search

OdysseasKr / neural-disaggregator

Code for NILM experiments using Neural Networks. Uses Keras/Tensorflow and the NILMTK.
MIT License
138 stars 57 forks source link

Calling diaggregate with a single monthly aggregate reading. #17

Closed cathreya closed 4 years ago

cathreya commented 4 years ago

This is probably more of an NILMTK question, but the disaggregate function in the shortseq2pointdisaggregator.py takes an Elecmeter object as a parameter. However, I am interested in getting the disaggregation of a single monthly aggregate. I tried to circumvent this by using something similar to the dataset converter code, by writing and reading from a hdf5 store, but I didn't manage to get anywhere with it. Is there any other way to achieve this?

OdysseasKr commented 4 years ago

Hi @cathreya! What do you mean by single monthly aggregate? Does this mean one value per month? Maybe you could use the sample_period attribute from NILMTK?

cathreya commented 4 years ago

Thanks for your reply! Instead of using values from the dataset, we want to to take total mains electricity usage per month from a user, and convert it to an Elecmeter input to the model.

OdysseasKr commented 4 years ago

In this case I am afraid that you have to write your own custom disaggregator. All models in this repo are designed to work with fast sampling rates (usually with a sampling period of 6 seconds).

cathreya commented 4 years ago

Suppose we had 6 seconds aggregates from the user, would we be able to create Elecmeter objects for those?

OdysseasKr commented 4 years ago

Aah ok I see. So you want to use a feed of user data but not a dataset? NILMTK is a toolkit for working with datasets so, since you couldn't figure out the converter, you can not use neural-disaggregator.

cathreya commented 4 years ago

That is a shame. Thank you for your help and your time.