Two component data representation using piecewise approximation and specific points for IoT

In these days, concept of IoT is widely spread from individual devices to large industrial systems. However, data handling from IoT devices is on developing stage. Delivering large amount of data to server which mainly operates data analysis is required and it becomes a burden to network. Our work has focused on reducing volume of data to transmit with data approximation in the domain of time series data. We suggest a data representation method which separates data into specific data point and piecewise approximation values. In addition, we show the difference between raw data and reduced data in distance and in time series classification accuracy.