Podcasts about glms

Our special guest, astrophysicist Rachel Losacca, explains the intricacies of galaxies, modeling, and the computational methods that unveil their mysteries. She shares stories about how advanced computational resources enable scientists to decode galaxy interactions over millions of years with true-to-life accuracy. Sid and Andrew discuss transferable practices for building resilient modeling systems. Prologue: Why it's important to bring stats back [00:00:03]Announcement from the American Statistical Association (ASA): Data Science Statement Updated to Include “ and AI”  Today's guest: Rachel Losacco [00:02:10]Rachel is an astrophysicist who's worked with major galaxy formation simulations for many years. She hails from Leiden (Lie-den) University and the University of Florida. As a Senior Data Scientist, she works on modeling road safety.  Defining complex systems through astrophysics [00:02:52]Discussion about origins and adoption of complex systemsDifficulties with complex systems: Nonlinearity, chaos and randomness, collective dynamics and hierarchy, and emergence.Complexities of nonlinear systems [00:08:20]Linear models (Least Squares, GLMs, SVMs) can be incredibly powerful but they cannot model all possible functions (e.g. a decision boundary of concentric circles)Non-linearity and how it exists in the natural worldChaos and randomness [00:11:30]Enter references to Jurassic Park and The Butterfly Effect“In universe simulations, a change to a single parameter can govern if entire galaxy clusters will ever form” - RachelCollective dynamics and hierarchy [00:15:45]Interactions between agents don't occur globally and often is mediated through effects that only happen on specific sub-scalesAdaptation: components of systems breaking out of linear relationships between inputs and outputs to better serve the function of the greater system   Emergence and complexity [00:23:36]New properties arise from the system that cannot be explained by the base rules governing the systemExamples in astrophysics [00:24:34]These difficulties are parts of solving previously impossible problemsConsider this lecture from IIT Delhi on Complex Systems to get a sense of what is required to study and formalize a complex system and its collective dynamics (https://www.youtube.com/watch?v=yJ39ppgJlf0)Consciousness and reasoning from a new point of view [00:31:45]Non-linearity, hierarchy, feedback loops, and emergence may be ways to study consciousness. The brain is a complex system that a simple set of rules cannot fully define.See: Brain modeling from scratch of C. Elgans What did you think? Let us know.Do you have a question or a discussion topic for the AI Fundamentalists? Connect with them to comment on your favorite topics: LinkedIn - Episode summaries, shares of cited articles, and more. YouTube - Was it something that we said? Good. Share your favorite quotes. Visit our page - see past episodes and submit your feedback! It continues to inspire future episodes.

TapirCast'in Popüler Konular ve Bilim Teknik serisinin bu bölümünde, Prof. Dr. Serhan Yarkan ve Sedanur Karabayram ile birlikte 2024 Türkiye Yerel Seçimleri'nde Yapay Zeka kullanımı üzerine konuşuyoruz. Bölümümüze, hızla gelişen teknolojiyle birlikte seçim süreçlerinin takibinin evrilip YouTube gibi dijital platformlar üzerinden gerçekleştirilmesini tartışarak başlıyoruz. Sonrasında, Cüneyt Özdemir'in seçim süreci yayınında, Yapay Zeka kullanımı ile sandıkların bir kısmının açılmasıyla elde edilen veri kullanılarak gerçekleştirilen gerçek zamanlı sonuç kestirimini sağlayan sistem üzerine konuşup bu sistemin geçmiş seçim sonuçlarından elde edilen veriler ile eğitilip test edilmesini ele alıyoruz. Daha sonra, seçim süreçlerinde sıkça gündeme gelen anket çalışmalarının arka planında yer alan yöntemlerei istatistiksel bakış açısıyla ele alıyoruz. Ardından, insanların sosyal medya hesaplarındaki davranışları üzerinden siyasi görüşlerinin kestirilip bu kestirimin genel seçim sonuçlarında kullanılmasını değerlendiriyoruz. Son olarak, seçim sonuçlarının kestiriminde kullanılan Yapay Zeka ve benzeri sistemlerin arka planında gerçekleştirilen araştırmaları ve geliştirilen teknolojileri açıklamak üzere generalized linear models (GLMs) ve ikili zaman serileri yaklaşımı (Binary Time Series Approach) kavramlarını ele alıp geliştirilen bu teknolojilerin geleceğini ve seçim süreçleri de dahil birçok farklı alanda nasıl kullanılacağını tartışarak bölümümüzü sonlandırıyoruz. Keyifli dinlemeler

In this week's episode Greg and Patrick take advantage of the recent expiration of a statute of limitations that legally allows them to talk about the multilevel model: what it is, when we might use it, and extremely cool extensions that it allows. Along the way they also discuss hostile federal judges, McNeish, airing of grievances, Gauss and Markov's corpses, Sesame Street, distributional baguettes, naivete, sentient GLMs, two pencil necks, Thor's Hammer, Willy Sutton, Siren's Song, peer groups of two, fighting good for an old guy, crazy town cool, 50 ducks, conceding a battle, and blushing corpses.Stay in contact with Quantitude! Twitter: @quantitudepod Web page: quantitudepod.org Merch: redbubble.com

This week on the PayneCast:- Chiefs Win!- Eric Beagles Resigns- GLMS wins NGAC- Ringgold/Columbus split, GL loses 2- Braves win 2 of 3 over the Orioles- UGA beats Tennessee 2 out of 3- Trevor Peek fights in Jacksonville on June 24 vs. Victor MartinezBe sure you leave us a review and a rating. You can follow us on Spotify, Apple Podcast, YouTube, Facebook, & Instagram! Please send your comments, topics, and ideas to thepaynecast1@gmail.com.

We speak with María Paula Fernández, Mattias Nystrom, and Stefan Adolf about the Golem Network, and the implications of creating a future internet based on distributed computing, with more focus on privacy.The Golem NetworkThe Golem Network is an accessible, reliable, open access and censorship-resistant protocol, which democratizes access to digital resources and connects users through a flexible, open source platform. As María Paula Fernández defines it, “At the highest level, Golem is a protocol for the exchange of digital resources. And not just digital resources like computing power, but also devices. It's a constellation of digital marketplaces where you can exchange with users across the globe.”To enable these exchanges, Golem works with a reference implementation called the Yagna daemon. Mattias explains the daemon's function, “The Yagna daemon helps you sort out your agreements as a requester, provider, or developer on the Golem Network.” Golem is anonymous and permissionless to join, so anyone can make use of the network and become a requester or provider. Golem uses Ethereum-based tokens, called GLMs, to enable the peer-to-peer exchanges.Golem solves problems with monopolies and deplatformingIf you need more computing power or storage, Golem can help — but the network also goes beyond those issues. The motivation for Golem is not to be at the mercy of corporations which often don't have your best interests in mind. María explains, “If the project you're running maybe doesn't fit with any of the tech giants like Google, AWS, Azure, Microsoft, or Instagram's terms and conditions, which can be really vague, you risk being taken down, and you lose your work. If you get deplatformed and that's your livelihood, that's a really big deal. That's why we need decentralized and censorship-resistant alternatives to the platforms that we have.”Try the Golem Network as a provider, requester, or developer. Platform.shLearn more about us.Get started with a free trial.Have a question? Get in touch!Platform.sh on social mediaTwitter @platformshTwitter (France): @platformsh_frLinkedIn: Platform.shLinkedIn (France): Platform.shFacebook: Platform.shWatch, listen, subscribe to the Platform.sh Deploy Friday podcast:YouTubeApple PodcastsBuzzsproutPlatform.sh is a robust, reliable hosting platform that gives development teams the tools to build and scale applications efficiently. Whether you run one or one thousand websites, you can focus on creating features and functionality with your favorite tech stack.

Podcast: Brain Inspired (LS 46 · TOP 1% what is this?)Episode: BI NMA 03: Stochastic Processes PanelPub date: 2021-07-22Panelists: Yael Niv.@yael_nivKonrad Kording@KordingLab.Previous BI episodes:BI 027 Ioana Marinescu & Konrad Kording: Causality in Quasi-Experiments.BI 014 Konrad Kording: Regulators, Mount Up!Sam Gershman.@gershbrain.Previous BI episodes:BI 095 Chris Summerfield and Sam Gershman: Neuro for AI?BI 028 Sam Gershman: Free Energy Principle & Human Machines.Tim Behrens.@behrenstim.Previous BI episodes:BI 035 Tim Behrens: Abstracting & Generalizing Knowledge, & Human Replay.BI 024 Tim Behrens: Cognitive Maps. This is the third in a series of panel discussions in collaboration with Neuromatch Academy, the online computational neuroscience summer school. In this episode, the panelists discuss their experiences with stochastic processes, including Bayes, decision-making, optimal control, reinforcement learning, and causality. The other panels: First panel, about model fitting, GLMs/machine learning, dimensionality reduction, and deep learning.Second panel, about linear systems, real neurons, and dynamic networks.Fourth panel, about basics in deep learning, including Linear deep learning, Pytorch, multi-layer-perceptrons, optimization, & regularization.Fifth panel, about “doing more with fewer parameters: Convnets, RNNs, attention & transformers, generative models (VAEs & GANs).Sixth panel, about advanced topics in deep learning: unsupervised & self-supervised learning, reinforcement learning, continual learning/causality.The podcast and artwork embedded on this page are from Paul Middlebrooks, which is the property of its owner and not affiliated with or endorsed by Listen Notes, Inc.

This is the first in a series of panel discussions in collaboration with Neuromatch Academy, the online computational neuroscience summer school. In this episode, the panelists discuss their experiences with model fitting, GLMs/machine learning, dimensionality reduction, and deep learning.

Comet provides a self-hosted and cloud-based meta machine learning platform allowing data scientists and teams to track, compare, explain and optimize experiments and models. Backed by thousands of users and multiple Fortune 100 companies, Comet provides insights and data to build better, more accurate AI models while improving productivity, collaboration and visibility across teams. Register for future sessions here: http://bit.ly/comet-ml-oh Checkout Comet ML by visiting: https://www.comet.ml/ Or on Twitter: https://twitter.com/CometML Connect with Ayodele LinkedIn: https://www.linkedin.com/in/ayodeleodubela/ Twitter: https://twitter.com/DataSciBae Check out her course on LinkedIn Learning: https://www.linkedin.com/learning/supervised-learning-essential-training/supervised-machine-learning-and-the-technology-boom [00:00:09] Friendly banter between the hosts [00:02:17] Did you learn anything new this week? [00:02:56] What is MLOps? [00:09:33] How MLOps is used in the finance industry [00:11:04] Topics to brush up on if you're looking to get into the finance space as a data scientist [00:12:06] Resources for learning about GLMs [00:15:00] The struggle of being a data scientist [00:17:19] What to do when you feel like there is so much to learn and not enough time [00:18:56] A few key things you should focus on when you're breaking into the field [00:22:01] The hardest part about SQL [00:25:59] What skills do I try to showcase in my portfolio project? [00:30:42] How am I supposed to gain business acumen when I don't have a job? [00:37:30] How do I get my profile noticed? [00:38:52] Understand how to develop KPIs and how your model impacts them [00:44:05] How would you split your time amongst different activities when doing a project? [00:51:17] There are multiple algorithms to use, how do I choose? [00:53:33] How to deal with these crazy job descriptions? [01:01:09] How do I position myself as a valuable candidate for a job? [01:07:19] Get people to do mock interviews with you [01:07:48] Convincing business stakeholders of your results when they want to follow their gut [01:15:17] Resources for picking up some nonobvious skills you need as a data scientist [01:16:13] Template methodology for problem framing: https://www.comet.ml/reports-templates/project-scope-template/reports/template/project-scope [01:18:14] Why you need to get as much hands on practice as you can

Singer / Songwriter the always talented Hollie Buhagiar joins us on the show this week to tell us all about her journey and how she got there.

We are back with a BRAND NEW episode of Totally Local! Aiming to bring you the very best in local music from the past and present.

Richard Gomila joins us on the show this week as we have an in-depth conversation on his career accompanied by five great tracks.

Link to bioRxiv paper: http://biorxiv.org/cgi/content/short/2020.09.01.276105v1?rss=1 Authors: Sarmashghi, M., Jadhav, S. P., Eden, U. Abstract: Point process generalized linear models (GLMs) provide a powerful tool for characterizing the coding properties of neural populations. Spline basis functions are often used in point process GLMs, when the relationship between the spiking and driving signals are nonlinear, but common choices for the structure of these spline bases often lead to loss of statistical power and numerical instability when the signals that influence spiking are bounded above or below. In particular, history dependent spike train models often suffer these issues at times immediately following a previous spike. This can make inferences related to refractoriness and bursting activity more challenging. Here, we propose a modified set of spline basis functions that assumes a flat derivative at the endpoints and show that this limits the uncertainty and numerical issues associated with cardinal splines. We illustrate the application of this modified basis to the problem of simultaneously estimating the place field and history dependent properties of a set of neurons from the CA1 region of rat hippocampus, and compare it with the other commonly used basis functions. We have made code available in MATLAB to implement spike train regression using these modified basis functions. Copy rights belong to original authors. Visit the link for more info

Bringing you the very best in local music from the past and present. Totally Local is sponsored by Chestertons Gibraltar.

Half Gibraltarian and Big Wednesday's own Caitlin Lavagna joins us on the show to tell us more about her band, their future and handpicks five great tracks :)

Link to bioRxiv paper: http://biorxiv.org/cgi/content/short/2020.07.23.218073v1?rss=1 Authors: Turra, G. L., Schneider, L., Liedgens, L., Deponte, M. Abstract: Leishmania parasites include important pathogens and model organisms and are even used for the production of recombinant proteins. However, functional genomics and the characterization of essential genes are often limited in Leishmania because of low-throughput technologies for gene disruption or tagging and the absence of components for RNA interference. Here, we tested the T7 RNA polymerase-dependent CRISPR-Cas9 system by Beneke et al. and the glmS ribozyme-based knock-down system in the model parasite Leishmania tarentolae. We successfully deleted two reference genes encoding the flagellar motility factor Pf16 and the salvage-pathway enzyme adenine phosphoribosyltransferase, resulting in immotile and drug-resistant parasites, respectively. In contrast, we were unable to disrupt the gene encoding the mitochondrial flavoprotein Erv. Cultivation of L. tarentolae in standard BHI medium resulted in a constitutive down-regulation of an episomal mCherry-glmS reporter by 40 to 60%. For inducible knock-downs, we evaluated the growth of L. tarentolae in alternative media and identified supplemented MEM, IMDM and McCoys 5A medium as candidates. Cultivation in supplemented MEM allowed an inducible, glucosamine concentration-dependent down-regulation of the episomal mCherry-glmS reporter by more than 70%. However, chromosomal glmS-tagging of the genes encoding Pf16, adenine phosphoribosyltransferase or Erv did not reveal a knock-down phenotype. Our data demonstrate the suitability of the CRISPR-Cas9 system for the disruption and tagging of genes in L. tarentolae as well as the limitations of the glmS system, which was restricted to moderate efficiencies for episomal knock-downs and caused no detectable phenotype for chromosomal knock-downs. Copy rights belong to original authors. Visit the link for more info

A great friend of ours here at the GLMS in Chris Cortes is our guest this week. It might be our longest episode YET but the anecdotes, stories as well as local music will be pleasing to your ears, we guarantee it.

A great friend of ours here at the GLMS in Chris Cortes is our guest this week. It might be our longest episode YET but the anecdotes, stories as well as local music will be pleasing to your ears, we guarantee it.

We're in for a treat as One Minute Silence's own Glen Diani joins us on SOG! Glen talks in-depth about those he was inspired by and to how it all happened with OMS!

Tyketto's own Danny Vaughn joins us EXCLUSIVELY on SOG this week and as always and like every other guest, he has carefully handpicked five great tracks from his career.

Another week, another SOG! We welcome Darrell Alman to the show and he takes us on his journey across the Gibraltar music scene and beyond. Like any other guest, he has handpicked five great local music tracks that mean a lot to him.

Elie Massias joins on the show and tells us more about his musical journey. He also handpicks five great tracks from his musical career.

After the great feedback on last week's episode, we decided to do another. This week Patrick Murphy joins us on the show and we talk all things local music and he also handpicks five great tracks.

Jamie Chiappe joins us on the show and we talk all things local music as well as handpicks five great tracks from the Gibraltar music scene.

Will Townes proposes a new, simpler way to analyze scRNA-seq data with unique molecular identifiers (UMIs). Observing that such data is not zero-inflated, Will has designed a PCA-like procedure inspired by generalized linear models (GLMs) that, unlike the standard PCA, takes into account statistical properties of the data and avoids spurious correlations (such as one or more of the top principal components being correlated with the number of non-zero gene counts). Also check out Will’s paper for a feature selection algorithm based on deviance, which we didn’t get a chance to discuss on the podcast. Links: Feature selection and dimension reduction for single-cell RNA-Seq based on a multinomial model (F. William Townes, Stephanie C. Hicks, Martin J. Aryee, Rafael A. Irizarry) GLM-PCA for R GLM-PCA for Python scry: an R package for feature selection by deviance (alternative to highly variable genes) Droplet scRNA-seq is not zero-inflated (Valentine Svensson)

We're back and with a great guest! Joining us this week is King Calaway's own Simon Dumas. Simon has had a great year and tells us more about it. We also take a trip down memory lane and look to the future.

We are back with BRAND NEW episodes of Straight Outta Gibraltar! Joining us this week on the season premiere is Thrifty Malone's own Paul 'Zooney' Alexander.

Bassist Jason Belilo is our guest this week! Listen in for some great anecdotes as well some great local music.

Formerly of Sympathy For The Blonde, John Barnett joins us this week!

We are back! Season 2 is upon us and we have GAMPA Principal Christian Santos joining us on the show for an in-depth chat about his career as well as some great music.

Nolan Frendo, Tristan Tonna, Aaron Ignacio and Justin Pou from Jetstream are our guests. Listen in for a lot of frolics, laughs, confessions as well as play SIX of their own tracks.

Afterhours' own Robert Perez is our guest and he chooses five KILLER tracks from the local music scene and more.

Local music legend Adrian Pisarello joins us on the pod this week and he tells us all about his adventures throughout the local music scene and more!

We wanted to keep to our format as to bringing you a BRAND NEW episode bi-weekly but we couldn't resist! Joining us this week is Reach's own Giles Ramirez! We discuss all aspects of Reach's BRAND NEW album Calling All Stations and EXCLUSIVELY play five tracks from the album too.

The Views' own Nigel Canepa joins us on the pod this week! He tells us all about being a part of the then Youth Movement that emerged onto the local music scene and more.

Welcome to the season premiere of Straight Outta Gibraltar! We are joined by Layla Bugeja who tells us all about her career thus far as well as her experiences.     

First ever GLMS signed Artist Jesse Sampere joins us on SOG this week! We discuss all aspects of his local music career thus far and play music from those performing this weekend at the London Gibraltar Friends Music Festival!

First ever GLMS signed Artist Jesse Sampere joins us on SOG this week! We discuss all aspects of his local music career thus far and play music from those performing this weekend at the London Gibraltar Friends Music Festival!

Superwookie's own Jonathan Sacramento is our guest this week for a fun chat covering his early days rehearsing at Bayside's Drama Studio alongside Jetstream's own Stu Whitwell and Murder of Crows' drummer Roin Sampere, his band from when he was studying abroad at Cardiff University called The Betty Ford Jazz Quartet, GLMS, opening up for Marillion in 2007 as well as his adventures when forming part of No Direction & SPF etc. Credit to Nikki Tacon

Many response variables are handled poorly by regression models when the errors are assumed to be normally distributed. For example, modeling the state damaged/not damaged of cells after treated with a certain chemical; and modeling the number of insects caught by a certain kind of trap. These types of situations can often be modeled well by a large class of regression models called generalized linear models (GLM). We will go over some of the basic statistical concepts of GLM and how it is relates to regression using normal errors. We will also go through some data analysis examples of GLMs in popular software such as R and SAS (possibly JMP if time allows) and explain how we interpret some of the output from each software. If time allows, the Bayesian approach to GLM will also be discussed. Course files available here: www.lisa.stat.vt.edu/?q=node/2321.

In many applications, the response variable is not Normally distributed. GLM can be used to analyze data from various non-Normal distributions. In this short course, we will introduce two most common GLM models: Logistic Regression for binary (yes/no or 0/1) data and Poisson Model for count data. Examples of fitting GLM models using JMP and interpretation of outputs are also provided. Course files available here: www.lisa.stat.vt.edu/?q=node/1177.

Die funktionelle Magnetresonanztomographie (fMRT) ist ein nicht-invasives Verfahren, das es ermöglicht, die Reizverarbeitung im Gehirn zu visualisieren. Für die statistische Analyse der Daten werden standardmäßig hypothesengeleitete Auswerteverfahren auf Basis des allgemeinen linearen Modells (general linear model, GLM) verwendet. Da der Signalanstieg in den Arealen sehr gering ist, muss, um in einem Experiment, das mit der Methode des GLMs ausgewertet werden soll, die Ruhe- und Aktivitätsbedingung mehrere Male wiederholt werden, um das erhaltene Signal besser vom Hintergrundrauschen trennen zu können. In bestimmten Fällen neuronaler Aktivität kann ein Stimulus nicht mehrere Male wiederholt werden, zum Beispiel nach Präsentation eines Reizes durch den ein sensorisches System für längere Zeit verändert wird (z.B. durch Adaptation). Die Methode der Independent Component Analysis (ICA) ist ein datengeleitetes Auswerteverfahren, das es ermöglicht Bilddaten so auszuwerten, dass a priori keine Hypothese über den Verlauf der Zeitkurve oder die Signalintensität bekannt sein muss. In dieser Arbeit konnte gezeigt werden, dass die ICA als datengeleitete Analysemethode reliable Ergebnisse sowohl im Bereich der trigeminalen Stimulation in fMRT Experimenten, als auch im relativ neuen Forschungsfeld der pharmakologischen Stimulation in der funktionellen Bildgebung liefert. Das Auswerteverfahren der ICA stellt somit eine vielversprechende Bereicherung der fMRT-Methodik dar.

Background: Body mass index (BMI) data usually have skewed distributions, for which common statistical modeling approaches such as simple linear or logistic regression have limitations. Methods: Different regression approaches to predict childhood BMI by goodness-of-fit measures and means of interpretation were compared including generalized linear models (GLMs), quantile regression and Generalized Additive Models for Location, Scale and Shape (GAMLSS). We analyzed data of 4967 children participating in the school entry health examination in Bavaria, Germany, from 2001 to 2002. TV watching, meal frequency, breastfeeding, smoking in pregnancy, maternal obesity, parental social class and weight gain in the first 2 years of life were considered as risk factors for obesity. Results: GAMLSS showed a much better fit regarding the estimation of risk factors effects on transformed and untransformed BMI data than common GLMs with respect to the generalized Akaike information criterion. In comparison with GAMLSS, quantile regression allowed for additional interpretation of prespecified distribution quantiles, such as quantiles referring to overweight or obesity. The variables TV watching, maternal BMI and weight gain in the first 2 years were directly, and meal frequency was inversely significantly associated with body composition in any model type examined. In contrast, smoking in pregnancy was not directly, and breastfeeding and parental social class were not inversely significantly associated with body composition in GLM models, but in GAMLSS and partly in quantile regression models. Risk factor specific BMI percentile curves could be estimated from GAMLSS and quantile regression models. Conclusion: GAMLSS and quantile regression seem to be more appropriate than common GLMs for risk factor modeling of BMI data.

Bartholomew Cubbins on RNA Episode 5: Three current RNA papers Watch the video(pops) A quick and dirty review of three current RNA papers:1. 14mer RNA aptamer targeting FMN.2. Gene regulation through nuclear retention.3. Ligand requirements of the glmS ribozyme.Enjoy,-BC

In this paper we model the life-history of LTC patients using a Markovian multi-state model in order to calculate premiums for a given LTC-plan. Instead of estimating the transition intensities in this model we use the approach suggested by Andersen et al. (2003) for a direct estimation of the transition probabilities. Based on the Aalen-Johansen estimator, an almost unbiased estimator for the transition matrix of a Markovian multi-state model, we calculate so-called pseudo-values, known from Jackknife methods. Further, we assume that the relationship between these pseudo-values and the covariates of our data are given by a GLM with the logit as link-function. Since the GLMs do not allow for correlation between successive observations we use instead the "Generalized Estimating Equations" (GEEs) to estimate the parameters of our regression model. The approach is illustrated using a representative sample from a German LTC portfolio.

Generalized linear models (GLMs) and semiparametric extensions provide a flexible framework for analyzing the claims process in non-life insurance. Currently, most applications are still based on traditional GLMs, where covariate effects are modelled in form of a linear predictor. However, these models may already be too restrictive if nonlinear effects of metrical covariates are present. Moreover, although data are often collected within longer time periods and come from different geographical regions, effects of space and time are usually totally neglected. We provide a Bayesian semiparametric approach, which allows to simultaneously incorporate effects of space, time and further covariates within a joint model. The method is applied to analyze costs of hospital treatment and accommodation for a large data set from a German health insurance company.

In this paper the data of a forest health inventory are analysed. Since 1983 the degree of defoliation (damage), together with various explanatory variables (covariates) concerning stand, site, soil and weather, are recorded by the second of the two authors, in the forest district Rothenbuch (Spessart, Bavaria). The focus is on the space and time dependencies of the data. The mutual relationship of space-time functions on the one side and the set of covariates on the other is worked out. To this end we employ generalized linear models (GLMs) for ordinal response variables and employ semiparametric estimation approaches and appropriate residual methods. It turns out that (i) the contribution of space-time functions is quantitatively comparable with that of the set of covariates, (ii) the data contain much more (timely and spatially) sequential structure than smooth space-time structure, (iii) a fine analysis of the individual sites in the area can be carried out with respect to predictive power of the covariates.