KNN算法分析及研究，将训练元组和测试元组看作是n维（若元组有n的属性）空间内的点，给定一条测试元组，搜索n维空间，找出与测试元组最相近的k个点（即训练元组），最后取这k个点中的多数类作为测试元组的类别。-Yuan KNN algorithm analysis and research, the training set and test set as a n d have the n properties (if a tuple) within the space of points, given a test tuples, search n dimensional space, find out and test a tuple closest point of k (i.e., training tuples), finally take most of these k points such as testing a tuple categories.