The Research On Attribute Weight Based On Rough Set In The Forest Ecosystem Health Evaluation

Forest health assessment has always been a hot problem in the study on forest management, which its research has provided scientific basis for sustainable forest management. And to determine the weigh of assessment indexes is indispensable in the process of forest health assessment. It can not only directly affect the final evaluation; also can directly affect the final decision results. In forest health evaluation, there were some problems in the previous method to determine the weight, such as too much subjective factors, the complex calculation and so on.Rough set theory to show the idea of "talk" with objective data is that it only dealt information with data, without prior knowledge. Combining with the characteristics of the uncertainty of forest health evaluation index, this paper puts forward using rough set to determine weights in forest health assessment. With objective data driven, the method of rough set determined weights of each index factor affect the forest health objective by analyzing internal rules through the forest system attribute data inquiry system. It has a strong theoretical and practical significance to improve the accuracy of forest health assessment result. In this paper, the main research contents and research conclusions are as follows:(1) In view of the shortcomings of commonly used method for determining weights, this paper studies the important degree of each attribute by mining method to determine the weight of the rough set objectively. According to the lack of decision attribute, knowledge representation system of rough set theory can be divided into information system and decision-making system. And this paper summarizes several kinds of definition of attribute importance. Through example calculation, the results show that:in the information system, with ignored the importance degree of attribute itself, the determining weight methods based on knowledge granularity and the information appeared that some phenomenon of attribute weights was zero. The paper improved the methods by attributing their own importance. So it made the result more reasonable. The running time of the method based on knowledge granularity is superior to other methods in the information system.(2) According to the decision attribute unknown in Daweishan nature reserve, this paper applied the improved method based on knowledge granularity to determine the weighing values for assessment index of Dawei mountain forest health. Results showed that this method mine correlation and important degree between each factor directly which are from the data of influence of forest health indicators. It did not depend on experts’experiences, the greater the amount of information, the more objective weight is. The results showed that the weight of13indexes were basically consistent with the expert scoring results. But there were deviation, such as crown density and fire danger rating index is not consistent.(3) Based on rough sets and the knowledge of the forest health assessment, the research designed and developed the forest health assessment system. The system took forest health evaluation index system, input and output of evaluation index, index data calculation, weight determination and evaluation method as one. This system was verified by example, the result showed that it made the forest health assessment more objective and effective and reduced the deficiency of the traditional method.