Predicting the Number of People In a Room
Keywords: indoor people number, SVM, deep learning, machine learning
TL;DR: Beginner，Preliminary work
Abstract: At present, the energy consumption of the whole building process and the energy consumption of building operation in China accounts for a high proportion of the country's total energy consumption, and reducing the energy consumption of building operation is an important part of saving energy, reducing greenhouse gas emissions, mitigating climate change and achieving carbon neutrality. Building energy consumption is mainly composed of two parts: cooling system energy consumption and heating system energy consumption, both of which are related to indoor cooling and heating loads. A large part of the indoor load is generated by indoor personnel, and the load has a lagging effect. Therefore, if the number of indoor personnel can be predicted in advance, the start/stop time and operating power of the cooling and heating systems can be determined in advance, so as to avoid wastage of cold and heat, and thus to achieve the purpose of energy saving and emission reduction. In the past, there are fewer studies on the time series prediction of indoor occupancy, so this paper will take a specific room as the object of study, based on its occupancy monitoring data, use different methods such as Support Vector Machine (SVM), Auto Regressive Integrated Moving Average (ARIMA), Random Forest, Deep Neural Networks, etc. to predict the number of indoor occupants, and evaluate the advantages and disadvantages of the effectiveness of various prediction methods.
Submission Number: 55
Loading