Model Estimation and Data Analysis: Nonlinear Regression and Loglinear Regression Models

LIMDEP software offers numerous tools for nonlinear regression including generalized method of moments (GMM), nonlinear least squares regression and nonlinear systems of equations. LIMDEP’s features for loglinear models include generalized linear models (GLM), fractional response models, many forms of binominal and multinomial regression and more.

Nonlinear Least Squares Regression

  • Estimators
    • Nonlinear instrumental variables
    • Efficient GMM
    • Weighted nonlinear least squares
  • Model specification
    • Recursive function definition
    • Analytic or numeric derivatives
    • All algebraic operators and any level of parentheses
    • Matrix bilinear and quadratic forms
    • Mathematical functions (exp, sin, phi, gma, etc.)
    • Univariate and bivariate normal integrals
    • Hermite and Laguerre quadrature
    • Simulation estimation
    • Fixed parameters
    • Several algorithms
  • Robust standard errors (White, Newey-West)
  • Box-Cox regression
  • Nonlinear 2SLS, IV, GMM
  • Retain all results for post estimation

Nonlinear Systems of Equations

  • Nonlinear least squares
  • Instrumental variables
  • Efficient GMM estimation

Loglinear Models (GLM)

  • Weibull regression
  • Inverse Gaussian regression
  • Exponential regression
  • Gamma regression
  • Beta regression
  • Binomial regression model (All are supported for cross section, fixed effects, random effects, random parameters, latent class formulations.)