f7bc62ca710f2433273c437be40241f133af2cf4,bambi/models.py,Model,build,#Model#,57

Before Change

        self.R2Y = pd.Series({x:sm.OLS(self.y.data, sm.add_constant(X.drop(x, axis=1))).fit().rsquared
            for x in list(X.columns)})
        self.SDX = X.std()
        self.SDY = self.y.data.std()
        self.meanX = X.mean(0)

        for t in self.terms.values():
            t._setup() 

After Change

            "r2_y":pd.Series({x:pd.stats.api.ols(y=self.y.data.squeeze(), x=X.drop(x, axis=1)).r2
                for x in list(X.columns)}),
            "sd_x":X.std(),
            "sd_y":self.y.data.std(),
            "mean_x":X.mean(axis=0),
            "mean_y":self.y.data.mean()
        }

        // save some info possibly useful for diagnostics, and send to ModelResults
        // mat = correlation matrix of X, w/ diagonal replaced by X means
        mat = X.corr()
        for x in list(mat.columns): mat.loc[x,x] = default_prior_info["mean_x"][x]
        self._diagnostics = {
            // the Variance Inflation Factors (VIF), which is possibly useful for diagnostics
            "VIF":1/(1 - default_prior_info["r2_x"]),
            "corr_mean_X":mat
        }

Instances