基于时序分析的人员在室预测模型及其参数优化*
Prediction model and parameter optimization of room occupancy based on time sequence analysis
摘要：
分析了人员在室的时序规律，提出了基于时序分析的人员在室预测模型及其参数优化方法。将体现人员在室规律的时刻信息与历史在室概率作为输入训练模型，实现了人员在室预测。以混淆矩阵、正确率与均方根误差（E）作为指标，提出了人员在室预测模型的检验方法。选取珠海市某办公楼内1间办公室累计1年的人员在室数据作为实测案例，验证了人员在室预测模型，结果表明，人员在室预测的综合正确率为89.4%，E为0.325 9。
Abstract：
Analyses the time sequence of the room occupancy, and puts forward a prediction model and a parameter optimization method based on the time sequence analysis. Taking the time information and the historical probability of room occupancy as the input training model, realizes the room occupancy prediction. Taking the confusion matrix, accuracy and root mean square error (E) as the indicators, proposes a test method of room occupancy prediction model. Selects a one-year data of an office in an office building in Zhuhai as an actual case to verify the prediction model, and the results show that the comprehensive accuracy rate of the prediction is 89.4%, and E is 0.325 9.
Keywords:timesequenceanalysis,occupancybehavior,roomoccupancy,predictionmodel,accuracy,rootmeansquareerror