Sprinklr/Intent_Prediction

意图预测数据集是一个针对商业环境中的对话进行意图识别模型测试的数据集。该数据集是使用Gemini-2.0-flash合成生成的，包含模拟的客服中心对话。数据集包括三个不同质量级别的议题分类（taxonomy），可用于对接触中心对话中的分类发现过程进行基准测试。数据集以英文为语言，遵循Apache-2.0许可。它包含三个分类，分别具有不同数量的意图和层级，并提供了相应的数据文件格式。数据集由Spinklr AI策划。

The Intent Prediction dataset is designed for testing intent identification models on business context conversations. It is synthetically generated using Gemini-2.0-flash and includes simulated contact center conversations. The dataset comprises three taxonomies of varying quality and levels for benchmarking the taxonomy discovery process in contact center conversations. The dataset is in English and licensed under Apache-2.0. It includes three taxonomies with different numbers of intents and levels, and details on the data file formats. The dataset is curated by Spinklr AI.