Data For Gaussian-Process-Based Emulators For Building Performance Simulation

Rastogi, Parag;Khan, Mohammad Emtiyaz;Andersen, Marilyne

Description

The ZIP folder contains the MAT files you need to rerun the experiment described inRastogi, Parag, Mohammad Emtiyaz Khan, and Marilyne Andersen. 2017. “Gaussian-Process-Based Emulators for Building Performance Simulation.” In Proceedings of BS 2017. San Francisco, CA, USA: IBPSA.-------------------------------------------------------------The two m-scripts (MATLAB) help you to load the results reported in the paper. Make sure to CHECK the file paths inside the scripts, especially to the MAT files. Usually, the paths should be fine if you update the variable pathMATfolder inside the script RunThis.m .There are two types of MAT files inside the folder called "Data" :1. Original data (building simulations) --> gpdata_BaseSimulation.mat2. Errors and predictions - errs_BaseSimulation_N_M.mat and ystore_BaseSimulation_N_M.mat --> The first contains all the error quantities and the second the 'y' predictions. The number N represents the run number (subset of master training data set sampled for the given run). The number M can take only two values - 1 or 2. The models for heating load are represented by 1 and for cooling by 2.3. Metadata - trainN* --> These files contain metadata for setting up the plots.See github repository https://github.com/paragrastogi/GPregressionInBS.git for more scripts. See www.paragrastogi.com or www.ibpsa.org for the conference paper.

Citations (0)

Mentions (0)

Metrics

Dataset Index

0.5

FAIR Score

79%

Citations

0

Mentions

0

Metrics Over Time

Publication Details

DOI

Publisher

Zenodo

License

Creative Commons Attribution 4.0

Open Access

Assigned Domain

Subfield

Management Science and Operations Research

Field

Decision Sciences

Domain

Social Sciences

Confidence Score

97%

Source

Open Alex

Keywords

building simulationgaussian process regressionbuilding energy efficiency

Normalization Factors

FT

57.69

CTw

1.00

MTw

1.00