small doc changes

docs/Compatibility with scikit-learn.rst

@@ -8,8 +8,7 @@ Compatibility with scikit-learn
 
 New to lifelines in version 0.21.3 is a wrapper that allows you to use lifeline's regression models with scikit-learn's APIs.
 
-.. note:: The X variable still needs to be a DataFrame, and should contains the event-occurred column (``event_col``) if it exists.
-
+.. note:: the API and functionality is still experimental. Please report any bugs or features on our Github issue list.
 
 
 .. code:: python
@@ -19,7 +18,7 @@ New to lifelines in version 0.21.3 is a wrapper that allows you to use lifeline'
     from lifelines import CoxPHFitter
     from lifelines.datasets import load_rossi
 
-    X = load_rossi().drop('week', axis=1)
+    X = load_rossi().drop('week', axis=1) # keep as a dataframe
     Y = load_rossi().pop('week')
 
     CoxRegression = sklearn_adapter(CoxPHFitter, event_col='arrest')
@@ -30,13 +29,16 @@ New to lifelines in version 0.21.3 is a wrapper that allows you to use lifeline'
     print(sk_cph)
 
     """
-    CoxPHFitter(alpha=0.05, penalizer=1.0, strata=None, tie_method='Efron')
+    SkLearnCoxPHFitter(alpha=0.05, penalizer=1.0, strata=None, tie_method='Efron')
     """
 
     sk_cph.predict(X)
     sk_cph.score(X, Y)
 
 
+.. note:: The X variable still needs to be a DataFrame, and should contains the event-occurred column (``event_col``) if it exists.
+
+
 If needed, the original lifeline's instance is available as the ``lifelines_model`` attribute.
 
 .. code:: python

docs/Survival Regression.rst

@@ -929,7 +929,7 @@ Model selection based on predictive power
 
 If censoring is present, it's not appropriate to use a loss function like mean-squared-error or
 mean-absolute-loss. Instead, one measure is the concordance-index, also known as the c-index. This measure
-evaluates the accuracy of the ordering of predicted time. It is in fact a generalization
+evaluates the accuracy of the *ranking* of predicted time. It is in fact a generalization
 of AUC, another common loss function, and is interpreted similarly:
 
 * 0.5 is the expected result from random predictions,
@@ -959,6 +959,8 @@ Fitted survival models typically have a concordance index between 0.55 and 0.75
     from lifelines.utils import concordance_index
     print(concordance_index(rossi['week'], -cph.predict_partial_hazard(rossi), rossi['arrest']))
 
+.. note:: Remember, the concordance score evaluates the relative rankings of subject's event times. Thus, it is scale invariant (i.e. you can multiple by a positive constant, or add a constant, and the rankings won't change). A model maximized for concordance-index does not necessarily give good predicted *times*, but will give good predicted *rankings*.
+
 
 However, there are other, arguably better, methods to measure the fit of a model. Included in ``print_summary`` is the log-likelihood, which can be used in an `AIC calculation <https://en.wikipedia.org/wiki/Akaike_information_criterion>`_, and the `log-likelihood ratio statistic <https://en.wikipedia.org/wiki/Likelihood-ratio_test>`_. Generally, I personally loved this article by Frank Harrell, `"Statistically Efficient Ways to Quantify Added Predictive Value of New Measurements" <http://www.fharrell.com/post/addvalue/>`_.
 

docs/conf.py

@@ -15,6 +15,7 @@
 from datetime import date
 import sys
 import os
+import lifelines
 
 # If extensions (or modules to document with autodoc) are in another directory,
 # add these directories to sys.path here. If the directory is relative to the
@@ -40,6 +41,7 @@
 
 exclude_patterns = ["_build", "jupyter_notebooks/.ipynb_checkpoints/*.ipynb"]
 
+
 # Add any paths that contain templates here, relative to this directory.
 templates_path = ["_templates"]
 
@@ -63,8 +65,9 @@
 #
 # The short X.Y version.
 
-version = "0.21.3"
-# The full version, including dev info
+# The short X.Y version.
+version = lifelines.__version__
+# The full version, including alpha/beta/rc tags.
 release = version
 
 # The language for content autogenerated by Sphinx. Refer to documentation

docs/lifelines.fitters.rst

@@ -112,7 +112,7 @@ lifelines.fitters.log\_logistic\_aft\_fitter module
 
 
 lifelines.fitters.piecewise\_exponential\_regression\_fitter module
------------------------------------------------------
+----------------------------------------------------------------------
 
 .. automodule:: lifelines.fitters.piecewise_exponential_regression_fitter
     :members: