Analyzing Co-occurrence Data
Analyzing Co-occurrence Data
复制标题
分析共现数据
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Philip Durrant
中科院分区：
文献类型：
--
作者：
S. Gries;Philip Durrant
在本章中，我们概述了共发生数据的定量方法。我们首先简要概述了不同类型的共发生的术语，这些概述在语料库研究中是突出的，然后讨论了一些用于量化共发生的关联措施的计算。我们提出了两个代表性的案例研究，其中一项探索词汇搭配和学习者的能力，这是动词具有参数结构结构的另一个创造性用途。此外，我们强调了大多数广泛使用的措施实际上是如何从将语料库的关联视为回归建模的一个实例中脱颖而出的，并讨论了较新的发展以及关联量度研究的潜在改进，例如使用关联的方向测量，而不是不批评的频率，以及关联度量，类型频率和熵中的关联 - 强度信息。
In this chapter, we provide an overview of quantitative approaches to co-occurrence data. We begin with a brief terminological overview of different types of co-occurrence that are prominent in corpus-linguistic studies and then discuss the computation of some widely-used measures of association used to quantify co-occurrence. We present two representative case studies, one exploring lexical collocation and learner proficiency, the other creative uses of verbs with argument structure constructions. In addition, we highlight how most widely-used measures actually all fall out from viewing corpus-linguistic association as an instance of regression modeling and discuss newer developments and potential improvements of association measure research such as utilizing directional measures of association, not uncritically conflating frequency and association-strength information in association measures, type frequencies, and entropies.