You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello,
I have seen that OmniXAI should work for time series forecasting. However, I'm a bit confused on what the output of the predict_function should be and I have not found any example in the documentation for forecasting. In the documentation I read this: predict_functionshould be an Timeseries instance. The outputs of predict_function are anomaly scores (higher scores imply more anomalous) for anomaly detection or predicted values for forecasting.
( I was trying to adapt this example to do forecasting.)
When I create a ShapTimeseries with mode='forecasting' and call explain(test_x) on it, the predict_function receives as input a Timeseries instance that has 100 samples from the time series data, that each have the length of my test set. So is the expected output one prediction for each of the time series? And should that be the next time step after the time series? Or should I make the length of my test set equal to 1 and create a prediction for this time stamp? Or ignore the true values that are passed and create predictions for each of the time stamps in these time series?
Thank you in advance for your answer!
The text was updated successfully, but these errors were encountered:
Hello,
I have seen that OmniXAI should work for time series forecasting. However, I'm a bit confused on what the output of the predict_function should be and I have not found any example in the documentation for forecasting. In the documentation I read this:
predict_function
should be an Timeseries instance. The outputs of predict_function are anomaly scores (higher scores imply more anomalous) for anomaly detection or predicted values for forecasting.( I was trying to adapt this example to do forecasting.)
When I create a ShapTimeseries with mode='forecasting' and call
explain(test_x)
on it, the predict_function receives as input a Timeseries instance that has 100 samples from the time series data, that each have the length of my test set. So is the expected output one prediction for each of the time series? And should that be the next time step after the time series? Or should I make the length of my test set equal to 1 and create a prediction for this time stamp? Or ignore the true values that are passed and create predictions for each of the time stamps in these time series?Thank you in advance for your answer!
The text was updated successfully, but these errors were encountered: