9439f5215fde61ee40910b0320ff4dd8d28571e2,fancyimpute/MICE.py,MICE,perform_imputation_round,#MICE#,121

Before Change

            n_missing_for_this_col = n_missing_for_each_column[col_idx]
            if n_missing_for_this_col > 0:  // if we have any missing data at all
                observed_row_mask_for_col = observed_mask[:, col_idx]
                other_cols = np.array(list(range(0, col_idx)) + list(range(col_idx + 1, n_cols)))
                output = X_filled[observed_row_mask_for_col, col_idx]
                if (n_cols - int(self.add_ones) > self.n_nearest_columns):
                    // probability of column draw is proportional to absolute 
                    // pearson correlation

After Change

            abs_correlation_matrix = np.abs(np.corrcoef(X_filled.T))

        n_missing_for_each_column = missing_mask.sum(axis=0)
        ordered_column_indices = np.arange(n_cols)
        for col_idx in visit_indices:
            missing_mask_col = missing_mask[:, col_idx]  // missing mask for this column
            n_missing_for_this_col = n_missing_for_each_column[col_idx]
            if n_missing_for_this_col > 0:  // if we have any missing data at all
                observed_row_mask_for_col = observed_mask[:, col_idx]
                output = X_filled[observed_row_mask_for_col, col_idx]

                other_cols = np.concatenate([
                    ordered_column_indices[:col_idx],
                    ordered_column_indices[col_idx + 1:]
                ])

                if n_original_cols > self.n_nearest_columns:
                    // probability of column draw is proportional to absolute
                    // pearson correlation

Instances