Data Prediction Service Model Based on Sequential Pattern Mining in Sensor-Cloud

With the development of cloud computing and Internet of Things, they gradually merged together to form a new system called sensor-cloud. Aiming at the characteristics of Sensing-as-a-Service of sensor-cloud, this paper studies the features of sensing data, a data mining schedule is introduced into data processing and transmission during the service process in sensor-cloud. A sequential pattern mining algorithm is used to predict data sequences that appear more frequently during the service period. By providing prediction data to system user, energy consumption would be saved, and service time would be reduced. We propose a patterns mining and matching scheme that is suitable for sensor-cloud system. By generating predicted data in different tasks to match the existing sensed data with sequential patterns, the prediction model provides a scheme for task migration in sensor-cloud. Simulate result shows that our model is quiet effective in sensor-cloud. It can reduce energy consumption and improve response time under a high accuracy.