MRGIR: Open geographical information retrieval using MapReduce
Zhiang Wu, Bo Mao, Jie Cao · 2011
City objects recommendation based on characteristics of users, location, time and weather is a challenging issue in geographical information retrieval (GIR). In the meanwhile, city objects recommendation is a computation-intensive and data-intensive application. Cloud computing has gained significant attention in recent years to process the large volume of data. MapReduce framework is currently a most dominant technology in cloud computing. Augmented User-based Collaborative Filtering (AUCF) algorithm which can effective deal with hybrid variable types is proposed firstly. Then, MapReduce for GIR (MRGIR) is presented and AUCF is implemented within MRGIR as an example. The MRGIR is implemented in Hadoop which is an open source framework for MapReduce. Experimental results shows that with moderate number of map tasks, the execution time of GIR algorithms (i.e., AUCF) can be reduced remarkably.