Compile aML with GCC 4.4.3 (64-bit)

I was able to successfully compile aML with GCC 4.4.3, which is part of Ubuntu 10.04. I only need to make changes on two files to avoid warning messages: "encrypt.f" (replacing "data str95/'Y,h=(ufSBF&H6W!LZ|boq1yJsX^Ax+K3U5r[mDlNOg7jn;-V:z]k/e *)v#2GQC$t''{c\\">%@~w?8dMREiT0a*9}<`\\._4PIp'/" with "data str95/'Y,h=(ufSBF&H6W!LZ|boq1yJsX^Ax+K3U5r[mDlNOg7jn;-V:z]k/e *)v#2GQC$t''{c\">%@~w?8dMREiT0a*9}<`\._4PIp'/") and "getdatainfo.f" (replacing "data backslash/'\\'/" with "data backslash/'\'/")

64-bit aML

I was able to compile and run aML under 64-bit Ubuntu 9.10. The only problem was "maktab", so I replaced my binary with the one supplied by the official web site. I will do some benchmark and see if these binaries are faster.

Intel Fortran and aML

I am happy to discover that aML can be successfully compiled on Ubuntu using the non-commercial Intel Fortran compiler v.11.

Starting value for aML

Feeding aML good starting values is crucial for fast convergence. I followed the suggestions offered by the aML manual for a while. Now I no longer need to struggle with the manual: it becomes a fun game that I am really good at.

The secret is to build up complicated models piece by piece, use estimates from the simpler models as the starting values of the more complicated ones; and use stepwise procedure to get good estimates for the intercepts, then fixed effects, then random effects.

Compile aML under Cygwin using g77

I finally got the time to setup a cygwin environment on my computer and tried to compile aML. After several trial and error, I got it successfully built and tested. The procedure is pretty straightforward, with only one thing to keep in mind: choose g77 3.3 instead of 3.4 otherwise the generated binary will issue an error message "core dump" (and this is probably why aML cannot be directly built on Ubuntu 8.04). I only tested on my 32-bit machine and will try it on my 64-bit Vista office machine on Monday.

This is exciting!

Semi-automatic table making

Both Stata and R have nice table making facilities, but aML is still lacking this. Here is how I convert aML outputs into Word tables.

1. Estimate aML models;
2. Use "maktab" facilities to put different models into one output;
3. Copy and paste the results into a text file;
4. Use Excel to read in the file;
5. Copy and paste the Excel worksheet into Word;
6. Edit the Word manually from there.

When I have time, I may write a python script to automate this process; but for now, this seems to work reasonably well.

Predict, simulate, and graph aML results

The multilevel and multiprocess software aML is quite amazing. It handles a much greater number of models than most other packages, and it is fast to reach convergence given suitable starting values. However, it does not have lots of facilities for results presentation.

I have used this software for several of my own research. I think it will be highly desirable to streamline the communication between aML and Stata or R. I can think of the following steps:

1. Estimate a mode in aML;
2. Extract point estimates and variance/covariance from the output file using Python;
3. Transfer the extracted results to R;
4. Predict, simulate, and graph the results with Zelig or other facilities built into R.
   

aML reported random effects

The random effects that aML reports are standard deviation and correlation, not variance and covariance.

Compiling aML under Ubuntu 7.04, again!

I decided to give to it another try. I wrote to Stan, reporting the problem I had when trying to compiling aML on my Ubuntu 7.04 box. He pointed it out that it is likely to be a bug in the compiler. Then I ask myself: why not try a different compiler?

Some google search shows that alternative FORTRAN 77 under linux includes PGI, Intel, Absoft, among others. PGI offers 30 days trial, so I decided to give it a try. The pgf77 generates correct binary for "aml", "bigaml", and "hugeaml", but it creates problems for "mktab", a utility to create tabular results out of aML output files. I need to re-compile this utility using the old g77 compiler.

In short, an easy solution to the problem will be to use pgf77 to generate the main binaries and to use the g77 to generate the auxiliary binaries, then put them together in one place (in the system path).

I have checked using both the provided samples and my own data, so far so good.

More aML stuff

This software review is also a good introduction of aML: http://www.cmm.bristol.ac.uk/learning-training/multilevel-m-software/reviewaml.pdf

Dan Powers (http://www.la.utexas.edu/research/faculty/dpowers/) also has some aML-related meterials, scattering around at different places.

Mplus vs. aML

I am working on a joint model of miscarriage and child mortality. The idea is that higher rate of miscarriage may end up with a more highly selected newborns, who are less vulnerable to premture death. This is a Heckman selection type of model, with the exception that the main equation is a hazard model instead of a continuous model. I have been trying to make Mplus to do the job, as hinted by Bengt (http://www.statmodel.com/discussion/messages/11/740.html?1155563326). Then I got a direct response from Linda that this type of model cannot be estimated using Mplus (http://www.statmodel.com/discussion/messages/14/2459.html?1185842830). I guess this leaves me no other alternatives but to go to aML.

aML is a fine software. I used it in 2004 for a chapter of my dissertation (at that time it cost about 1,000 bucks). Now it is free and open-source, and everyone can look at the code (in FORTRAN) to see how certain things are done. I am a little surprised that it has not attracted more attention.

What I would really like to see is a wrapper program in Stata and can 1) export data to aML, 2) automatically generate sensible starting values and feed them to aML, and 3) gather coefficients estimated in aML for post-estimation manipulation.

If nobody has already started working on this, I might end up doing it myself.