gregversteeg / corex Goto Github PK

https://github.com/gregversteeg/CorEx
Basic Usage Example needs "import numpy as np"

changed my version of corex.py
#from scipy.misc import logsumexp # Tested with 0.13.0
from scipy.special import logsumexp # Tested with 1.16.0

If I understand correctly, logsumexp was moved from scipy.misc to scipy.special, which is raising the error mentioned in the title when trying to import corex?

One of the failing message, for line 264 of corex.py is copied below, I am using python 3.5.1

print "Warning: Data matrix values should be consecutive integers starting with 0,1,..."
^
SyntaxError: Missing parentheses in call to 'print'

Hi,

Any progress on CorEx with continuous variables? It would be great as most of my data is continuous variables.

Thanks for the great work!

Hi, CorEx is a great work! And I am recently reading the paper , and I get trouble in understanding the optimization part in Sec. A.
The original objective is like this:

And because of:

Then:

Why the red-box part can be extracted? I think $p(x)p(y| x)$ is different from $p(x_i)p(y| x_i)$. Can you offer more details? Thanks!

Working on a Mac with anaconda. Did not have sfdp.
Did conda install graphviz
Now get the trianglation library missing below.
Suggestions appreciated.

(base) katherinepaseman@Bills-MacBook-Pro-3 bio_corex % python vis_corex.py data/test_data.csv
vis_corex.py:719: DeprecationWarning: 'U' mode is deprecated
with open(filename, 'rU') as csvfile:
Time for first layer: 0.05
TC at layer 0 is: 2.078
TC at layer 1 is: -0.000
Groups in sorted_groups.txt
Pairwise plots among high TC variables in "relationships"
/opt/anaconda3/lib/python3.7/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.
warnings.warn(msg, UserWarning)
/opt/anaconda3/lib/python3.7/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.
warnings.warn(msg, UserWarning)
/opt/anaconda3/lib/python3.7/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.
warnings.warn(msg, UserWarning)
/opt/anaconda3/lib/python3.7/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.
warnings.warn(msg, UserWarning)
/opt/anaconda3/lib/python3.7/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.
warnings.warn(msg, UserWarning)
/opt/anaconda3/lib/python3.7/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.
warnings.warn(msg, UserWarning)
weight threshold is -0.000000 for graph with max of 12.000000 edges
non-isolated nodes,edges 6 4
Error: remove_overlap: Graphviz not built with triangulation library
Error: remove_overlap: Graphviz not built with triangulation library
non-isolated nodes,edges 6 4
Error: remove_overlap: Graphviz not built with triangulation library
Error: remove_overlap: Graphviz not built with triangulation library
(base) katherinepaseman@Bills-MacBook-Pro-3 bio_corex %

Hi,
I'd like to apply your method to my research problem (computational systems biology).
I'm trying to run fit method with my data (2d array, dim: 30 x 13800) and keep getting the following error:

Warning: Data matrix values should be consecutive integers starting with 0,1,...
Traceback (most recent call last):
  File "projectrepo/green/corex_fernandez/test.py", line 17, in <module>
    lv31.fit(arx)
  File "/home/fewpills/projectrepo/green/corex_fernandez/corex.py", line 202, in fit
    self.fit_transform(X)
  File "/home/fewpills/projectrepo/green/corex_fernandez/corex.py", line 229, in fit_transform
    self.update_marginals(X_event, self.p_y_given_x)  # Eq. 8
  File "/home/fewpills/projectrepo/green/corex_fernandez/corex.py", line 289, in update_marginals
    self.log_marg = self.calculate_p_y_xi(X_event, p_y_given_x) - self.log_p_y
  File "/home/fewpills/projectrepo/green/corex_fernandez/corex.py", line 299, in calculate_p_y_xi
    pseudo_counts = 0.001 + np.dot(X_event, p_y_given_x).transpose((1,0,2))  # n_hidden, n_events, dim_hidden
MemoryError

I've tried to run it both with python 2.7 and 3.4 as a script.
I'd be grateful for any suggestions.

I get the following warning when running your example with four patterns having five variables:

/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/numpy/core/_methods.py:59: RuntimeWarning: Mean of empty slice.
warnings.warn("Mean of empty slice.", RuntimeWarning)

Wondering if this is serious or can be safely ignored.

And by the way, does this implementation support sparse matrices in input?

Re 2015 paper, I don't see the S&P components in any data directory; and quantquote does not seem to have it. Showing me a data source would be helpful.
Thanks.