A Distributed and Standard based Data Warehouse and Mining System for Large-Scale Sensor Data

Intelligent Automation, Inc. (IAI), along with our collaborators propose to develop a Distributed and Standard based Data Warehouse and Mining System for Large-Scale Sensor Data (DSDWM). DSDWM is a synergy of SensorML based database system, streamed and high-dimensional data cubing approaches. Our stream cubing architecture is designed to handle streamed sensor data with quick response and partial computation of data cubes by popular-path cubing. The DSDWM innovations are: (1) SensorML schema database to develop a solid and extendable data storage system for large amount of complex sensor data (2) unique innovative high-dimensional data cubing technology, (3) internal data mining tool that adopts a generic framework of data mining, to provide advanced data exploitation and analysis capability beyond cubing technology, and (4) AJAX on mobile devices to enable asynchronous communication so that the bandwidth is saved. DSDWM offers the key advantage that multi-source sensor input data can be transferred to minimal cubing thus saving computation and storage expense significantly for high-dimensional cubes and prediction cubes provide an automatic and intelligent way to build cubes with optimized accuracy for a large range of prediction models.