Update example.

Author: tsalo

examples/02_meta-analysis/plot_meta-analysis_walkthrough.py

@@ -145,28 +145,24 @@ def var_to_ci(y, v, n):
 # The two combination models in PyMARE are Stouffer's and Fisher's Tests.
 #
 # Notice that these models don't use :class:`pymare.core.Dataset` objects.
-stouff = pymare.estimators.Stouffers(input='z')
-stouff.fit(y=z[:, None])
-stouff_summary = stouff.summary()
+stouff = pymare.estimators.StoufferCombinationTest()
+stouff.fit(z[:, None])
 print('Stouffers')
-print('z: {}'.format(stouff_summary.z))
-print('p: {}'.format(stouff_summary.p))
+print('p: {}'.format(stouff.params_["p"]))
 print()
 
-fisher = pymare.estimators.Fishers(input='z')
-fisher.fit(y=z[:, None])
-fisher_summary = fisher.summary()
+fisher = pymare.estimators.FisherCombinationTest()
+fisher.fit(z[:, None])
 print('Fishers')
-print('z: {}'.format(fisher_summary.z))
-print('p: {}'.format(fisher_summary.p))
+print('p: {}'.format(fisher.params_["p"]))
 
 ###############################################################################
 # Now we have a fixed effects model
 # `````````````````````````````````````````````````````````````````````````````
 # This estimator does not attempt to estimate between-study variance.
 # Instead, it takes ``tau2`` (:math:`\tau^{2}`) as an argument.
 wls = pymare.estimators.WeightedLeastSquares()
-wls.fit(dset)
+wls.fit_dataset(dset)
 wls_summary = wls.summary()
 results['Weighted Least Squares'] = wls_summary.to_df()
 print('Weighted Least Squares')
@@ -184,31 +180,31 @@ def var_to_ci(y, v, n):
 # can use either maximum-likelihood (ML) or restricted maximum-likelihood (REML)
 # to iteratively estimate it.
 dsl = pymare.estimators.DerSimonianLaird()
-dsl.fit(dset)
+dsl.fit_dataset(dset)
 dsl_summary = dsl.summary()
 results['DerSimonian-Laird'] = dsl_summary.to_df()
 print('DerSimonian-Laird')
 print(dsl_summary.to_df().T)
 print()
 
 hedge = pymare.estimators.Hedges()
-hedge.fit(dset)
+hedge.fit_dataset(dset)
 hedge_summary = hedge.summary()
 results['Hedges'] = hedge_summary.to_df()
 print('Hedges')
 print(hedge_summary.to_df().T)
 print()
 
 vb_ml = pymare.estimators.VarianceBasedLikelihoodEstimator(method='ML')
-vb_ml.fit(dset)
+vb_ml.fit_dataset(dset)
 vb_ml_summary = vb_ml.summary()
 results['Variance-Based with ML'] = vb_ml_summary.to_df()
 print('Variance-Based with ML')
 print(vb_ml_summary.to_df().T)
 print()
 
 vb_reml = pymare.estimators.VarianceBasedLikelihoodEstimator(method='REML')
-vb_reml.fit(dset)
+vb_reml.fit_dataset(dset)
 vb_reml_summary = vb_reml.summary()
 results['Variance-Based with REML'] = vb_reml_summary.to_df()
 print('Variance-Based with REML')
@@ -219,15 +215,15 @@ def var_to_ci(y, v, n):
 # using ``y`` and ``n``, but assumes within-study variance is homogenous
 # across studies.
 sb_ml = pymare.estimators.SampleSizeBasedLikelihoodEstimator(method='ML')
-sb_ml.fit(dset)
+sb_ml.fit_dataset(dset)
 sb_ml_summary = sb_ml.summary()
 results['Sample Size-Based with ML'] = sb_ml_summary.to_df()
 print('Sample Size-Based with ML')
 print(sb_ml_summary.to_df().T)
 print()
 
 sb_reml = pymare.estimators.SampleSizeBasedLikelihoodEstimator(method='REML')
-sb_reml.fit(dset)
+sb_reml.fit_dataset(dset)
 sb_reml_summary = sb_reml.summary()
 results['Sample Size-Based with REML'] = sb_reml_summary.to_df()
 print('Sample Size-Based with REML')