Although there have been a great ma

Although there have been a great many applications of machine learning in text class ification for information retrieval, there have been relatively few within an e ducational context, and most of them have been aimed at inferring dialog acts , for example [14]. More similar approaches to ours inc lude Larkey’s [15] com parison of k- nearest neighbor, naıve Bayes and linear regression classifiers in as signing grades to student essays. Sathiyamurthy and Geetha [16] built a text classification system which used part-of-speech tagging to aligne- learning
documents acc ording to an ACM domain ontology, allowing the documents to be classified according to Bloom’s taxonomy [17]. Yilm aze let al [18] used an SVM algorithm to perform text categorization for automatically aligning curricular documents with state and federal science benchmarks.