libTLDA
libTLDA copied to clipboard
wmkouw
Library of transfer learners and domain-adaptive classifiers.
The `cross_val_predict` method used for domain discrimination in the instance weighting class outputs classes (not probabilities) by default. It appears the intended behavior requires passing `method="predict_proba"` to the `cross_val_predict` method....
https://github.com/wmkouw/libTLDA/blob/0c66ec2327d191b88fca0803a7c74b0bb05afd42/libtlda/iw.py#L290 ix = np.argmin(d, axis=1) To count target samples within each source Voronoi cell, ix should be the index of the source sample if there exists a target sample that...
X: memory usage: 705.4+ MB, Z: memory usage: 350.5+ MB, I encountered this error ( MemoryError: Unable to allocate 1010. GiB for an array with shape (368104, 368104) and data...
> Increase the number of iterations (max_iter) or scale the data as shown in: > https://scikit-learn.org/stable/modules/preprocessing.html > Please also refer to the documentation for alternative solver options: > https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression >...
PackagesNotFoundError when I try ro install the package.
Running the StructuralCorrespondenceClassifier.fit() function returns : ValueError: Complex data not supported My source domain training set is a BoW representation of separate emails. The same holds for the target domain...
When I use ImportanceWeightedClassifier (IW) iwe=kda, the console shows that C:\Users\Administrator\Anaconda3\envs\libtlda\lib\site-packages\scipy\stats\kde.py:262: RuntimeWarning: invalid value encountered in true_divide result = result / self._norm_factor Besides, FLDA also has this problem (divided by...
A couple of people have sent emails that they expected certain scikit methods to be present, e.g. predict_proba. Plan is to make sure that all my classifiers have at least:...
A few classifiers will be added soon and some restructuring will be necessary. The plan at this point is to form the following 4 modules: - iwm (importance-weighted models) -...
Gong, Zhang, Liu, Tao, Glymour, Schölkopf. Domain adaptation with conditional transferable components. ICML, 2016. Code: http://mingminggong.xyz/publications.html