Native Language Identification on Text and Speech

TL;DR

This paper introduces an ensemble SVM-based system for native language identification using text and speech data, achieving high accuracy and ranking third in a shared task competition.

Contribution

It presents a novel ensemble approach combining multiple SVM classifiers trained on character n-grams for native language identification from both text and speech.

Findings

Achieved 83.58% accuracy in the NLI shared task

Ranked 3rd among participating teams

Effective use of character n-grams for classification

Abstract

This paper presents an ensemble system combining the output of multiple SVM classifiers to native language identification (NLI). The system was submitted to the NLI Shared Task 2017 fusion track which featured students essays and spoken responses in form of audio transcriptions and iVectors by non-native English speakers of eleven native languages. Our system competed in the challenge under the team name ZCD and was based on an ensemble of SVM classifiers trained on character n-grams achieving 83.58% accuracy and ranking 3rd in the shared task.