Cross validation script

[manics]

Adds a cross-validation script, based on https://github.com/wnd-charm/wnd-charm/blob/master/examples/classifier_cross_validation.py, see #18.

Requires wnd-charm/wnd-charm#26

Note the cross validation script is curently broken:

In [17]: e.PredictedValueAnalysis()
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-17-8ed307c64bfa> in <module>()
----> 1 e.PredictedValueAnalysis()

/Users/simon/omero-test/venv/lib/python2.7/site-packages/pychrm-0.2-py2.7-macosx-10.8-x86_64.egg/pychrm/FeatureSet.pyc in print_method_wrapper(*args, **kwargs)
    201                                 sys.stdout = backup
    202                 else:
--> 203                         retval = method_that_prints_output( *args, **kwargs )
    204                 return retval
    205

/Users/simon/omero-test/venv/lib/python2.7/site-packages/pychrm-0.2-py2.7-macosx-10.8-x86_64.egg/pychrm/FeatureSet.pyc in PredictedValueAnalysis(self)
   4062                         vals = np.array( [result.predicted_value for result in self.accumulated_individual_results[filename] ])
   4063                         self.ground_truth_values.append( self.accumulated_individual_results[filename][0].ground_truth_value )
-> 4064                         self.predicted_values.append( np.mean(vals) )
   4065                         self.individual_stats[filename] = ( len(vals), np.min(vals), np.mean(vals), \
   4066                                                             np.max(vals), np.std(vals) )

/usr/local/lib/python2.7/site-packages/numpy/core/fromnumeric.pyc in mean(a, axis, dtype, out, keepdims)
   2486
   2487     return _methods._mean(a, axis=axis, dtype=dtype,
-> 2488                             out=out, keepdims=keepdims)
   2489
   2490

/usr/local/lib/python2.7/site-packages/numpy/core/_methods.pyc in _mean(a, axis, dtype, out, keepdims)
     55                         out=ret, casting='unsafe', subok=False)
     56     else:
---> 57         ret = ret / float(rcount)
     58     return ret
     59

TypeError: unsupported operand type(s) for /: 'NoneType' and 'float'

[manics]

@colettace This is based on the cross-validation script you suggested, on my test data PredictedValueAnalysis() throws an error (which is why I've temporarily put a try-catch around it).