Peter G. N. Scheyen
1993
Evaluation of TTP Parser: A Preliminary Report
Tomek Strzalkowski
|
Peter G. N. Scheyen
Proceedings of the Third International Workshop on Parsing Technologies
TTP (Tagged Text Parser) is a fast and robust natural language parser specifically designed to process vast quantities of unrestricted text. TTP can analyze written text at the speed of approximately 0.3 sec/sentence, or 73 words per second. An important novel feature of TTP parser is that it is equipped with a skip-and-fit recovery mechanism that allows for fast closing of more difficult sub-constituents after a preset amount of time has elapsed without producing a parse. Although a complete analysis is attempted for each sentence, the parser may occasionally ignore fragments of input to resume “normal” processing after skipping a few words. These fragments are later analyzed separately and attached as incomplete constituents to the main parse tree. TTP has recently been evaluated against several leading parsers. While no formal numbers were released (a formal evaluation is planned later this year), TTP has performed surprisingly well. The main argument of this paper is that TTP can provide a substantial gain in parsing speed giving up relatively little in terms of the quality of output it produces. This property allows TTP to be used effectively in parsing large volumes of text.