Research And Application Of Set Pair Analysis Theory In Water And Soil Resources

The changes of soil and water resources were affected by natural and human that has the characteristics multidimensional,nonlinearity and lack of samples. In order to solve the multiple complexities of water and soil resources, the new set pair analysis model has been established in water and soil resources comprehensive evaluation and prediction in this research.There are four achievements in this research:1. Improved Set Pair Analysis Model based on Projection Pursuit (RAGAPP-SPA).The relational degree can be obtained by using the optimum projection vector as the index weight and perfected the deficiency that adjacent relational degree ofⅤdegree is -1, and assumed the existence of six degree index. The evaluation results proved that the weight can be determined objectively by using the RAGAPP-SPA and accorded with the actual law. The grade can be divided carefully and the tested data can be utilized greater, and can be applied to research a variety of non-linear and non-normal high-dimensional data.2. Due to the uncertain value problem of variation coefficient i in Set Pair Analysis, this paper used the correlation degree of variation degree and identical degree and opposite degree as the i value, respectively, to calculate i by using absolute correlation degree calculation method of grey theory, and obtained the satisfactory results. The results showed that it was feasible to calculate i based on absolute correlation degree of the grey theory, and provided a new idea and method for solving variation coefficient i in Set Pair Analysis.3. Traditional matter-element model was improved by using set pair analysis in order to reduce the complexity of original model and achieve the uncertainty analysis of matters characteristics, and the improved model was used to evaluate and obtained the satisfactory results. Results show that set pair matter-element model was feasible for evaluating and expanding the application areas based on set pair matter-element model and providing a new way for water resources evaluation.4. Both set pair analysis and statistic test of correlation coefficients can be used to determine the main physics vectors for affecting the changes of water resources. Set pair analysis method can be used to quantificational compute the similarity between the nonce and history main physics vectors from the views of identical, discrepant and contrary sides. The weighted average of the water resources values of the nearest neighbor history samples can be regarded as the predicted value of the nonce water resources, then the set pair analysis was established based on similarity forecast model of water resources change that named SPA-SF. The application results of SPA-SF showed that the statistic and physical concepts of SPA-SF were distinct, its computation method was visual, its precision was high, and the scheme of modeling SPA-SF was simple, effective and perfect. So the SPA-SF can be applied for forecasting different hydrological time series with plenty representative history samples.In this research, the organic combination between theory and practice was achieved by using the above models. At the same time, the multiple complex and nonlinear study on soil and water resources obtained satisfactory results, and provided new technologies for solving some practical problems currently, and provided new idea for soil and water resources system research.