TransBEAM contains a set of classes and methods to analyze and compare different state-of-the-art data-driven building energy modeling techiques. The module also implements two ways to implement transfer learning: data transfer and model transfer, to make use of both the simulation data and the field data. The current version is still in development stage with two models being implemented to predict building energy consumption: Random Forest and Feed-Forward Network. In the current version, the library can only handle timeseries data in tabular format. In the future versions, we plan to support much more advanced techniques, including other data formats.
- Setup the conda environment in your system:
- Once conda is installed, open Miniconda/Anaconda shell and go to the installation directory.
- Git clone the repository:
git clone https://stash.pnnl.gov/scm/~jain432/transBEAM.git - Go to the installed module folder:
cd <installation_folder>\transBEAM - Create conda environment using
environment.yml:conda env create -f environment.yml - Activate the conda environment:
conda activate bem
- Go to the examples folder:
cd <installation_folder>\transBEAM\examples - Run the jupyter notebook:
jupyter notebook
Milan Jain, Khushboo Gupta, Arun Sathanur, Vikas Chandan, and Mahantesh M Halappanavar. "Transfer-Learnt Models for Predicting Electricity Consumption in Buildings with Limited and Sparse Field Data." American Control Conference (ACC) 2021.
Released under the 3-Clause BSD license (see License.rst)