IFRAT: An IoT Field Recognition Algorithm Based on Time-Series Data
Shuai Guo, Zhongwen Guo, Zhijin Qiu, Yingjian Liu, Yu Wang · 2017
The increasing interest in integration of Internet of Things (IoT) heterogeneous data has resulted in the introduction of a variety of systems designs and schema matching algorithms. However, numerous algorithms for schema matching fail to process automatically and efficiently because of the unknown of data source's schema. In this paper, we attempt to solve this problem by introducing a new algorithm that could recognize the field of data source through its large collections of time-series data. By knowing the field forward, we could get the basic schema, which makes great contribution to schema matching afterwards. Our algorithm has a good advantage at extracting characteristics of time series data and cluster them by using the self-organizing map (SOM). Then we apply clustering results to recognize IoT fields and devices when a new unknown dataset is coming. We demonstrate the utility and efficiency of our algorithm with a set of comprehensive experiments on real datasets from several fields. The results show that our algorithm has good performance and efficiency.