mdymczyk / iot-pipeline Goto Github PK

Hi there,

You apply the sliding window to construct the training data for the autoencoder model which is converted to dim(rows-window+1, columns*window). My understanding is that the last (window-1) observations are lost, the number of observations (data points) is reduced from rows to (rows-window+1).

After competing the anomaly error, how can we tell if the anomaly happens within the last (window-1) observations?

File "", line 28
elif:
^
SyntaxError: invalid syntax

Hi there,

Firstly thanks for sharing such a great example.

However, I don't think the following codes will be run in your python script. They are just directly copied from the R script.
newRow = c(inp[idx,], t(inp[(idx+1):(idx+window-1),]))
if idx %% 10000 == 0: print(idx)