A Computational Learning Theoretic Approach to Birdsong Syntax Analysis

我们提出了一个有效的自动机为基础的方法来提取动物行为的连续序列数据的行为单元和规则。引入原有的扩展，我们集成了两个基本的方法-Ngram模型和Angluin的机器学习算法到行为学数据挖掘框架。它使我们能够获得最简单的有限自动机表示的行为规则，接受（或产生）的最小的一组可能的行为模式从动物行为的序列数据。用这种方法，我们演示了如何使用真实的鸟鸣数据的行为学数据挖掘工程，并使用人工鸟鸣数据由计算机程序生成的这种方法进行实验评估。这些结果表明，我们的行为学数据挖掘有效地工作，即使是嘈杂的行为学数据，通过适当地设置参数。此外，我们展示了一个案例研究，使用孟加拉雀的歌曲，表明我们的方法成功地抓住了核心结构的歌唱行为，如循环和分支。

We propose an efficient automata-based approach to extract behavioral units and rules from continuous sequential data of animal behavior. Introducing original extensions, we integrate two elemental methods-the Ngram model and the Angluin's machine learning algorithm into an ethological data mining framework. It allows us to obtain the simplest finite automaton representation of behavioral rule that accepts (or generates) the smallest set of possible behavioral patterns from sequential data of animal behavior. With this method, we demonstrate how the ethological data mining works using real birdsong data and performs experimental evaluations of this method using artificial birdsong data generated by a computer program. These results suggest that our ethological data mining effectively works even for noisy ethological data by appropriately setting the parameters. In addition, we demonstrate a case study using the Bengalese finch song, showing that our method successfully grasps the core structure of the singing behavior such as loops and branchings.