add CI bound length for sensitivity results

Author: PhilippBach

scripts/irm/irm_apo_sensitivity.py

@@ -114,6 +114,7 @@
                     causal_contrast_model.sensitivity_analysis(cf_y=cf_y, cf_d=cf_d, rho=rho, level=level ,null_hypothesis=theta)
                     cover_lower = theta >= causal_contrast_model.sensitivity_params['ci']['lower']
                     cover_upper = theta <= causal_contrast_model.sensitivity_params['ci']['upper']
+                    ci_bound_width = causal_contrast_model.sensitivity_params['ci']['upper'][0]- causal_contrast_model.sensitivity_params['ci']['lower'][0]
                     rv = causal_contrast_model.sensitivity_params['rv']
                     rva = causal_contrast_model.sensitivity_params['rva']
                     bias_lower = abs(theta - causal_contrast_model.sensitivity_params['theta']['lower'])
@@ -135,6 +136,7 @@
                         "RVa": rva,
                         "Bias (Lower)": bias_lower,
                         "Bias (Upper)": bias_upper,
+                        "CI Bound Length": ci_bound_width,
                         "Learner g": learner_g_name,
                         "Learner m": learner_m_name,
                         "level": level,
@@ -156,6 +158,7 @@
          "RVa": "mean",
          "Bias (Lower)": "mean",
          "Bias (Upper)": "mean",
+         "CI Bound Length": "mean",
          "repetition": "count"}
     ).reset_index()
 print(df_results)

scripts/irm/irm_ate_sensitivity.py

@@ -101,6 +101,7 @@
                 dml_irm.sensitivity_analysis(cf_y=cf_y, cf_d=cf_d, rho=rho, level=level, null_hypothesis=theta)
                 cover_lower = theta >= dml_irm.sensitivity_params['ci']['lower']
                 cover_upper = theta <= dml_irm.sensitivity_params['ci']['upper']
+                ci_bound_width = dml_irm.sensitivity_params['ci']['upper'][0]- dml_irm.sensitivity_params['ci']['lower'][0]
                 rv = dml_irm.sensitivity_params['rv']
                 rva = dml_irm.sensitivity_params['rva']
                 bias_lower = abs(theta - dml_irm.sensitivity_params['theta']['lower'])
@@ -118,6 +119,7 @@
                         "RVa": rva,
                         "Bias (Lower)": bias_lower,
                         "Bias (Upper)": bias_upper,
+                        "CI Bound Length": ci_bound_width,
                         "Learner g": learner_g_name,
                         "Learner m": learner_m_name,
                         "level": level,
@@ -135,6 +137,7 @@
          "RVa": "mean",
          "Bias (Lower)": "mean",
          "Bias (Upper)": "mean",
+         "CI Bound Length": "mean",
          "repetition": "count"}
     ).reset_index()
 print(df_results)