Time Series - Cleansing a Dataset
Overview
Cleansing a time series dataset is not much different
cleansing
any dataset, assuming that the dataset is
stationary.
When the dataset is not stationary, additionaly complications can arise. For example, if the dataset represents asset prices, but one value is missing,
removing the missing data record and then calculating returns (or some form of
differencing)
will likely produce a return value that is larger and more volatile on average with the other returns. This is because, the return computed between the two
values on either side of the missing value will represent a return over 2 periods, instead of one.
Topics
