Publications by Wingfeet
change in weight of cars plot
Based on last week’s faster algorithm I wanted to finish with car weights. Unfortunately a fail again. By now it is a fail of myself, it needs a bit more dedication and grunt than I am willing and able to give for this blog. This week I added some extra functions around the existing functions so I could harvest final results prett...
4821 sym 8 img
Rain in Netherlands during the past 100 years
Climate has my interest. But discussions on climate change seem to be focused on temperature. In real life, we look at temperature, rain, sunshine and wind. I was therefor happy to find a load of rain data on Royal Netherlands Meteorological Institute. In this post I plot some of the data.DataThe actual page for the data is http://ww...
5747 sym 6 img
Did the pattern of rain change in the last 100 years?
Last week I showed rain data from six stations in Netherlands years 1906 till now. The obvious next question is; did it change? A surprisingly difficult question. The data is not normal distributed, but it is time-correlated, location correlated.The dataAs described before; the data are from KNMI; ROYAL NETHERLANDS METEOROLOGICAL INSTITUTE. Th...
4549 sym 4 img
A JAGS calculation on pattern of rain January 1906-1915 against 2003-2012
Two weeks ago I showed rain data from six stations in Netherlands years 1906 till now. Last week I showed that frequency of days with and without rain differed between December 1906-1915 and December 2003-2012. This week I am considering the same data as t-distributed with a left truncation at 0 mm rain. This can be modeled very easy in JAGS. T...
4089 sym
More rainfall calculations – REML
I wanted to have a look at various REML methods for a long time. The rainfall data seemed a nice example. On top of that, FreshBiostats had a blog post ‘Mixed Models in R: lme4, nlme, or both?’. So lme4 it is.The data As described before; the data are from KNMI; ROYAL NETHERLANDS METEOROLOGICAL INSTITUTE. The actual page for ...
3064 sym 4 img
Exercise in REML/Mixed model
I want to build a bit more experience in REML, so I decided to redo some of the SAS examples in R. This post describes the results of example 59.1 (page 5001, SAS(R)/STAT User guide 12.3 link). Following the list from freshbiostats I will analyze using lme4 and MCMCglm.DataThe data is a split plot design. To quote ‘The split-plot d...
8544 sym
More REML exercise
Last week I tried exercise 1 of the SAS(R) proc mixed with R libraries lme4 and MCMCglm. So this week I aimed for exercise 2 but ended up redoing exercise 1 with nlme.Exercise 2 results gave me problems with library lme4 and latter parts of the exercise had special correlation structures which seem not feasible with lme4. So library nlme came in...
5674 sym
Mixed models exercise 2. Repeated measurements
Continuing my exploration of mixed models, I now understand what is happening in the second SAS(R)/STAT example for proc mixed (page 5007 of the SAS/STAT 12.3 Manual). It is all about correlation between the time-points within subjects. The data as such is simple, size measurements of children at ages 8, 10, 12 and 14. The subject of ...
10672 sym
Mixed models; Random Coefficients, part 1
Continuing with my exploration of mixed models I am now at the first part of random coefficients: example 59.5 for proc mixed (page 5034 of the SAS/STAT 12.3 Manual). This means I skipped examples 59.3 (plotting the likelihood) and 59.4 (known G and R). The latter I might want to do later, though I find this to be quite a strong prior...
8250 sym 2 img
Mixed models; Random Coefficients, part 2
Continuing from random coefficients part 1, it is time for the second part. To quote the SAS/STAT manual ‘a random coefficients model with error terms that follow a nested structure‘. The additional random variable is monthc, which is a factor containing the months and nested under batch. Hence there is one additional statement i...
8066 sym