1,721,034 research outputs found
An introduction to model selection
No full textThis paper is an introduction to model selection intended for nonspecialists who have knowledge of the statistical concepts covered in a typical first (occasionally second) statistics course. The intention is to explain the ideas that generate frequentist methodology for model selection, for example the Akaike information criterion, bootstrap criteria, and cross-validation criteria. Bayesian methods, including the Bayesian information criterion, are also mentioned in the context of the framework outlined in the paper. The ideas are illustrated using an example in which observations are available for the entire population of interest. This enables us to examine and to measure effects that are usually invisible, because in practical applications only a sample From the population is observed. The problem of selection bias, a hazard of which one needs to be aware in the context of model selection; is also discussed. (C) 2000 Academic Press
A comparison of several time-series models for assessing the value at risk of shares
No full textThe objective of this investigation was to assess the suitability of some standard time-series models to perform a specific task in the context of recent change in banking regulations in Germany. The task is to estimate the value at risk (VaR) associated with financial assets on a daily basis. The procedure employed by the supervisory authorities to monitor whether a model used for this purpose adequately performs this task is outlined. Nine time-series models were investigated using share prices from the Frankfurt Stock exchange. The models were compared in terms of criteria that are derived from the new regulations. The results are reported. Copyright (C) 2001 John Wiley & Sons, Ltd
Estimating the spatial distribution in forest stands by counting small angles between nearest neighbours
No full textEstimating survival functions for the main tree species based on time series data from the forest condition survey in Rheinland-Pfalz, Germany
No full textThe expected increase of calamities due to climate change has generated a growing demand from forest planners for applicable approaches to quantify production risks. Survival functions which provide the probability that a forest stand exceeds a certain age could be particularly useful in this context. However, for Central Europe only a few investigations have derived site and tree species specific parameters for survival functions that are applicable in larger regions and for longer periods. The German National Forest Condition Survey collects time series data that are spatially representative and that cover a relative long time period. Using the data from Rheinland-Pfalz, we illustrate how survival analysis can be used to estimate the survival function. The Kaplan-Meier estimate for the main tree species (groups) reveals an increasing dropout probability with increasing age, which is typical for forest stands. As expected, spruce stands are most vulnerable against hazards, showing a survival probability of 73% at age 100. The Weibull distribution proves to be an adequate parametric model for describing the survival times of all tree species groups. Adding site factor variables in an accelerated failure time model, which was tested for spruce, reveals plausible estimates. However, the simultaneous treatment of different hazards turns out to be problematic. In order to avoid these limitations, and to improve the survival models based on forest condition surveys, we recommend that the causes of removal be identified as accurately as possible, and in greater detail, compared to the current instructions. The application of the methodology presented here for data from other Federal States is certainly worth considering
Measuring Vulnerability to Poverty Using Long-Term Panel Data
No full textMeasuring Vulnerability to Poverty Using Long-Term Panel Data Author & abstract Download & other version 16 References 4 Citations Related works & more Corrections Author Listed: Katja Landau (Georg-August-University Göttingen) Stephan Klasen (Georg-August-University Göttingen) Walter Zucchini (Georg-August-University Göttingen) Registered: Stephan Klasen Abstract We investigate the accuracy of ex ante assessments of vulnerability to income poverty using cross-sectional data and panel data. We use long-term panel data from Germany and apply di fferent regression models, based on household covariates and previous-year equivalence income, to classify a household as vulnerable or not. Predictive performance is assessed using the Receiver Operating Characteristics (ROC), which takes account of false positive as well as true positive rates. Estimates based on cross-sectional data are much less accurate than those based on panel data, but for Germany, the accuracy of vulnerability predictions is limited even when panel data are used. In part this low accuracy is due to low poverty incidence and high mobility in and out of poverty
Estimating survival functions for the main tree species based on time series data from the forest condition survey in Rheinland-Pfalz, Germany
No full textThe expected increase of calamities due to climate change has generated a growing demand from forest planners for applicable approaches to quantify production risks. Survival functions which provide the probability that a forest stand exceeds a certain age could be particularly useful in this context. However, for Central Europe only a few investigations have derived site and tree species specific parameters for survival functions that are applicable in larger regions and for longer periods. The German National Forest Condition Survey collects time series data that are spatially representative and that cover a relative long time period. Using the data from Rheinland-Pfalz, we illustrate how survival analysis can be used to estimate the survival function. The Kaplan-Meier estimate for the main tree species (groups) reveals an increasing dropout probability with increasing age, which is typical for forest stands. As expected, spruce stands are most vulnerable against hazards, showing a survival probability of 73% at age 100. The Weibull distribution proves to be an adequate parametric model for describing the survival times of all tree species groups. Adding site factor variables in an accelerated failure time model, which was tested for spruce, reveals plausible estimates. However, the simultaneous treatment of different hazards turns out to be problematic. In order to avoid these limitations, and to improve the survival models based on forest condition surveys, we recommend that the causes of removal be identified as accurately as possible, and in greater detail, compared to the current instructions. The application of the methodology presented here for data from other Federal States is certainly worth considering
A Web-based rainfall atlas for southern Africa
No full textWe describe the development of a Web-based rainfall atlas for southern Africa, a decision support system for the management of water resources. The rainfall atlas, which is accessible online at the URI http://134.76.173.220/ rainfall/index.hunl, was constructed in a number of phases over some 20 years. In the first phase, a 16 parameter model was developed, validated for representative sites, and then fitted to daily rainfall data from 2550 sites. Eight years later the estimates of the model parameters were updated, extended to 5070 sites, and interpolated on a grid of 1 minute of a degree of latitude and longitude over the entire region of southern Africa, namely South Africa, Lesotho and Swaziland. The method of kriging with external drift was used for the interpolation. The interpolated estimates were used to generate long artificial daily rainfall sequences at a spatial resolution of 1 minute of degree square. The sequences were used to compute rainfall-related statistics, such as percentiles of annual and monthly rainfall distributions, probabilities associated with droughts, and additional measures relating to the timing and intensity of rainfall. The final phase, information transfer, was the construction of the Atlas website, which offers online access to a wide range of rainfall-related statistics and some 5000 maps. To facilitate the computation of statistics that are not available in the database, the Atlas enables users to generate, and then import, artificial sequences online, for any grid point in southern Africa. These sequences are designed to mirror the properties of real rainfall records (seasonality, serial dependence, distributional properties, etc.) at the required grid point and can be used to compute quantities of interest empirically. Copyright (c) 2006 John Wiley & Sons, Ltd
Estimating the spatial distribution in forest stands by counting small angles between nearest neighbours
No full textMeasuring Vulnerability to Poverty Using Long-Term Panel Data
No full textMeasuring Vulnerability to Poverty Using Long-Term Panel Data Author & abstract Download & other version 16 References 4 Citations Related works & more Corrections Author Listed: Katja Landau (Georg-August-University Göttingen) Stephan Klasen (Georg-August-University Göttingen) Walter Zucchini (Georg-August-University Göttingen) Registered: Stephan Klasen Abstract We investigate the accuracy of ex ante assessments of vulnerability to income poverty using cross-sectional data and panel data. We use long-term panel data from Germany and apply di fferent regression models, based on household covariates and previous-year equivalence income, to classify a household as vulnerable or not. Predictive performance is assessed using the Receiver Operating Characteristics (ROC), which takes account of false positive as well as true positive rates. Estimates based on cross-sectional data are much less accurate than those based on panel data, but for Germany, the accuracy of vulnerability predictions is limited even when panel data are used. In part this low accuracy is due to low poverty incidence and high mobility in and out of poverty
A model for the diameter-height distribution in an uneven-aged beech forest and a method to assess the fit of such models
No full textThis paper illustrates the application of a mixture model to describe the bivariate diameter-height distribution of trees growing in a pul e, uneven-aged beech forest. A mixture of two bivariate normal distributions is considered but the methodology is applicable to mixtures of other distributions. The model was fitted to diameter-height observations for 1242 beech trees in the protected forest Dreyberg (Solling, Germany). A considerable advantage of the model, apart from the fact that it happens to fit this large data set unusually well, is that the individual parameters all have familiar interpretations. The bivariate Johnson S-BB distribution was also fitted to the data for the purpose of comparing the fits. A second issue discussed in this paper is concerned with the general question of assessing the fit of models for bivariate data. We show how a device called "pseudo-residual" enables one to investigate the fit of a bivariate model in new ways and in considerable detail. Attractive features of pseudo-residuals include the fact that they are not difficult to interpret; they can be: computed using generally available statistical software and, most important of all, they enable one to examine the fit of a model by means of simple graphs
- …
