4d874309b36dff213fe66475eb7aec6d01bd973d,panel/iv/model.py,IVGMM,fit,#IVGMM#Any#Any#Any#,422

Before Change

        cov_estimator = IVGMMCovariance(x, y, z, params, w,
                                        cov_type, **cov_config)

        results = {"cov_type": cov_type,
                   "weight_mat": w,
                   "weight_type": self._weight_type,
                   "weight_config": self._weight_type,
                   "iterations": iters,
                   "j_stat": self._j_statistic(params, w)}
        pe = self._post_estimation(params, cov_estimator)
        results.update(pe)

        return IVGMMResults(results, self)

After Change

        y, x, z = self._y, self._x, self._z
        nobs = y.shape[0]
        weight_matrix = self._weight.weight_matrix
        w = inv(z.T @ z / nobs)
        _params = params = self.estimate_parameters(x, y, z, w)
        eps = y - x @ params

        iters, norm = 1, 10 * tol
        while iters < iter_limit and norm > tol:
            w = inv(weight_matrix(x, z, eps))
            params = self.estimate_parameters(x, y, z, w)
            eps = y - x @ params
            delta = params - _params
            xpz = x.T @ z / nobs
            if iters == 1:
                v = (xpz @ w @ xpz.T) / nobs
                vinv = inv(v)
            _params = params
            norm = delta.T @ vinv @ delta
            iters += 1

        cov_estimator = IVGMMCovariance(x, y, z, params, w,
                                        cov_type, **cov_config)

        results = self._post_estimation(params, cov_estimator)
        gmm_pe = self._gmm_post_estimation(params, w, cov_type, iters)
        results.update(gmm_pe)

        return IVGMMResults(results, self)

Instances