Application Reasearch Of Smart Home System Based On Data Mining

In recent decades,with the improvement of the economic level of people’s material life requirements are also gradually raised,smart home has emerged accordingly,but the original smart home is not really smart,with the progress of people’s living standards also more and more unable to meet people’s requirements.With the development of Internet technology and artificial intelligence technology matures,the development of smart home as if encountered a new dawn,how can a smart home control from the hands of people actually handed over to smart home is a big difficult problem which needs to be settled,although the road will be hard and long,step by step,but we can make some small change.This paper proposes a new solution based on the phenomenon that the smart home is not smart enough to help the smart home to be able to adapt to the changes in user behavior.The smart home in this paper is grouped in the way of scene mode,and the state of smart home devices in scene mode is correspondingly changed through the change of user behavior.This paper proposes an adaptive model of smart home scene mode,which can quickly and accurately adjust the state and list of devices in the scene mode,aiming at liberating users’ hands and improving user experience at the same time.In order to build the scene mode adaptive model,first of all,based on the collected user data preprocessing,and then equipment for user data correlation analysis and correlation analysis of user behavior characteristics,to design appropriate scene mode for the user,and according to the correlation between user behavior patterns in the situation of faster adaptive adjustment.Secondly,if there are obvious and traceless changes in the user behavior,the sliding window method is used to determine the effectiveness of the event according to the probability of the event,so as to determine whether to update the list of smart home devices in situational mode.Then,the operation state of the equipment needs to be predicted.In this paper,two most commonly used smart home prediction models are selected,which are based on BP neural network and support vector machine.The weight-value and threshold-value of BP neural network prediction model is selected quickly by genetic algorithm.Through many experiments and comparisons,the results show that the prediction model based on BP neural network is more accurate than that based on support vector machine.Introducing,testing of the correlation of association rule mining the association rules between the equipment into prediction model based on BP neural network,and see if there are any more to improve forecast accuracy,it can be seen that the introduction of the association rules did not bring the accuracy in forecasting model qualitative leap,but based on BP neural network prediction model does have the accuracy improvement.Based on the above-mentioned scene mode prediction model combining Apriori association rules and BP neural network,the prediction model prediction accuracy is as high as 92.83%.Finally,a user simulation experiment is conducted on the situational adaptive model in this paper,and the experimental results show that the accuracy of the situational adaptive model proposed in this paper has been greatly improved,especially that the adaptive model can make corresponding adjustments faster when some relevant changes of user behaviors occur.The prediction accuracy of the model reached 93%.