BAMP - Bayesian Age-Period-Cohort Modelling and Prediction - Manual
 
BAMP
Bayesian Age-Period-Cohort Modelling and Prediction

Manual

Detailed information on the software, the data and the output can be found in Schmid, Held: Bayesian Age-Period-Cohort Modeling and Prediction - BAMP, Journal of Statistical Software, Vol. 21, Issue 8, Oct 2007. This page gives a short summary of the basic information.

 Windows user: Please see below.


Quick steps

After you have installed BAMP, edit the bamp.ini file or write your own ini-file. Then start the program by typing
bamp ini-file
or (if you use bamp.ini) just type
bamp
On window-based OS you can also start BAMP with a double click on the BAMP-icon. BAMP then will use the bamp.ini file

Input-Files
BAMP needs inputfiles with the population data and the number of deaths. Both files have to be matrices of number of agegroups x number of periods (or transposed, see
dataorder). The data has to be seperated by blanks.

Predictions
If you want to get predictions of future deathrates, set predictions to 1. Set number of predictions to the number of periods for which you want the predictions. If you also have population projection, set predictions to 3.
In order to test the model you can also make prediction for existing data. Then the input files must include this data, the matrices have to be of order number of agegroups x (number of periods + number of predictions). Set deviance to 2 and BAMP will calculate the deviance for the predicted periods.
 

Description of bamp.ini

 

cancer data  Path + File of number of deaths
population data Path + File of population number
output folder Path of Folder for output files
temp folder Path of Folder for temporarily files
output file File for general output; 0 = console
dataorder 1 if data for one period is in a row, 0 if in a column
number of agegroups Number of agegroups
number of periods Number of periods
periods per agegroup Number of periods in one agegroup
predictions 0 = no predictions, 1 = predictions, 2 = predictions for existing data (see deviance)
number of predictions Number of periods to be predicted
number of iterations Number of iterations (including burn in)
burn in Number of iterations for burn in (including tuning)
step Thinning parameter for MCMC simulation
tuning Number of iterations done for tuning
samples 1 if you wish to get all samples, 0 else; samples will be written to "temp folder"
deviance 0 = don't compute deviance, 1 = compute deviance, 2 = compute deviance also for predicted periods (only if predictions = 2)
age block 0 = don't compute age effect, 1 = use RW 1, 2 = use RW 2
age hyperpar.a / b. hyperparameters for Priors of precision
period block 0 = don't compute period effect, 1 = use RW 1, 2 = use RW 2
period hyperpar.a / b. hyperparameters for Priors of precision
cohort block 0 = don't compute cohort effect, 1 = use RW 1, 2 = use RW 2
cohort hyperpar.a and b. hyperparameters for Priors of precision
z hyperpar.a and b. hyperparameters for Priors of precision of overdispersion parameter
quantile 1 to 5 Quantile of output data; -1 no output

Microsoft Windows:  

We recommend to convert bamp.ini to unix format using dos2unix after editing the file with WordPad. Unpack dos2unix to the bamp folder. After editing and saving bamp.ini, drop the bamp.ini file on the dos2unix icon. After this, BAMP should run smoothly on all Windows versions.

 

 
Top
     



Website Baker is released under the GNU General Public License.
This template is made by DanFuh.

Valid XHTML 1.0 Transitional Valid CSS!