label a few points

labeled_N = 4
ys = np.array([-1]*len(ytrue)) # -1 denotes unlabeled point

Are unlabeled points a workaround for train/test data points?
Ie, model.fit(x_trn, y_trn). model.score(x_tst, y_tst)

Traceback (most recent call last):

File "/project/peaclab-mon/pyenv/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 3319, in run_code
exec(code_obj, self.user_global_ns, self.user_ns)

File "", line 1, in
from frameworks.CPLELearning import CPLELearningModel

File "/usr2/collab/esencan/semisup-learn/frameworks/CPLELearning.py", line 119
except Exception, e:
^
SyntaxError: invalid syntax

It seems like

self.plattlr.predict_proba

Is called (line 63, within self.predict_proba) before self.plattlr is set on line 78.

This causes models like a generic SVC to fail:

AttributeError: 'SelfLearningModel' object has no attribute 'plattlr'

This is what I got after run the heart dataset example, seems, the semi-supervised learning does not help. I did not edit anything though, not sure where this inconsistency comes from.
supervised log.reg. score 0.555555555556
self-learning log.reg. score 0.377777777778
..n....n....n.n max_iter exceeded.
CPLE semi-supervised log.reg. score 0.555555555556
..nn..n...n..n. max_iter exceeded.
CPLE semi-supervised RBF SVM score 0.555555555556

Could you please update a new version that is compatible with Python 3? Thx!

I got a wrong of " No module named 'frameworks'", but when I used “pip” to install the "frameworks",
return the "SyntaxError: invalid syntax", is "frameworks" not a module? do everyone have some suggestions? Thanks!

I believe this should be scikit-learn instead of sklearn

I get the following error, why.....?
'QN_S3VM_Dense' object has no attribute '_QN_S3VM_Dense__kernel'

ATT, although a simple

python setup.py build
python setup.py install

seems work.

What about the robustness of 'SelfLearningModel' and 'CPLELearningModel'? when I tested them on "iris" and "leukemia" datasets, the performance of these models was much bad than the supervised model, which very confused me.

This model use the labeled samples to train , but when test the acc , this model use the labeled samples again. Is the acc would higher?
I'm a newer to machine-learning.
I think the score function can't use all data ,but should use the data which not be trained.
Looking forward for your reply.

when doing

from frameworks.CPLELearning import CPLELearningModel

I get

File "/home/MYUSERNAME/anaconda3/lib/python3.6/site-packages/semisup_learn-0.0.1-py3.6.egg/frameworks/CPLELearning.py", line 119 except Exception, e: ^ SyntaxError: invalid syntax

I am using python 3.6 and have installed the module by

python setup.py install

I find this project very interesting, and applaud the code. Though, I ended up re-writing many parts due to some typos in python 3, and have been unable to find a missing qns3vm module.

maybe this is all due to the fact that I'm in Python 3.

good work, otherwise though 👍

it has an error when running the example.py

urllib2.HTTPError: HTTP Error 404: Dataset 'lung-cancer-ontario' not found on mldata.org.

I have used the CPLELearningModel to solve the 2-class unlabled data successfully, but when I try to use it to solve multi-classes problem, I get error, so I wonder does this model support semi-supervision on multi-classes?