CondTSC

CondTSC是由上海交通大学开发的一个针对时间序列分类的数据集浓缩框架。该数据集包含60条数据，旨在通过双域匹配技术生成一个较小的合成数据集，以保持与完整真实数据集相当的性能。CondTSC通过多视图数据增强、双域训练和双代理目标匹配来增强时间域和频率域的数据浓缩过程。该数据集主要应用于时间序列分析，旨在解决深度学习任务中时间序列数据管理的问题，特别是在训练深度神经网络时。

CondTSC is a dataset condensation framework for time series classification developed by Shanghai Jiao Tong University. It generates a small-scale synthetic dataset with 60 data samples via dual-domain matching technology, which is designed to retain performance comparable to that of the full real-world dataset. CondTSC enhances the data condensation process in both time and frequency domains through multi-view data augmentation, dual-domain training, and dual-proxy objective matching. Primarily applied to time series analysis, this framework aims to address the issues of time series data management in deep learning tasks, particularly during the training of deep neural networks.