The Study On Preference Sensitive Decision Tree Algorithm And Its Application In Household Environment Monitoring And Control Question

With the continuous improvement of living standards, people's requirements for the home environment are increasing, which allow smart home concept to come into being. On the basis of the traditional residential function, smart home could also finish the interconnection, self-management and self-control of household equipments through the Internet of things technology, finally, it provides a kind of safe, comfortable and efficient residential environment for inhabitants. As an important issue in the field of smart home study, household environment monitoring could complete the corresponding control of household environment by real-time detecting indoor environmental conditions. Therefore, the article attempts to do some relative research, and the study work as follows:(1) this paper puts forward a kind of preference sensitive decision tree algorithm (PSDTA), (2) this paper developes a set of household sleep environment monitoring and forecasting system.The behavior prediction problem with preference in home environment monitoring are being studied in this paper, and it is abstracted into decision problems. By collecting and analyzing the data of a series of behavioral movements, time and sound in the process that people enter the room, preference sensitive decision tree algorithm is proposed to predict the behavior of the occupants. The algorithm reconstructed the attribute selection factors by introducing the concept of "class-preference degree", "EP increment" and so on. Before the start of the study, this paper uses a kind of feature pre-selection methods to eliminate the redundancy of attribute, and solves the problem of without considering the correlation between the non class attribute in the construction process of traditional decision tree. The standard UCI data set and the real data of sleep environment are selected to carry out the experimental test, the experimental results show that this algorithm can reduce the size of the decision tree effectively. Further more, the algorithm can not only achieve the high precision prediction of preference class, but also ensure the decision tree's overall accuracy. Finally, a better predict result of residents sleep tendency in the experiment was obtained.According to PSDTA algorithm idea, this paper designs and develops a set of household sleep environment monitoring and forecasting system. The system integrates various techniques, like embedded technology, ZigBee technology, wireless location technology, WI-FI technology and infrared technology. In the developing environment of the IAR Embedded Workbench IDE and IAR EWARM, respectively C language is used to development of ZigBee node and intelligent gateway, otherwise, in the Eclipse environment, Java language is used in the development of the Android client monitoring software, then above all-mentioned can be run in no less than Windows 7.0 operating system. The main functions of the system include system automatically triggered function, environmental monitoring function and sleep preference prediction function. The system automatically triggering function mainly determines the real-time distance between habitant and household environment, and then determines whether automatically trigger system. Environmental monitoring function is responsible for monitoring all kinds of environmental data and automatic control related equipment. Preference prediction function of sleep could be input all kinds of environment date, movements date into PSDTA model and forecast the residents tend to sleep, then the system controls those room equipment, and adjusts the environment to the preference state according to the prediction results. Test results show that the design of hardware and software have better compatibility and various function effect in line with expectations, besides, the system has good stability and feasibility.From what I have mentioned above, the issue of household behavior preference in environmental monitoring is studied by this paper, which puts forward a kind of solution named PSDTA algorithm to solve the problem, and develops a set of sleeping environment monitoring and forecasting system. This paper provides new ideas and new methods to solve the problems of insufficient human-facing experience, ungenerous intelligence and high cost in current smart home.