pysindy_add.txt

def get_coef_list(self):
        if hasattr(self, 'coef_list'):
            return self.coef_list
        else:
            return [self.coefficients()]

def validate_coef_list(self, *validation_data, target_index=0, criterion='aic'):
        if len(validation_data) == 2:
            X_pre, y_pre = validation_data
        else:
            X_pre, y_pre = self.feature_library.cached_xp_full, self.cached_x_dot
        # Consider directly from model.get_coef_list() ไม่คิด coefficients ใหม่ จากข้อมูลทั้งหมด
        # print_pde(model.get_coef_list()[np.argmin(np.sum((np.squeeze(np.tensordot(X_pre, np.array(model.get_coef_list()).T, axes=([-1], [0])), axis=1)-y_pre)**2, axis=0))].reshape(-1,1), model.get_feature_names())
        if len(X_pre.shape) < 3:
            X_pre = np.expand_dims(X_pre, 0)
        coef_list = np.squeeze(np.array(self.get_coef_list())[:, target_index:target_index+1, :], axis=1)
        if hasattr(self.optimizer, 'threshold'):
            print("hasattr threshold in optimizer...")
            # th = self.optimizer.threshold
            th = max(self.optimizer.threshold, np.nextafter(0, 1))
            all_indices = set([tuple(np.where(np.abs(coef_list[i])>=th)[0]) for i in range(len(coef_list))])
        else:
            all_indices = set([tuple(np.nonzero(coef_list[i])[0]) for i in range(len(coef_list))])
        out = []
        for b in range(X_pre.shape[0]):
            ols_models = [(OLS(y_pre[:, target_index:target_index+1], X_pre[b][:, init_feature]).fit(), init_feature) for init_feature in all_indices]
            best_ols_index = np.argmin([getattr(m, criterion) for m, _ in ols_models])
            print_pde(ols_models[best_ols_index][0].params.reshape(-1,1), 
                    np.array(self.get_feature_names())[list(ols_models[best_ols_index][1])], ut=f"{self.feature_names[target_index]}_t")
            out.append(ols_models[best_ols_index])
        return out