Detecting English-French Cognates Using Orthographic Edit Distance

Detecting English-French Cognates Using Orthographic Edit Distance Qiongkai Xu1,2, Albert Chen1, Chang Li1 1The Australian National University, College of Engineering and Computer Science 2National ICT Australia, Canberra Research Lab {xuqiongkai, u5708995, spacegoing}@gmail.com Abstract Identification of cognates is an important component of computer assisted second language learning systems. We present a simple rule-based system to recognize cognates in English text from the perspec- tive of the French language. At the core of our system is a novel similarity measure, orthographic edit distance, which incorpo- rates orthographic information into string edit distance to compute the similarity be- tween pairs of words from different lan- guages. As a result, our system achieved the best results in the ALTA 2015 shared task. 1 Introduction Cognates words are word pairs that are similar in meaning, spelling and pronunciation between two languages. For example, “age” in English and “ˆage” in French are orthographically similar while ”father” in English and ”Vater” in German are phonetically similar. There are three types of cognates: true cognates, false cognates and semi- cognates. True cognates may have similar spelling or pronunciation but they are mutual translations in any context. False cognates are orthographi- cally similar but have totally different meanings. Semi-cognates are words that have the same mean- ing in some circumstances but a different mean- ing in other circumstances. Finding cognates can help second language learners leverage their back- ground knowledge in their first language, thus im- proving their comprehension and expanding their vocabulary. In this paper, we propose an automatic method to identify cognates in English and French with the help of the Google Translator API1. Our method calculates the similarity of two words based solely 1https://code.google.com/p/google-api-translate-java/ on the sequences of characters involved. After ex- ploring n-gram similarity and edit distance sim- ilarity, we propose an orthographic edit distance similarity measure which leverages orthographic information from source language to target lan- guage. Our approach achieved first place in the ALTA 2015 shared task. 2 Related Work There are many ways to measure the similarity of words from different languages. Most popu- lar ones are surface string based similarity, i.e. n- gram similarity and edit distance. An n-gram is a contiguous sequence of n items, normally letters, from a given sequence. There are many popular measures that use n-grams such as DICE (Brew et al., 1996), which uses bi-grams, and Longest Common Subsequence Ratio (LCSR) (Melamed, 1999). LCSR was later found to be a special case of n-gram similarity by Kondrack (Kondrak, 2005), who developed a general n-gram frame- work. He provided formal, recursive definitions of n-gram similarity and distance, together with efficient algorithms for computing them. He also proved that in many cases, using bi-grams is more efficient than using other n-gram methods. Since LCSR is only a tri-gram measure, using bi- gram similarity and distance can easily outperform LCSR in many cases. Instead of computing common n-grams, word similarity can be also measured using edit dis- tance. The edit distance between two strings is the minimum number of …

Content truncated to 500 words