|
|||||
|
|
| 自适应最近邻预测(英文) | |
| 引用本文： | 张衍华,陈宇明,陈桂景. 自适应最近邻预测(英文)[J]. 安徽大学学报(自然科学版), 1999, 23(2): 9-16,24 |
| 作者姓名： | 张衍华 陈宇明 陈桂景 |
| 作者单位： | 1. 山东建筑工程学院,济南,250013 2. 安徽大学计算机系 3. 安徽大学数学系 |
| 基金项目： | 国家自然科学基金;19671001; |
| 摘 要： | 研究了在最近邻预测中如何选择最近邻的阶数的问题。考虑了两种根据训练样本来确定近邻阶数ｋｎ与ｋｎ的方法。还证明了使用这些近邻阶数来进行预测可以达到最优预测以及进行了随机模拟计算，计算结果与文中理论相当吻合，并且揭示了最优近邻阶数ｋ０ｎ与样本量ｎ之间关系。 |
| 关 键 词： | 最近邻预测，预测风险函数，最优近邻阶数，自适应最近邻预测 |
Adaptive Nearest Neighbour Prediction |
|
| Zhang Yanhua. Adaptive Nearest Neighbour Prediction[J]. Journal of Anhui University(Natural Sciences), 1999, 23(2): 9-16,24 | |
| Authors: | Zhang Yanhua |
| Affiliation: | Zhang Yanhua (Shandong Architectural Institute Jinan 250013) Chen Yuming (Department of Computer Science Anhui University) Chen Guijing (Department of Mathematics Anhui University) |
| Abstract: | Here the problem how to choose the order number of Nearest Neighbour prediction has been studied. We consider two kinds of order numbers:adaptive order number k * nand the order number k nwhich independent of sample. The asymptotical optimality of these prediction procedures have been proved.Our studying of Monte Carlo simulation support this theory well, and explore the relation between k 0 nand training sample size n . |
| Keywords: | NN-prediction Bayesian risk optimal order of nearest neighbour strong consistency. |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |