The feasibility of auto profiling online

Pienemann, M; Lin, B; Chung, L

doi:10.1007/978-3-642-02937-0_17

The feasibility of auto profiling online

Lookup NU author(s): Professor Manfred Pienemann, Bi-jar Lin

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

In this paper we will demonstrate the feasibility of using Auto Profiling (AP) online to analyse an informant's linguistic data and his/ her language distributions. AP is an extension of the linguistic profiling approach that was originally carried out manually. This procedure was further developed into a computer-based procedure (Rapid Profile) (RP) that assists the analyst, although this is still essentially a manual observation-based procedure. The automatic profiling approach presented in this work is the first fully automatic procedure for linguistic profiling. It accepts free-style written interlanguage input and automatically assigns a C-structure and aspects of the L2 morphology to the interlanguage input. On this basis the system is capable of reliably assessing the stage of acquisition of the input sample. In addition, it gives feedback to the specific errors contained in the input, and can also be used for automatically tagging L2 corpora. At the end of this paper, we will provide an analysis of real data to demonstrate that the Auto Profiling algorithm is not only based on the application of Processability Theory (PT) to Specific L2s and on the logic of RP, but also can be used online to profile an informant's data in real time under time constraints. © 2009 Springer-Verlag Berlin Heidelberg.