Cubic B-spline and Generalised Linear Models for COVID-19 Patients in Thailand

Large-scale climate processes such as the El Nino Southern Oscillation (ENSO), Indian Ocean Dipole (IOD) and Southern Annular Mode (SAM) influence the hydro-climatology of Southeast Australia (SEA). In the present study, we show that low-flow events in many catchments in SEA are significantly influenced by variability in these climate drivers. Extreme value distributions and Generalised Linear Models (GLMs) are used here to model low-flow characteristics such as intensity, duration and frequency with respect to these climate drivers. Further, we study how the future projections of ENSO, IOD and SAM are likely to evolve under climate change by examining the projected values of their representative indices and how they will impact low-flow events in the region. It is found that the future dry phases of these climate drivers are likely to be more dry than those in the historic period. This in turn is expected to lead to intensification of low-flow events in the future, resulting in lower availability of fresh water during occurrences of the dry phases of these climate drivers. Thus, climate change in the future is expected to significantly influence future low-flow events in the region thereby making it even more crucial for water managers to adequately manage and ensure water availability. Keywords: low-flows, ENSO, IOD, SAM, Extreme Value Theory, Generalised Linear Models, Southeast Australia, CMIP5, RCP8.5.

Download Full-text

Integral projection models

10.1093/oso/9780198838609.003.0010 ◽

2021 ◽

pp. 181-196

Author(s):

Edgar J. González ◽

Dylan Z. Childs ◽

Pedro F. Quintana-Ascencio ◽

Roberto Salguero-Gómez

Keyword(s):

Population Dynamics ◽

Environmental Variables ◽

Linear Models ◽

Vital Rates ◽

Matrix Population Models ◽

Generalised Linear Models ◽

Integral Projection Models ◽

Integral Projection ◽

Over Time

Integral projection models (IPMs) allow projecting the behaviour of a population over time using information on the vital processes of individuals, their state, and that of the environment they inhabit. As with matrix population models (MPMs), time is treated as a discrete variable, but in IPMs, state and environmental variables are continuous and are related to the vital rates via generalised linear models. Vital rates in turn integrate into the population dynamics in a mechanistic way. This chapter provides a brief description of the logic behind IPMs and their construction, and, because they share many of the analyses developed for MPMs, it only emphasises how perturbation analyses can be performed with respect to different model elements. The chapter exemplifies the construction of a simple and a more complex IPM structure with an animal and a plant case study, respectively. Finally, inverse modelling in IPMs is presented, a method that allows population projection when some vital rates are not observed.

Download Full-text

Generalised Linear Models For Continuous Data

Introducing Quantitative Geography ◽

10.4324/9780203973356-17 ◽

2005 ◽

pp. 200-246

Keyword(s):

Linear Models ◽

Continuous Data ◽

Generalised Linear Models

Download Full-text

Lifetime, 5-year and past-year prevalence of homelessness in Europe: a cross-national survey in eight European nations

BMJ Open ◽

10.1136/bmjopen-2019-033237 ◽

2019 ◽

Vol 9 (11) ◽

pp. e033237 ◽

Cited By ~ 1

Author(s):

Owen Taylor ◽

Sandrine Loubiere ◽

Aurelie Tinland ◽

Maria Vargas-Moniz ◽

Freek Spinnewijn ◽

...

Keyword(s):

Service Use ◽

Linear Models ◽

Pairwise Comparison ◽

Response Rates ◽

Computer Assisted ◽

Generalised Linear Models ◽

The Past ◽

History Of ◽

Nationally Representative ◽

Difference Test

ObjectivesTo examine the lifetime, 5-year and past-year prevalence of homelessness among European citizens in eight European nations.DesignA nationally representative telephone survey using trained bilingual interviewers and computer-assisted telephone interview software.SettingThe study was conducted in France, Ireland, Italy, the Netherlands, Poland, Portugal, Spain and Sweden.ParticipantsEuropean adult citizens, selected from opt-in panels from March to December 2017. Total desired sample size was 5600, with 700 per country. Expected response rates of approximately 30% led to initial sample sizes of 2500 per country.Main outcome measuresHistory of homelessness was assessed for lifetime, past 5 years and past year. Sociodemographic data were collected to assess correlates of homelessness prevalence using generalised linear models for clustered and weighted samples.ResultsResponse rates ranged from 30.4% to 33.5% (n=5631). Homelessness prevalence was 4.96% for lifetime (95% CI 4.39% to 5.59%), 1.92% in the past 5 years (95% CI 1.57% to 2.33%) and 0.71% for the past year (95% CI 0.51% to 0.98%) and varied significantly between countries (pairwise comparison difference test, p<0.0001). Time spent homeless ranged between less than a week (21%) and more than a year (18%), with high contrasts between countries (p<0.0001). Male gender, age 45–54, lower secondary education, single status, unemployment and an urban environment were all independently strongly associated with lifetime homelessness (all OR >1.5).ConclusionsThe prevalence of homelessness among the surveyed nations is significantly higher than might be expected from point-in-time and homeless service use statistics. There was substantial variation in estimated prevalence across the eight nations. Coupled with the well-established health impacts of homelessness, medical professionals need to be aware of the increased health risks of those with experience of homelessness. These findings support policies aiming to improve health services for people exposed to homelessness.

Download Full-text

Redefining the deviance objective for generalised linear models

British Actuarial Journal ◽

10.1017/s1357321712000190 ◽

2012 ◽

Vol 17 (3) ◽

pp. 491-509

Author(s):

A. C. Lovick ◽

P. K. W. Lee

Keyword(s):

Linear Models ◽

Patent Application ◽

Generalised Linear Models ◽

Number Of Factors ◽

Efficient Calculation ◽

Unseen Data ◽

Data Points ◽

New Material ◽

Price Comparison Sites ◽

Generalised Linear Modelling

AbstractThis paper defines the ‘Case Deleted’ Deviance - a new objective function for evaluating Generalised Linear Models, and applies this to a number of practical examples in the pricing of general insurance. The paper details practical approximations to enable the efficient calculation of the objective, and derives modifications to the standard Generalised Linear Modelling algorithm to allow the derivation of scaled parameters from this measure to reduce potential over fitting to historical data. These scaled parameters improve the predictiveness of the model when applied to previously unseen data points, the most likely being related to future business written. The potential for over fitting has increased due to number of factors now used, particularly in pricing personal lines business and the advent of price comparison sites which has increased the penalties of mis-estimation. New material in this paper has been included in a UK patent application No. 1020091.3.

Download Full-text

Multiscale generalised linear models for nonparametric function estimation

Biometrika ◽

10.1093/biomet/92.1.119 ◽

2005 ◽

Vol 92 (1) ◽

pp. 119-133 ◽

Cited By ~ 16

Author(s):

Eric D. Kolaczyk ◽

Robert D. Nowak

Keyword(s):

Linear Models ◽

Function Estimation ◽

Generalised Linear Models ◽

Nonparametric Function ◽

Nonparametric Function Estimation

Download Full-text

A Sarmanov Distribution with Beta Marginals: An Application to Motor Insurance Pricing

Mathematics ◽

10.3390/math8112020 ◽

2020 ◽

Vol 8 (11) ◽

pp. 2020

Author(s):

Catalina Bolancé ◽

Montserrat Guillen ◽

Albert Pitarque

Keyword(s):

Beta Distribution ◽

Relative Frequency ◽

Linear Models ◽

Speed Limit ◽

Negative Dependence ◽

Night Time ◽

Generalised Linear Models ◽

Insurance Pricing ◽

High Quantiles ◽

Motor Insurance

Background: The Beta distribution is useful for fitting variables that measure a probability or a relative frequency. Methods: We propose a Sarmanov distribution with Beta marginals specified as generalised linear models. We analyse its theoretical properties and its dependence limits. Results: We use a real motor insurance sample of drivers and analyse the percentage of kilometres driven above the posted speed limit and the percentage of kilometres driven at night, together with some additional covariates. We fit a Beta model for the marginals of the bivariate Sarmanov distribution. Conclusions: We find negative dependence in the high quantiles indicating that excess speed and night-time driving are not uniformly correlated.

Download Full-text