estimator for the hazard function will be presented for time-discrete survival analysis. The estimator is derived from a likelihood function based upon time-discrete counting processes. With martingale techniques asymptotic properties of the estimator of the cumulative hazard function are shown. Since we consider a nonparametric approach no exploratory variables are considered in the empirical example. For analyzing the remigrant behavior of different foreign nations (Italy, Yugoslavia, Greece, Spain and Turkey) the Socio-Economic Panel (SOEP) is used as a data basis. The estimations are carried out with a module of PRODISA, a program package developed for the analysis of time-discrete duration and panel data for the nonparametric and (semi)parametric case.     

Gruppen leiten und beraten

Coaching and leading of groups is on one hand a basal competence in higher education, on the other hand it makes high demands to academic developers, who as ??new professionals?? support the development of this competence. The following contribution is concerned with a continuous education for professional academic developers, which places with this objective coaching as consultation format and methodology for leading groups in the center. This program, conceptualized by the authors, is the first of its kind in Germany. It started at the center for higher education of the Technical University Dortmund in April 2011 and will be completed in June 2012.     

Regularised Manova for High‐Dimensional Data

The traditional and readily available multivariate analysis of variance (MANOVA) tests such as Wilks' Lambda and the Pillai–Bartlett trace start to suffer from low power as the number of variables approaches the sample size. Moreover, when the number of variables exceeds the number of available observations, these statistics are not available for use. Ridge regularisation of the covariance matrix has been proposed to allow the use of MANOVA in high‐dimensional situations and to increase its power when the sample size approaches the number of variables. In this paper two forms of ridge regression are compared to each other and to a novel approach based on lasso regularisation, as well as to more traditional approaches based on principal components and the Moore‐Penrose generalised inverse. The performance of the different methods is explored via an extensive simulation study. All the regularised methods perform well; the best method varies across the different scenarios, with margins of victory being relatively modest. We examine a data set of soil compaction profiles at various positions relative to a ridgetop, and illustrate how our results can be used to inform the selection of a regularisation method.