Conclusions

Abstract

Entropy guided transformation learning is a machine learning algorithm for classification tasks. In this book, we detail how ETL generalizes transformation based learning by solving the TBL bottleneck: the construction of good template sets. ETL relies on the use of the information gain measure to select feature combinations that provide effective template sets. In this work, we also present ETL committee, an ensemble method that uses ETL as the base learner. We describe the application of ETL to four language independent NLP tasks: part-of-speech tagging, phrase chunking, named entity recognition and semantic role labeling. Overall, we successfully apply it to thirteen different corpora in six different languages: Dutch, English, German, Hindi, Portuguese and Spanish. Our extensive experimental results demonstrate that ETL is an effective way to learn accurate transformation rules. In all experiments, ETL shows better results than TBL with hand-crafted templates. Our experimental results also demonstrate that ETL Committee is an effective way to improve the ETL effectiveness. We believe that by avoiding the use of handcrafted templates, ETL enables the use of transformation rules to a greater range of classification tasks.