[英]Can I perform multivariate time series forecasting on yearly data
I wanted to build an LSTM model that can work on a yearly dataset.我想构建一个可以处理年度数据集的 LSTM model。 I found that most of the articles or resources about LSTM use DateTime instead of the year, but my dataset only has year and cannot drill down to month-day...
我发现大多数关于 LSTM 的文章或资源都使用 DateTime 而不是年份,但我的数据集只有年份,不能深入到月日......
Is LSTM able to work on a small yearly dataset? LSTM 是否能够处理小型的年度数据集? Or there is any model more suitable for the task?
或者有没有更适合这个任务的model?
The Heisenberg uncertainty principle states that:海森堡测不准原理指出:
a given function cannot be arbitrarily compact both in time and frequency, defining an “uncertainty” lower bound.给定的 function 在时间和频率上都不能任意紧凑,从而定义了“不确定性”下限。
Or with other words, if you want to forecast patterns on a daily basis, you need at least data on a daily basis, if you want data on monthly basis you need data on monthly basis if you indeed want to predict a yearly forecast, then you can use your yearly data but you will not be able to drill further down.或者换句话说,如果你想每天预测模式,你至少需要每天的数据,如果你想要每月的数据,你需要每月的数据,如果你确实想预测年度预测,那么您可以使用您的年度数据,但您将无法进一步向下钻取。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.