Sanders Nathan C, Chin Steven B
Department of Linguistics, Indiana University.
J Quant Linguist. 2009 Feb 1;16(1):96-114. doi: 10.1080/09296170802514138.
Phonological distance can be measured computationally using formally specified algorithms. This work investigates two such measures, one developed by Nerbonne and Heeringa (1997) based on Levenshtein distance (Levenshtein, 1965) and the other an adaptation of Dunning's (1994) language classifier that uses maximum likelihood distance. These two measures are compared against naïve transcriptions of the speech of pediatric cochlear implant users. The new measure, maximum likelihood distance, correlates highly with Levenshtein distance and naïve transcriptions; results from this corpus are easier to obtain since cochlear implant speech has a lower intelligibility than the usually high intelligibility of the speech of a different dialect.
语音距离可以使用形式化指定的算法进行计算测量。这项工作研究了两种这样的测量方法，一种是由内尔博内和赫林加（1997年）基于莱文斯坦距离（莱文斯坦，1965年）开发的，另一种是邓宁（1994年）语言分类器的改编版，它使用最大似然距离。将这两种测量方法与小儿人工耳蜗使用者语音的原始转录进行比较。新的测量方法，即最大似然距离，与莱文斯坦距离和原始转录高度相关；由于人工耳蜗语音的可懂度低于不同方言语音通常较高的可懂度，因此从该语料库获得结果更容易。
