Let's say that I have a multivariate time series dataframe of sensors. Each sensor has its serial number, group, several statistics and column failure. Data about sensor's statistics are collected daily. When sensor is broken, failure cell of last good record is set to 1 (otherwise it is 0).

Failure records are very rare, less than 1%.

How to oversample minority class (sensors that have failed) in this case?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.