Predicting Home User Activities in a Secured Smart home Using Internet of Things and Deep Learning

Abstract: The innovation Smart home is a home where appliances can be remotely controlled and monitored with the help of a network connection. Despite the numerous advantages of smart home, the security of smart homes has been a concern to home users also, smart home can improve in predicting user activities in the home. In this research work, we were able to predict home user activities in a secured smart home. A software interface was provided for One Time Password (OTP) authentication and for controlling home appliances. The use of deep learning model enables us to predict accurately, when appliances homes should be ON or OFF. The prediction was made from previous appliances historical dataset. The home security was enhanced with one time password and decision tree to detect dictionary attacks. The result gotten from the prediction system showed series of predictions made by the system at different time interval. The security result shows how effective the smart home security system is in mitigating attacks in the home.

Keywords: Smart home, Security, One Time Password, Internet of Things, Machine Learning.