加权空间模糊动态聚类算法在土壤肥力评价中的应用

 【目的】改进和提高空间模糊聚类算法。【方法】首先利用层次分析法得到各属性的权值,然后将权值与空间模糊动态聚类法相结合,最后利用概率统计中的F分布来确定最佳分类,以提高空间模糊聚类算法的智能性。【结果】加权空间模糊动态聚类算法与基于模糊等价关系的传递闭包方法进行比较表明,当λ取0.993时,F值最大,分类效果最好。此时,加权的F值为4.898,未加权的F值为2.957,说明加权的类间的差距比未加权的明显,即该算法聚类准确率要明显高于未加权的模糊聚类算法。【结论】将其改进的算法运用到精准农业的土壤肥力评价中,试验结果与实际情况相符,证明了该算法的有效性。

Application of Weighted Spatially Fuzzy Dynamic Clustering Algorithm in Evaluation of Soil Fertility

【Objective】 As a traditional fuzzy clustering algorithm has its shortages, an improved algorithm was presented in the paper. 【Method】 First, access to the weighted value of each attribute using AHP, and then add weighted value to the spatial fuzzy dynamic clustering algorithm, finally, use F- distribution of probability statistics to determine the best classification number in order to improve the algorithm intelligence. 【Result】 The weighted spaces fuzzy dynamic cluster algorithm was compared with the fuzzy equivalent relations transitive closure algorithm, the result shows that F-value is the largest and the classification results is best when λ=0.993. This time, the weighted F- value is 4.898, unweighted F- value is 2.957, that shows the weighted gap between class are more obvious that unweighted one, that is, the accuracy of the clustering algorithm is significantly higher than unweighted fuzzy clustering algorithm. 【Conclusion】 Tests show that the clustering algorithm’s accurate rate is higher than the unweighted fuzzy clustering algorithm.