16px-feed-icon Veröffentlichungen View this page in English



Automotive Quality analysis is mainly based on structured information like damage and part codes, which are gathered in dealerships. Besides this structured data, there is usually a free-form text containing the voice of the customer (VoC), which can be used for example for industrial quality methods like Six Sigma. The major challenges in processing the textual information are the specialized technical vocabulary, the unusual syntactic grammar and the textual quality (misspellings, abbreviations, omission of function words). In this paper we will present a system consisting of NLP modules to face these challenges and to extract the relations between components, symptoms and other concepts using unsupervised pos-tagging and unsupervised parsing.

Type: Inproceedings

Author: Hänig, C. and Schierle, M.
Title: Relation Extraction based on Unsupervised Syntactic Parsing
Booktitle: Proceedings of the conference on Text Mining Services (TMS 2009)
Year: 2009
AUTHOR = {Hänig, C. and Schierle, M.},
TITLE = {Relation Extraction based on Unsupervised Syntactic Parsing},
BOOKTITLE = {Proceedings of the conference on Text Mining Services (TMS 2009)},
YEAR = {2009}