Abstract:

With the development of technology, the whole world is like a global village and English plays an important role as a global language in the world. Therefore, enhancing our English level not only improves our communication skill for international competition, but also becomes a tool to learn knowledge. Currently, however, there is still no reliable tool to judge the readability of English article to help learners select proper articles to read. So, here we present a tool to analyze the readability of English article with Bayesian Networks (BNs) and adapt the GEPT’s vocabulary difficulty levels and the amount of words per sentence as the parameters of this tool. The resource of our study comes from 110 GEPT’s Elementary, Intermediate and High-Intermediate level’s reading comprehension articles. We use these articles to build our Bayesian Networks (BNs) structure and compare the effectiveness with Flesch-Kincaid Reading Ease and KNN algorithm. The result indicates that our proposed method is more effective than the other tool in predicting the readability of English article.

Keywords:

Bayesian Networks(BNs), Readability, GEPT, Flesch-Kincaid Reading Ease, KNN algorithm.