The prediction of building energy consumption is an important task in building design, retrofit and operation. A well-designed building and its energy system may lead to energy conservation and CO2 reduction. This book deals with up-to-date artificial intelligence models and optimization techniques related to this application.
In this book, we start with a review of recently developed models on predicting building energy consumption, including elaborate and simple engineering methods, statistical methods and artificial intelligence methods, especially ANNs and SVMs. This previous work includes solving all levels of energy analysis with appropriate models, optimizing model parameters, treating inputs for better performance, simplifying the problems and comparing different models. We then summarize the advantages and disadvantages of each model. Among existing methods, artificial intelligence models are attracting more and more attention in the research community.
Then, this book attempts to apply SVMs in predicting building energy consumption. We present the SVM principles in depth, as well as some important extensions, including SVC, SVR, one-class SVM, multiclass SVM and transductive SVM. These models have demonstrated superiority in many sorts of applications due to ideas of maximum margin, regularization and the kernel method.
Before using SVMs in our application, ...