This work presents Entropy Guided Transformation Learning (ETL), a new machine learning algorithm for classification tasks.
ETL generalizes Transformation Based Learning (TBL) by automatically solving the TBL bottleneck: the construction of good
template sets. ETL uses the information gain in order to select the feature combinations that provide good template sets.
We describe the application of ETL to two language independent Text Mining preprocessing tasks: part-of-speech tagging and
phrase chunking. We also report our findings on one language independent Information Extraction task: named entity recognition.
Overall, we successfully apply it to six different languages: Dutch, English, German, Hindi, Portuguese and ...
Read Full Article