ClovaCall

ClovaCall是一个大规模的韩语目标导向对话语音语料库，专门用于接触中心的自动语音识别服务。该数据集包含约60,746对简短句子和相应的口语表达，主要集中在餐厅预订领域。数据集的创建过程涉及通过众包平台Crowdworks收集和筛选句子，并通过电话记录口语表达。ClovaCall数据集适用于多种基于接触中心的预订服务，因为其内容涵盖了预订服务中常用的词汇和表达。该数据集的应用旨在提高目标导向对话场景中自动语音识别的准确性。

ClovaCall is a large-scale Korean goal-oriented conversational speech corpus specifically designed for automatic speech recognition services in contact centers. This dataset contains approximately 60,746 pairs of short sentences and corresponding spoken utterances, mainly focusing on the restaurant reservation domain. The creation of this dataset involved collecting and filtering sentences via the crowdsourcing platform Crowdworks, as well as recording spoken utterances through telephone calls. The ClovaCall dataset is applicable to a variety of contact center-based reservation services, as its content covers the commonly used vocabulary and expressions in reservation scenarios. The application of this dataset aims to improve the accuracy of automatic speech recognition in goal-oriented conversational contexts.