sjjsy / sor-nilm

A repository to house an NFHMM algorithm implementation and a report and presentation related to a NILM related project on the SOR course.
4 stars 1 forks source link

AttributeError: 'NFHMMDisaggregator' object has no attribute '_pre_disaggregation_checks' #2

Open makkarss929 opened 4 years ago

makkarss929 commented 4 years ago

Can't run code due to AttributeError: 'NFHMMDisaggregator' object has no attribute '_pre_disaggregation_checks',Please refer to image for more details Screenshot (681)

makkarss929 commented 4 years ago

Screenshot (681)

makkarss929 commented 4 years ago

` == Running NFHMM with HP=10, SI=30, SP=10...

== nfhmm.train(dataset.buildings[1].elec)

== nfhmm.disaggregate(dataset.buildings[1].mains(), output)


AttributeError Traceback (most recent call last)

in 29 output = HDFDataStore(str(nfhmm_outfile), 'w') 30 print('\n== nfhmm.disaggregate(dataset.buildings[%d].mains(), output)' % (disag_building)) ---> 31 nfhmm.disaggregate(dataset.buildings[disag_building].elec.mains(), output, max_appliances=3, sample_period=sp) 32 output.close() 33 ~\Anaconda3\envs\keras-gpu\lib\site-packages\nilmtk_2\nilmtk\disaggregate\nfhmm_wrapper.py in disaggregate(self, mains, output_datastore, max_appliances, **load_kwargs) 67 """ 68 # Run any required pre disaggregation checks ---> 69 load_kwargs = self._pre_disaggregation_checks(load_kwargs) 70 71 # Set defaults AttributeError: 'NFHMMDisaggregator' object has no attribute '_pre_disaggregation_checks' `
sjjsy commented 4 years ago

Hi @makkarss929, pleased to hear of your interest toward NILM and our work. Unfortunately we haven't spent much time on ensuring others can easily understand the details, test and build upon this work. Also we had no intention to maintain it as it is a project done for educational and proof of concept type exploratory use. Since it's three years since we last worked on this it would take me some time to be able to solve your problem.

May I ask what is your goal? The NILMTK includes some algorithms that can be more easily understood and applied and there are some that perform better than this unsupervised NFHMM algorithm provided you have good training data. Or are you specifically interested in this because it is unsupervised? Where/how are you planning to apply it?

makkarss929 commented 4 years ago

Dear Sir, I hope you’re having a great week!

My name is Tarun Makkar,B.Tech 3rd year ,and I work as ML intern at The Knight Labs,Noida,UP,India working for linc.world,Madrid,Spain.I recently worked on Energy disaggregation using submetered data. Currently, I am working on Unsupervised version and I became familiar with your work when I saw your piece about Energy Disaggregation i.e sor-nilm.

If you’re open to it, I’d love to connect on LinkedIn to learn more about your experience.

Really looking forward to keeping in touch.

Thanking You Yours Truly Tarun Makkar 7056311813

On Wed, Feb 19, 2020 at 1:24 PM sjjsy notifications@github.com wrote:

Hi @makkarss929 https://github.com/makkarss929, pleased to hear of your interest toward NILM and our work. Unfortunately we haven't spent much time on ensuring others can easily understand the details, test and build upon this work. Also we had no intention to maintain it as it is a project done for educational and proof of concept type exploratory use. Since it's three years since we last worked on this it would take me some time to be able to solve your problem.

May I ask what is your goal? The NILMTK includes some algorithms that can be more easily understood and applied and there are some that perform better than this unsupervised NFHMM algorithm provided you have good training data. Or are you specifically interested in this because it is unsupervised? Where/how are you planning to apply it?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sjjsy/sor-nilm/issues/2?email_source=notifications&email_token=AJ5PO76E2TSBZGUJOBOIAATRDTQSXA5CNFSM4KWWBRWKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEMGW7BQ#issuecomment-588083078, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJ5PO75Z5VB6UEW2NXQI36LRDTQSXANCNFSM4KWWBRWA .

sjjsy commented 4 years ago

Hey Tarun, nice to learn of your interest! Hope our repository inspires you to continue deeper into your study of energy disaggregation and unsupervised learning! I do recommend getting familiar with the newest version of NILMTK and the algorithms there and perhaps then trying to get the NFHMM algorithm we developed in R working within your setup and with your data. Perhaps then contributing to NILMTK by adding support for unsupervised algorithms. Perhaps you can fit the work to be a part of a university course project. Good luck!