Update on error plotting posterior_predictive plots and the Gelman-Rubin statistic
AT
Hello HDDM users,
I am new to hddm and am working through the hddm tutorial posted on http://ski.clps.brown.edu/hddm_docs/tutorial_python.html on a Windows 10 Dell laptop. I am hoping someone can shed light on 2 different error messages that I am getting – one with the Gelman-Rubin statistic and the second when plotting posterior_predictive plots.
The Gelman-Rubin Stat error:
The code works OK until I get to instantiate the model step. At this point I get a RuntimeWarning message, but the code was harmless and prints the expected output. I saw in a previous conversation that someone else had asked about the RuntimeWarning error and was told not to worry about the warning as it is harmless.
C:\Users\athapar\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\scipy\optimize\optimize.py:2116: RuntimeWarning: invalid value encountered in double_scalars
tmp2 = (x - v) * (fx - fw)
I then run the code for the Gelman-Rubin Statistic. I again get the error message but at this point, the code behaves oddly - it goes through and completes the run with the following error message RuntimeWarning. But then instead of printing the output shown in the tutorial, it repeats the loop, and then it begins to repeat it again (see below - I actually halted the third run)
C:\Users\athapar\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\scipy\optimize\optimize.py:2116: RuntimeWarning: invalid value encountered in double_scalars tmp2 = (x - v) * (fx - fw)
[-----------------100%-----------------] 5000 of 5000 complete in 545.1 sec
C:\Users\athapar\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\scipy\optimize\optimize.py:2116: RuntimeWarning: invalid value encountered in double_scalars tmp2 = (x - v) * (fx - fw)
[-----------------100%-----------------] 5000 of 5000 complete in 537.2 sec
C:\Users\athapar\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\scipy\optimize\optimize.py:2116: RuntimeWarning: invalid value encountered in double_scalars tmp2 = (x - v) * (fx - fw)
[ 1% ] 95 of 5000 complete in 10.0 secHalting at iteration 94 of 5000
The plot_posterior_predictive()
The second issue is with the plot_posterior_predictive(m.plot_posterior_predictive(figsize=(14, 10))) function.
The full error message for the plot_posterior_predictive() is provided below. I get a plot which shows a red plot that looks as one would expect but a flat blue plot. I also tried this with a different dataset from the hddm package and again the plot_posterior_predictive() call resulted in the same error message and a plot that displays a red line and a flat blue line.
Thanks, Anjali
C:\Users\athapar\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\kabuki\analyze.py:556: MatplotlibDeprecationWarning: Passing non-integers as three-element position specification is deprecated since 3.3 and will be removed two minor releases later. ax = fig.add_subplot(np.ceil(nrows), columns, subj_i+1)
AttributeError Traceback (most recent call last)<ipython-input-34-916850d95b91> in <module>----> 1 m.plot_posterior_predictive(figsize=(14, 10)) ~\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\hddm\models\base.py in plot_posterior_predictive(self, *args, **kwargs) 766 if 'value_range' not in kwargs: 767 kwargs['value_range'] = np.linspace(-5, 5, 100)--> 768 kabuki.analyze.plot_posterior_predictive(self, *args, **kwargs) 769 770 def plot_posterior_quantiles(self, *args, **kwargs): ~\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\kabuki\analyze.py in plot_posterior_predictive(model, plot_func, required_method, columns, save, path, figsize, format, num_subjs, **kwargs) 558 ax.set_title(str(bottom_node['subj_idx'])) 559 --> 560 plot_func(bottom_node['node'], ax, **kwargs) 561 562 if num_subjs is not None and i >= num_subjs: ~\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\kabuki\analyze.py in _plot_posterior_pdf_node(bottom_node, axis, value_range, samples, bins) 478 axis.hist(bottom_node.value.values, normed=True, color='r', 479 range=(value_range[0], value_range[-1]), label='data',--> 480 bins=bins, histtype='step', lw=2.) 481 482 axis.set_ylim(bottom=0) # Likelihood and histogram can only be positive ~\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\matplotlib\__init__.py in inner(ax, data, *args, **kwargs) 1445 def inner(ax, *args, data=None, **kwargs): 1446 if data is None:-> 1447 return func(ax, *map(sanitize_sequence, args), **kwargs) 1448 1449 bound = new_sig.bind(ax, *args, **kwargs) ~\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\matplotlib\axes\_axes.py in hist(self, x, bins, range, density, weights, cumulative, bottom, histtype, align, orientation, rwidth, log, color, label, stacked, **kwargs) 6813 if patch: 6814 p = patch[0]-> 6815 p.update(kwargs) 6816 if lbl is not None: 6817 p.set_label(lbl) ~\AppData\Local\Continuum\anaconda3\envs\py36\lib\site-packages\matplotlib\artist.py in update(self, props) 994 func = getattr(self, f"set_{k}", None) 995 if not callable(func):--> 996 raise AttributeError(f"{type(self).__name__!r} object " 997 f"has no property {k!r}") 998 ret.append(func(v)) AttributeError: 'Polygon' object has no property 'normed'
