A Hybrid Approach to Content Analysis for Automatic Essay Grading

We present CarmelTC, a novel hybrid text classification approach for automatic essay grading. Our evaluation demonstrates that the hybrid CarmelTC approach outperforms two “bag of words”approaches, namely LSA anda Naive Bayes, as well as a purely symbolic approach.

• Type: Conference
• Year: 2003
• Where: NAACL
• Authors: Carolyn Penstein Rosé, Antonio Roque, Dumisizwe Bhembe, Kurt VanLehn
• Keywords: Anda Naive Bayes, Hybrid Carmeltc Approach, NAACL 2003, NAACL 2007, Text Classification Approach