The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Changes in version 1.0.1

The new version follows the major revision of the paper in April 2022

• The count data model has changed. It is now possible to have several gamma parameters instead of gamma = 1 as assumed in the previous version.
• In the count data model identification, sigma is now set to 1, because there are several gamma’s.
• The network formation model now includes two unobserved effects instead of one.

Some function name has changed

• SARML has been replaced by sar.
• SARTML has been replaced by sart.
• CDnetNPL has been replaced by cdnet.
• simSARnet has been replaced by simsar.
• simTobitnet has been replaced by simsart.
• simCDnet has been replaced by simcdnet.
• netformation has been replaced by homophily.

SART model under rational expectations

It is now possible to estimate the SART model under rational expectations. In the previous version, the SART model is only available under complete information.

Changes in version 2.0.1

This version follows the major revision of the paper in September 2022. - The count data model includes a more flexible specification. Especially, it is possible to assume that the cut points are not equally spaced for large values of the dependent variable. - I also implement a network formation model with degree heterogeneity as fixed effects (see Yan et al., 2019). - Models under incomplete information are now estimated using LBFGS algorithm of the package RcppNumerical. Thus, the optimization is performed in C++ and is very fast compared to the version 1.0.1.

Changes in versions 2.0.2 and 2.0.3

Note and Warning found in the check for MACOS have been fixed

Changes in version 2.1.0

R defaulted to C++11 in R 4.0.0, to C++14 in R 4.2.0 and to C++17.

Changes in version 2.1.1

Fixed effect is allowed in the model SAR.

Changes in version 2.1.2

• I address the problem of single-agent subnetwork.
• AIC and BIC are added to the output of cdnet. They can be used to choose Rbar.

Changes in version 2.1.3

• homophily has been changed to homophily.re for the random effect models.
• homophily.FE has heen changed to homophily.fe.
• Random effects in homophily.re can be one side or two sides.
• Fixed effects in homophily.fe can be one side or two sides.
• Symmetric network models are included in homophily.re.
• Symmetric network models are included in homophily.fe.
• The function homophili.data is added to convert data between directed network models and symmetric network models.

Changes in version 2.2.0

• Following the revision of the paper Count Data Model with Social Interaction under Rational Expectations, cdnet now allows heterogeneity in peer effects. For example peer effects can be estimated between Blacks and Whites, Girls and Boys etc.
• cdnet allows heterogeneity in the cut-points. The cost function in the utility function is nonparametric and depends on some characteristics X such as gender, race, etc.

Changes in version 2.2.1

• The Newton-Raphson method has been added for estimating the homophily model.

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.