Rough set method to identify key factors affecting precipitation in Lhasa

Hydrological information is a digital description for hydrological processes, which reflects the interaction among different elements for the surface system on earth. Conventional hydrological information processing methods to deal with such kind of uncertainty with mathematical assumptions does not tally with objective facts. Rough set theory is an effective mathematical tool treating imprecise, uncertain and incomplete data. It is based on the classification mechanism and assumes that classification is an equivalence relation in specific space while no external parameters are needed. By analyzing 13 factors that may affect precipitation at Lhasa station from 1972 to 1995, eight key factors affecting precipitation are identified and 58 decision rules are generated in this study. Data from 1996 to 2001 are used to evaluate performance of rough set model and the result shows that 83% of the test sets can be classified correctly and thus the rough set model can effectively address uncertainty of various factors affecting precipitation.