Application of Rough Sets Conditional Information Entropy in Climatic Evaluation of Flue-Cured Tobacco Growing Areas

In order to analyze the application effect of rough sets conditional information entropy in climatic evaluation of flue-cured tobacco growing areas, climatic data of 18 typical tobacco growing areas in Yunnan were collected, evaluated by the analytic hierarchy process (AHP), standard deviation (SD), comprehensive weighting method based on game theory and rough sets conditional information entropy, and compared the results of weighting and evaluation. The results showed that: (1) The descriptive statistics of climate indices from different flue-cured tobacco growing areas showed that the mean temperature from root elongation period (20.17℃) to maturation period (19.51℃) was generally in a appropriate level. And the mean precipitation increased gradually from root elongation period to maturation period which was at a high level. And the sunshine hours in tobacco field were sufficient; (2) Among the AHP, SD method and comprehensive weighting method based on game theory, the index with highest weight coefficients were the precipitation at fast-growing period index (0.173), the mean temperature at fast-growing period index (0.143) and the mean temperature at fast-growing period index (0.152). The discrimination (η) was 0.079, 0.084 and 0.083, respectively; (3) The index with highest weight coefficient of rough sets conditional information entropy was the precipitation at fast-growing period index (0.239) and the η was 0.090; (4) The results of rank correlation coefficient of different evaluation methods showed that the consistency of rough sets conditional information entropy was better than other evaluation methods. Compared with other decision-making schemes, the rough sets conditional information entropy had higher discrimination and better rank correlations, and was suitable for comprehensive evaluation of climatic conditions in flue-cured tobacco growing areas.