A Graph-based Readability Assessment Method using Word Coupling

Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 411–420, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. A Graph-based Readability Assessment Method using Word Coupling Zhiwei Jiang, Gang Sun, Qing Gu∗, Tao Bai, Daoxu Chen State Key Laboratory for Novel Software Technology Nanjing University, Nanjing 210023, China jiangzhiwei@outlook.com, sungangnju@163.com, guq@nju.edu.cn, bt@xjau.edu.cn, cdx@nju.edu.cn Abstract This paper proposes a graph-based read- ability assessment method using word coupling. Compared to the state-of-the- art methods such as the readability for- mulae, the word-based and feature-based methods, our method develops a coupled bag-of-words model which combines the merits of word frequencies and text fea- tures. Unlike the general bag-of-words model which assumes words are indepen- dent, our model correlates the words based on their similarities on readability. By applying TF-IDF (Term Frequency and Inverse Document Frequency), the cou- pled TF-IDF matrix is built, and used in the graph-based classification framework, which involves graph building, merging and label propagation. Experiments are conducted on both English and Chinese datasets. The results demonstrate both ef- fectiveness and potential of the method. 1 Introduction Readability assessment is a task that aims to eval- uate the reading difficulty or comprehending easi- ness of text documents. It is helpful for education- ists to select texts appropriate to the reading/grade levels of the students, and for web designers to or- ganize texts on web pages for the users doing per- sonalized searches for information retrieval. Research on readability assessment starts from the early 20th century (Dale and Chall, 1948). Many useful readability formulae have been devel- oped since then (Dale and Chall, 1948; McLaugh- lin, 1969; Kincaid et al., 1975). Currently, due to the development of natural language processing, the methods on readability assessment have made a great progress (Zakaluk and Samuels, 1988; ∗Corresponding author. Benjamin, 2012; Gonzalez-Dios et al., 2014). The word-based methods compute word frequencies in documents to estimate their readability (Collins- Thompson and Callan, 2004; Kidwell et al., 2009). The feature-based methods extract text features from documents and train classification models to classify the readability (Schwarm and Osten- dorf, 2005; Feng et al., 2010; Franc¸ois and Fairon, 2012; Hancke et al., 2012). In this paper, we propose a graph-based method using word coupling, which combines the mer- its of both word frequencies and text features for readability assessment. We design a cou- pled bag-of-words model, which correlates words based on their similarities on sentence-level read- ability computed using text features. The model is used in a graph-based classification frame- work, which involves graph building, graph merg- ing/combination, and label propagation. We per- form experiments on datasets of both English and Chinese. The results demonstrate both effective- ness and potential of our method. The rest of this paper is organized as follows: Section 2 introduces backgrounds of our work. Section 3 presents the details of the method. Sec- tion 4 designs the experiments and explains the re- sults. Finally, Section 5 concludes the paper with planned future work. 2 Background In this section, …

Content truncated to 500 words