• The count data model has changed. It is now possible to have severalgamma parameters instead of gamma = 1 as assumed in the previousversion.
• 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 effectsinstead of one.

Some function name haschanged

• SARML has been replaced bysar.
• SARTML has been replaced bysart.
• CDnetNPL has been replaced bycdnet.
• simSARnet has been replaced bysimsar.
• simTobitnet has been replaced bysimsart.
• simCDnet has been replaced bysimcdnet.
• netformation has been replaced byhomophily.

SART model underrational expectations

It is now possible to estimate the SART model under rationalexpectations. In the previous version, the SART model is only availableunder complete information.

Changes in version 2.0.1

This version follows the major revision of the paper in September2022. - The count data model includes a more flexible specification.Especially, it is possible to assume that the cut points are not equallyspaced for large values of the dependent variable. - I also implement anetwork formation model with degree heterogeneity as fixed effects (seeYan et al.,2019). - Models under incomplete information are now estimated usingLBFGS algorithm of the package RcppNumerical. Thus, the optimization isperformed in C++ and is very fast compared to the version 1.0.1.

Changes in versions 2.0.2and 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 toC++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 tochoose Rbar.

Changes in version 2.1.3

• homophily has been changed tohomophily.refor the random effect models.
• homophily.FE has heen changed tohomophily.fe.
• Random effects inhomophily.re can be one side or twosides.
• Fixed effects inhomophily.fe can be one side or twosides.
• Symmetric network models are included inhomophily.re.
• Symmetric network models are included inhomophily.fe.
• The functionhomophili.data is added to convert databetween directed network models and symmetric network models.

Changes in version 2.2.0

• Following the revision of the paperCount Data Model with SocialInteraction under Rational Expectations,cdnet nowallows heterogeneity in peer effects. For example, peer effects can beestimated separately for groups such as Blacks and Whites, or Girls andBoys.
  
• cdnet now also allows heterogeneity in the cut-points.The cost function in the utility function is nonparametric and dependson characteristicsX such as gender, race, etc.

Changes in version 2.2.1

• Added the Newton–Raphson method for estimating the homophilymodel.

Changes in version 2.2.2

• Inhomophili.fe, elements in the estimate and gradientvectors are now labelled.
  
• The summary method forcdnet andsart nolonger displays marginal effects.
  
• Themeffects method has been added and can be appliedto the outputs ofcdnet andsart.
  This method computes direct, indirect, and total marginal effects.