Energy disaggregation/non-intrusive load monitoring
Budget: ₹37,500 – ₹75,000 INR
We are looking for someone who can modify the NILMTK toolkit to allow the following functionality
1. Write a data transformer for the input provided in a custom format to a format supported by NILMTK or write a dataset_converter class for our custom format. An initial version of a dataset_converter will be provided. (5% effort)
2. Use algorithms available NILMTK-contrib to train and predict on given data. (5% effort)
3. Build pipeline for comparing and saving best performing models for deployment use (given data for a 1 hour interval, return the disaggregated data and not just the metrics). (90% effort)
Useful links
* NILMTK repository: https://github.com/nilmtk/nilmtk
* NILMTK-contrib repository: https://github.com/nilmtk/nilmtk-contrib/
* Paper for reference about NILM: https://arxiv.org/pdf/2102.12578
1. Write a data transformer for the input provided in a custom format to a format supported by NILMTK or write a dataset_converter class for our custom format. An initial version of a dataset_converter will be provided. (5% effort)
2. Use algorithms available NILMTK-contrib to train and predict on given data. (5% effort)
3. Build pipeline for comparing and saving best performing models for deployment use (given data for a 1 hour interval, return the disaggregated data and not just the metrics). (90% effort)
Useful links
* NILMTK repository: https://github.com/nilmtk/nilmtk
* NILMTK-contrib repository: https://github.com/nilmtk/nilmtk-contrib/
* Paper for reference about NILM: https://arxiv.org/pdf/2102.12578