Hi,Doctor Xiao!
I have read your paper, and I try your code. But I can't under stand your timeseries format. To be specific, Your demo produce real batch data (256,133,2) and (256,) the number 2 I think is the timestramp and event label. the number 133 means there are 133 events happen randomly. However I can't understand why the timeseries data format is (256,133,7,4). Why is it 133? I think timeseries value is sampled in a fixed frequency, and its length is irrelevant with the numbef of the event occuring.
你好,
我是来自交大生医工学院的研究生,想使用您的研究帮助我做有关情绪检测方面的工作。
我读过您的文章,在代码运行过程中,不是很理解为什么时间序列的数据格式是这样的。打个比方,按照代码生成的数据 (256,133,2),事件序列的长度统一为133.数字2分别指发生的事件类型和时间点(我是这么理解的),事件是随机发生的。然而时间序列是按照固定频率采样得到的,我不是很理解为什么它的长度也被设置成了133。想请问应该如何理解您所设置的时间序列的数据格式(256,133,7,4)
谢谢!