remove boot coefs from irm model

Author: SvenKlaassen

doubleml/irm/tests/_utils_iivm_manual.py

@@ -205,7 +205,6 @@ def iivm_orth(g_hat0, g_hat1, m_hat, r_hat0, r_hat1, u_hat0, u_hat1, w_hat0, w_h
 def boot_iivm(y, d, z, thetas, ses, all_g_hat0, all_g_hat1, all_m_hat, all_r_hat0, all_r_hat1,
               all_smpls, score, bootstrap, n_rep_boot,
               n_rep=1, apply_cross_fitting=True, normalize_ipw=True):
-    all_boot_theta = list()
     all_boot_t_stat = list()
     for i_rep in range(n_rep):
         smpls = all_smpls[i_rep]
@@ -215,17 +214,15 @@ def boot_iivm(y, d, z, thetas, ses, all_g_hat0, all_g_hat1, all_m_hat, all_r_hat
             test_index = smpls[0][1]
             n_obs = len(test_index)
         weights = draw_weights(bootstrap, n_rep_boot, n_obs)
-        boot_theta, boot_t_stat = boot_iivm_single_split(
+        boot_t_stat = boot_iivm_single_split(
             thetas[i_rep], y, d, z,
             all_g_hat0[i_rep], all_g_hat1[i_rep], all_m_hat[i_rep], all_r_hat0[i_rep], all_r_hat1[i_rep],
             smpls, score, ses[i_rep], weights, n_rep_boot, apply_cross_fitting, normalize_ipw)
-        all_boot_theta.append(boot_theta)
         all_boot_t_stat.append(boot_t_stat)
 
-    boot_theta = np.hstack(all_boot_theta)
     boot_t_stat = np.hstack(all_boot_t_stat)
 
-    return boot_theta, boot_t_stat
+    return boot_t_stat
 
 
 def boot_iivm_single_split(theta, y, d, z, g_hat0_list, g_hat1_list, m_hat_list, r_hat0_list, r_hat1_list,
@@ -234,7 +231,6 @@ def boot_iivm_single_split(theta, y, d, z, g_hat0_list, g_hat1_list, m_hat_list,
     u_hat0, u_hat1, w_hat0, w_hat1, g_hat0, g_hat1, m_hat, r_hat0, r_hat1 = compute_iivm_residuals(
         y, d, g_hat0_list, g_hat1_list, m_hat_list, r_hat0_list, r_hat1_list, smpls)
 
-    m_hat_adj = np.full_like(m_hat, np.nan, dtype='float64')
     if normalize_ipw:
         m_hat_adj = _normalize_ipw(m_hat, d)
     else:
@@ -258,6 +254,6 @@ def boot_iivm_single_split(theta, y, d, z, g_hat0_list, g_hat1_list, m_hat_list,
                  + np.divide(np.multiply(z, w_hat1), m_hat_adj)
                  - np.divide(np.multiply(1.-z, w_hat0), 1.-m_hat_adj))
 
-    boot_theta, boot_t_stat = boot_manual(psi, J, smpls, se, weights, n_rep, apply_cross_fitting)
+    boot_t_stat = boot_manual(psi, J, smpls, se, weights, n_rep, apply_cross_fitting)
 
-    return boot_theta, boot_t_stat
+    return boot_t_stat

doubleml/irm/tests/_utils_irm_manual.py

@@ -174,7 +174,6 @@ def irm_orth(g_hat0, g_hat1, m_hat, p_hat, u_hat0, u_hat1, d, score):
 def boot_irm(y, d, thetas, ses, all_g_hat0, all_g_hat1, all_m_hat, all_p_hat,
              all_smpls, score, bootstrap, n_rep_boot,
              n_rep=1, apply_cross_fitting=True, normalize_ipw=True):
-    all_boot_theta = list()
     all_boot_t_stat = list()
     for i_rep in range(n_rep):
         smpls = all_smpls[i_rep]
@@ -184,17 +183,15 @@ def boot_irm(y, d, thetas, ses, all_g_hat0, all_g_hat1, all_m_hat, all_p_hat,
             test_index = smpls[0][1]
             n_obs = len(test_index)
         weights = draw_weights(bootstrap, n_rep_boot, n_obs)
-        boot_theta, boot_t_stat = boot_irm_single_split(
+        boot_t_stat = boot_irm_single_split(
             thetas[i_rep], y, d,
             all_g_hat0[i_rep], all_g_hat1[i_rep], all_m_hat[i_rep], all_p_hat[i_rep], smpls,
             score, ses[i_rep], weights, n_rep_boot, apply_cross_fitting, normalize_ipw)
-        all_boot_theta.append(boot_theta)
         all_boot_t_stat.append(boot_t_stat)
 
-    boot_theta = np.hstack(all_boot_theta)
     boot_t_stat = np.hstack(all_boot_t_stat)
 
-    return boot_theta, boot_t_stat
+    return boot_t_stat
 
 
 def boot_irm_single_split(theta, y, d, g_hat0_list, g_hat1_list, m_hat_list, p_hat_list,
@@ -233,9 +230,9 @@ def boot_irm_single_split(theta, y, d, g_hat0_list, g_hat1_list, m_hat_list, p_h
                         np.multiply(p_hat, (1.-m_hat_adj))) \
             - theta * np.divide(d, p_hat)
 
-    boot_theta, boot_t_stat = boot_manual(psi, J, smpls, se, weights, n_rep_boot, apply_cross_fitting)
+    boot_t_stat = boot_manual(psi, J, smpls, se, weights, n_rep_boot, apply_cross_fitting)
 
-    return boot_theta, boot_t_stat
+    return boot_t_stat
 
 
 def fit_sensitivity_elements_irm(y, d, all_coef, predictions, score, n_rep):

doubleml/irm/tests/test_iivm.py

@@ -92,17 +92,15 @@ def dml_iivm_fixture(generate_data_iivm, learner, score, normalize_ipw, trimming
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
-                                            res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                            res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
-                                            all_smpls, score, bootstrap, n_rep_boot,
-                                            normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
+                                res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                                res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
+                                all_smpls, score, bootstrap, n_rep_boot,
+                                normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_iivm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_iivm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_iivm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
 
     return res_dict
@@ -125,9 +123,6 @@ def test_dml_iivm_se(dml_iivm_fixture):
 @pytest.mark.ci
 def test_dml_iivm_boot(dml_iivm_fixture):
     for bootstrap in dml_iivm_fixture['boot_methods']:
-        assert np.allclose(dml_iivm_fixture['boot_coef' + bootstrap],
-                           dml_iivm_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_iivm_fixture['boot_t_stat' + bootstrap],
                            dml_iivm_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)

doubleml/irm/tests/test_iivm_classifier.py

@@ -84,17 +84,15 @@ def dml_iivm_classifier_fixture(generate_data_iivm_binary, learner, score, norma
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
-                                            res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                            res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
-                                            all_smpls, score, bootstrap, n_rep_boot,
-                                            normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
+                                res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                                res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
+                                all_smpls, score, bootstrap, n_rep_boot,
+                                normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_iivm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_iivm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_iivm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
 
     return res_dict
@@ -117,9 +115,6 @@ def test_dml_iivm_se(dml_iivm_classifier_fixture):
 @pytest.mark.ci
 def test_dml_iivm_boot(dml_iivm_classifier_fixture):
     for bootstrap in dml_iivm_classifier_fixture['boot_methods']:
-        assert np.allclose(dml_iivm_classifier_fixture['boot_coef' + bootstrap],
-                           dml_iivm_classifier_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_iivm_classifier_fixture['boot_t_stat' + bootstrap],
                            dml_iivm_classifier_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)

doubleml/irm/tests/test_iivm_subgroups.py

@@ -104,17 +104,15 @@ def dml_iivm_subgroups_fixture(generate_data_iivm, learner, score, normalize_ipw
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
-                                            res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                            res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
-                                            all_smpls, score, bootstrap, n_rep_boot,
-                                            normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
+                                res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                                res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
+                                all_smpls, score, bootstrap, n_rep_boot,
+                                normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_iivm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_iivm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_iivm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
 
     return res_dict
@@ -137,9 +135,6 @@ def test_dml_iivm_subgroups_se(dml_iivm_subgroups_fixture):
 @pytest.mark.ci
 def test_dml_iivm_subgroups_boot(dml_iivm_subgroups_fixture):
     for bootstrap in dml_iivm_subgroups_fixture['boot_methods']:
-        assert np.allclose(dml_iivm_subgroups_fixture['boot_coef' + bootstrap],
-                           dml_iivm_subgroups_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_iivm_subgroups_fixture['boot_t_stat' + bootstrap],
                            dml_iivm_subgroups_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)

doubleml/irm/tests/test_iivm_tune.py

@@ -157,17 +157,15 @@ def dml_iivm_fixture(generate_data_iivm, learner_g, learner_m, learner_r, score,
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
-                                            res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                            res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
-                                            all_smpls, score, bootstrap, n_rep_boot,
-                                            normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_iivm(y, d, z, res_manual['thetas'], res_manual['ses'],
+                                res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                                res_manual['all_m_hat'], res_manual['all_r_hat0'], res_manual['all_r_hat1'],
+                                all_smpls, score, bootstrap, n_rep_boot,
+                                normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_iivm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_iivm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_iivm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
 
     return res_dict
@@ -190,9 +188,6 @@ def test_dml_iivm_se(dml_iivm_fixture):
 @pytest.mark.ci
 def test_dml_iivm_boot(dml_iivm_fixture):
     for bootstrap in dml_iivm_fixture['boot_methods']:
-        assert np.allclose(dml_iivm_fixture['boot_coef' + bootstrap],
-                           dml_iivm_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_iivm_fixture['boot_t_stat' + bootstrap],
                            dml_iivm_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)

doubleml/irm/tests/test_irm.py

@@ -109,21 +109,18 @@ def dml_irm_fixture(generate_data_irm, learner, score, normalize_ipw, trimming_t
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_irm(y, d, res_manual['thetas'], res_manual['ses'],
-                                           res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                           res_manual['all_m_hat'], res_manual['all_p_hat'],
-                                           all_smpls, score, bootstrap, n_rep_boot,
-                                           normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_irm(y, d, res_manual['thetas'], res_manual['ses'],
+                               res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                               res_manual['all_m_hat'], res_manual['all_p_hat'],
+                               all_smpls, score, bootstrap, n_rep_boot,
+                               normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_irm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
         np.random.seed(3141)
         dml_irm_obj_ext.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_irm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_irm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_ext'] = dml_irm_obj_ext.boot_coef
         res_dict['boot_t_stat' + bootstrap + '_ext'] = dml_irm_obj_ext.boot_t_stat
 
     # sensitivity tests
@@ -163,12 +160,6 @@ def test_dml_irm_se(dml_irm_fixture):
 @pytest.mark.ci
 def test_dml_irm_boot(dml_irm_fixture):
     for bootstrap in dml_irm_fixture['boot_methods']:
-        assert np.allclose(dml_irm_fixture['boot_coef' + bootstrap],
-                           dml_irm_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
-        assert np.allclose(dml_irm_fixture['boot_coef' + bootstrap],
-                           dml_irm_fixture['boot_coef' + bootstrap + '_ext'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_irm_fixture['boot_t_stat' + bootstrap],
                            dml_irm_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)

doubleml/irm/tests/test_irm_classifier.py

@@ -83,17 +83,15 @@ def dml_irm_classifier_fixture(generate_data_irm_binary, learner, score, normali
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_irm(y, d, res_manual['thetas'], res_manual['ses'],
-                                           res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                           res_manual['all_m_hat'], res_manual['all_p_hat'],
-                                           all_smpls, score, bootstrap, n_rep_boot,
-                                           normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_irm(y, d, res_manual['thetas'], res_manual['ses'],
+                               res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                               res_manual['all_m_hat'], res_manual['all_p_hat'],
+                               all_smpls, score, bootstrap, n_rep_boot,
+                               normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_irm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_irm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_irm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
 
     return res_dict
@@ -116,9 +114,6 @@ def test_dml_irm_se(dml_irm_classifier_fixture):
 @pytest.mark.ci
 def test_dml_irm_boot(dml_irm_classifier_fixture):
     for bootstrap in dml_irm_classifier_fixture['boot_methods']:
-        assert np.allclose(dml_irm_classifier_fixture['boot_coef' + bootstrap],
-                           dml_irm_classifier_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_irm_classifier_fixture['boot_t_stat' + bootstrap],
                            dml_irm_classifier_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)

doubleml/irm/tests/test_irm_tune.py

@@ -123,17 +123,15 @@ def dml_irm_fixture(generate_data_irm, learner_g, learner_m, score, normalize_ip
 
     for bootstrap in boot_methods:
         np.random.seed(3141)
-        boot_theta, boot_t_stat = boot_irm(y, d, res_manual['thetas'], res_manual['ses'],
-                                           res_manual['all_g_hat0'], res_manual['all_g_hat1'],
-                                           res_manual['all_m_hat'], res_manual['all_p_hat'],
-                                           all_smpls, score, bootstrap, n_rep_boot,
-                                           normalize_ipw=normalize_ipw)
+        boot_t_stat = boot_irm(y, d, res_manual['thetas'], res_manual['ses'],
+                               res_manual['all_g_hat0'], res_manual['all_g_hat1'],
+                               res_manual['all_m_hat'], res_manual['all_p_hat'],
+                               all_smpls, score, bootstrap, n_rep_boot,
+                               normalize_ipw=normalize_ipw)
 
         np.random.seed(3141)
         dml_irm_obj.bootstrap(method=bootstrap, n_rep_boot=n_rep_boot)
-        res_dict['boot_coef' + bootstrap] = dml_irm_obj.boot_coef
         res_dict['boot_t_stat' + bootstrap] = dml_irm_obj.boot_t_stat
-        res_dict['boot_coef' + bootstrap + '_manual'] = boot_theta
         res_dict['boot_t_stat' + bootstrap + '_manual'] = boot_t_stat
 
     return res_dict
@@ -156,9 +154,6 @@ def test_dml_irm_se(dml_irm_fixture):
 @pytest.mark.ci
 def test_dml_irm_boot(dml_irm_fixture):
     for bootstrap in dml_irm_fixture['boot_methods']:
-        assert np.allclose(dml_irm_fixture['boot_coef' + bootstrap],
-                           dml_irm_fixture['boot_coef' + bootstrap + '_manual'],
-                           rtol=1e-9, atol=1e-4)
         assert np.allclose(dml_irm_fixture['boot_t_stat' + bootstrap],
                            dml_irm_fixture['boot_t_stat' + bootstrap + '_manual'],
                            rtol=1e-9, atol=1e-4)