... popularity as they constantly outper-form other learning algorithms for many NLP tasks.Unfortunately, once a model is trained, the de-cision function for kernel-based classifiers such asSVM ... Yamcha)and transforms them into our format. The trans-formed models can then be used by our efficient Javaimplementation of the method described in this pa-per. We supply wrappers for the interfaces ... Ohio, USA, June 2008.c2008 Association for Computational LinguisticssplitSVM: Fast, Space-Efficient, non-Heuristic, Polynomial KernelComputation for NLP Applications Yoav Goldberg and Michael...