Prof. Kenneth Train (Berkeley)
Creating Models from Data
This course will be used to do that using simulation models.
Three models are used
1. Analytical Model (Using a known probability disctribution function)
2. Actual Simulation using Probit
3. Mixed Probit (Breaking the unknown into two seperate models one Analytical and one Probit)
All models will be coded in GAUSS
Advantages of GAUSS
Here are some of the things you can do with GAUSS and optional Applications: Optimization, Maximum likelihood estimation, Linear programming, Loglinear models, EIGEN systems, Factorizations (QR Cholesky, LU), Decompositions (SVD and Schur), Equation Solving, Cumulative distribution functions, Autoregression, Time-series cross sectional models, Co-integration models, Rational expectation models, 2 & 3 stage least squares, ARIMA models, Bessel functions, Nonlinear systems of equations, Differential equations, Multinomial logit analysis, Probit analysis, Ordered probit and logit, Exponential duration model with censoring, Descriptive statistics, Limited dependent variable models, Covariance structure analysis, Curve fitting.
