| 一种基于近邻关系的新型离群评估算法 |
| |
| 引用本文： | 张洁玲.一种基于近邻关系的新型离群评估算法[J].福建工程学院学报,2017,0(6):591-596. |
| |
| 作者姓名： | 张洁玲 |
| |
| 作者单位： | 福建江夏学院电子信息科学学院 |
| |
| 摘 要： | 针对传统离群点检测算法的局限性进行研究，利用数据对象之间的相邻关系，提出了一种基于密度和距离相结合的离群检测算法，该算法解决了基于距离的离群检测算法不能准确识别局部离群点的问题，有效避免由于稀疏和密集簇过于邻近的而出现离群点误判的情况。通过在人工模拟数据及真实数据集上的实验测试证明改进算法的可行性，该算法能更有效地检测出数据集中的离群对象。
|
| 关 键 词： | CDD算法 k-近邻 离群评估 |
A new outlier evaluation algorithm based on the nearest neighbor relationship |
| |
| Zhang Jieling.A new outlier evaluation algorithm based on the nearest neighbor relationship[J].Journal of Fujian University of Technology,2017,0(6):591-596. |
| |
| Authors: | Zhang Jieling |
| |
| Institution: | School of Electrical and Information Science, Fujian Jiangxia University |
| |
| Abstract: | Aiming at the limitations of traditional outliers detection algorithm, a new outlier detection algorithm based on the combination of density and distance was proposed according to the neighboring relationships of the data. The new algorithm solves the problem that the distance-based algorithm cannot identify local outliers, and effectively avoids wrong detection of outliers when the sparse clusters and dense clusters are too close. Experiments on artificial and real datasets prove that the improved algorithm is feasible and it can detect the outliers in the datasets more effectively. |
| |
| Keywords: | CDD algorithm k-nearest neighbor outlier evaluation |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《福建工程学院学报》浏览原始摘要信息 |
| 点击此处可从《福建工程学院学报》下载免费的PDF全文 |
