Workshop constructing effects in RSiena

Teacher: Tom A.B. Snijders

The workshop will be online

This is a workshop for RSiena users. There are no financial costs.
The aim is that workshop participants will be able to create their own effects.
A prerequisite for the workshop is a basic knowledge of C++ programming.

How to prepare

The following are some things to prepare for the workshop.
In any case, I hope you are able to do the first two (know about inheritance and know how to install an R package from source).

If you do not know about inheritance in C++, then please study this (it does not take much time).
There is a lot of material about this on the web, e.g.:
It will be assumed you know about inheritance and the use of virtual functions.
Please prepare how to build an R package from source on your own computer. Of course, the R package will be RSiena.
Steps:
1. In R, make sure that you have installed the packages that are "dependencies" in RSiena:
  Matrix, lattice, parallel, MASS, methods, xtable, network, codetools.
2. If you are using a Windows computer, install RTools from https://cran.r-project.org/bin/windows/Rtools/.
3. Make a new directory for the workshop.
4. Download the source code of RSiena from https://github.com/stocnet/rsiena/releases.
  Here you will find five "assets"; download one of those indicated as "source code". Unpack the file; this will give a directory which has a name such as rsiena-1.4.19, where the version number might be changed; put this as a subdirectory in the directory for the workshop you just created.
5. If you are a Windows user, like me:
  I find it most convenient to work with three batch files.
  My R program is installed in directory "C:\ProgramFiles\R\R-4.4.1\bin\R.exe". You will have it in some other directory, and you will have to make the corresponding changes in the following batch files. My batch files are
  - buildsiena.bat
  - checksiena.bat; you will need this only to check that the code is correct (it will take a while, so do this only if you are unsure about this);
  - install.bat
  Download these three batch files, modify them (in a raw text editor) so that they mention the new version of RSiena and point to your own R installation; then first run buildsiena.bat and then installsiena.bat. Just for your interest, you also could run checksiena.bat.
6. If you are a Mac or Linux/Unix user, please find your way in the "further background" below.
To get a sense of some straightforward C++ classes that are used for creating effects, look at the following examples of C++ classes in RSiena.
You can find them in https://github.com/stocnet/rsiena/tree/main/src/model/effects:
1. AverageReciprocatedAlterEffect, which is used for the behavior effect avXAlt;
2. IndegreePopularityEffect, which is used for the network effects inPop, inPopSqrt, and inPop.c.
Also look at their definitions in the manual; calculating the effect statistics in these classes is done by the use of, respectively, egoStatistic and tieStatistic
For further background about constructing R packages, there is a lot of material on the web. If you are on a Mac or on Linux/Unix, and you are unsure how to proceed to install a package from source, you will find help in some of the material. Examples are:
- The official R documentation for building a package
- R Packages (2e) by Hadley Wickham and Jennifer Bryan
- R package primer (a minimal tutorial) by Karl Broman.
  The relevant part is "Building and installing an R package", but if you are new to this, you can look at what precedes. What follows is not important for this workshop.
- GeeksforGeeks: "How to Install an R Package from Source?" .
- Writing R packages in Rstudio (Tutorial adapted from stirlingcodingclub.github.io by Brad Duthie).

Tuesday November 5, 2024; 10.00-12.00 (time zone Amsterdam NL)

I will treat the following slide presentations :

We shall see how far I get.
In any case, we should come far enough so that you can make the assignment.

For the next meeting, there is the following assignment:

Effects in RSiena: assignment

Monday November 11, 2024; 14.00-16.00 (time zone Amsterdam NL)

Treat the assigment
Question time
Whatever is left
Question time again.

Further materials

Websites :
- The main Siena page
Manuals :
- RSiena manual (see Chapter 18)
Literature :
- Siena algorithms.

Back to the main Siena page

	Workshop constructing effects in RSiena Teacher: Tom A.B. Snijders The workshop will be online This is a workshop for RSiena users. There are no financial costs. The aim is that workshop participants will be able to create their own effects. A prerequisite for the workshop is a basic knowledge of C++ programming.


	How to prepare The following are some things to prepare for the workshop. In any case, I hope you are able to do the first two (know about inheritance and know how to install an R package from source). If you do not know about inheritance in C++, then please study this (it does not take much time). There is a lot of material about this on the web, e.g.: https://www.tutorialspoint.com/cplusplus/cpp_inheritance.htm https://www.tutorialspoint.com/cplusplus/cpp_function_overriding.htm https://en.cppreference.com/w/cpp/language/virtual It will be assumed you know about inheritance and the use of virtual functions. Please prepare how to build an R package from source on your own computer. Of course, the R package will be RSiena. Steps: In R, make sure that you have installed the packages that are "dependencies" in RSiena: Matrix, lattice, parallel, MASS, methods, xtable, network, codetools. If you are using a Windows computer, install RTools from https://cran.r-project.org/bin/windows/Rtools/. Make a new directory for the workshop. Download the source code of RSiena from https://github.com/stocnet/rsiena/releases. Here you will find five "assets"; download one of those indicated as "source code". Unpack the file; this will give a directory which has a name such as rsiena-1.4.19, where the version number might be changed; put this as a subdirectory in the directory for the workshop you just created. If you are a Windows user, like me: I find it most convenient to work with three batch files. My R program is installed in directory "C:\ProgramFiles\R\R-4.4.1\bin\R.exe". You will have it in some other directory, and you will have to make the corresponding changes in the following batch files. My batch files are buildsiena.bat checksiena.bat; you will need this only to check that the code is correct (it will take a while, so do this only if you are unsure about this); install.bat Download these three batch files, modify them (in a raw text editor) so that they mention the new version of RSiena and point to your own R installation; then first run buildsiena.bat and then installsiena.bat. Just for your interest, you also could run checksiena.bat. If you are a Mac or Linux/Unix user, please find your way in the "further background" below. To get a sense of some straightforward C++ classes that are used for creating effects, look at the following examples of C++ classes in RSiena. You can find them in https://github.com/stocnet/rsiena/tree/main/src/model/effects: AverageReciprocatedAlterEffect, which is used for the behavior effect avXAlt; IndegreePopularityEffect, which is used for the network effects inPop, inPopSqrt, and inPop.c. Also look at their definitions in the manual; calculating the effect statistics in these classes is done by the use of, respectively, egoStatistic and tieStatistic For further background about constructing R packages, there is a lot of material on the web. If you are on a Mac or on Linux/Unix, and you are unsure how to proceed to install a package from source, you will find help in some of the material. Examples are: The official R documentation for building a package R Packages (2e) by Hadley Wickham and Jennifer Bryan R package primer (a minimal tutorial) by Karl Broman. The relevant part is "Building and installing an R package", but if you are new to this, you can look at what precedes. What follows is not important for this workshop. GeeksforGeeks: "How to Install an R Package from Source?" . Writing R packages in Rstudio (Tutorial adapted from stirlingcodingclub.github.io by Brad Duthie).


	Tuesday November 5, 2024; 10.00-12.00 (time zone Amsterdam NL) I will treat the following slide presentations : Effects in RSiena: The R side Effects in RSiena on the C++ side We shall see how far I get. In any case, we should come far enough so that you can make the assignment. For the next meeting, there is the following assignment: Effects in RSiena: assignment


	Monday November 11, 2024; 14.00-16.00 (time zone Amsterdam NL) Treat the assigment Question time Whatever is left Question time again.


	Further materials Websites : The main Siena page Manuals : RSiena manual (see Chapter 18) Literature : Siena algorithms.
	Back to the main Siena page