Heterogeneous Natural Language Processing Tools via Language Processing Chains

One of the most recent developments in NLP is the emergence of linguistic annotation metasystems which make use of existing processing tools and implement pipelined architecture. In this paper we describe a system that offers a new perspective in exploiting NLP meta-systems by providing a common processing framework. This framework supports most of common NLP tasks by chaining tools that are able to communicate on the basis of common formats. As a demonstration of the effectiveness of the system to manage heterogeneous NLP tools, we developed an English processing chain, pipelining OpenNLP-based and C++ NLP implementations. Furthermore, we conducted experiments to test the stability and measure the performance of the English processing chain. A baseline processing chain for the Bulgarian language illustrates the capabilities of the system to support and manage processing chains for more languages.