孤立点

基于ICA与SVM的孤立点挖掘模型

Outlier Mining Model Based on ICA & SVM

本文针对SVM的模型、核函数的构造、SVM参数选择和孤立点检测四个方面进行了研究。

This paper does the researches on four areas : SVM model , kernel function constructing , SVM parameter selection and outlier detection .

设G是一个没有孤立点的简单图。

Let G be a simple graph with no isolated vertices .

基于孤立点和初始质心选择的k均值算法的改进与应用

Application of an improved k-means algorithm based on outliers and original clustering center

另外，本文的XML孤立点数据清理算法也能达到较高的准确率和查全率。

In addition , outlier data cleaning method for XML can achieve higher accuracy and the recall level .

首先，对kmeans算法中孤立点检测问题进行深入研究，提出了基于网格的数据预处理算法。

Firstly , after outlier detection problem of the k_means algorithm is studied deeply , a grid-based data pre-processing algorithm is proposed .

基于孤立点检测的RFID数据流清洗技术研究

Research on Cleaning Techniques over RFID Data Stream Based on Outlier Detection Methods

基于距离的分布式RFID数据流孤立点检测

Distant-based Outlier Detection for Distributed RFID Data Streams

基于统计聚类RBF神经网络的孤立点检测研究

A New Isolated Point Detecting Algorithm Based on Statistical Clustering RBF Neural Network

设G是有n个点的图，G的一个匹配是指G的一个生成子图，它的每个分支或是孤立点或是孤立边。

One match of G is a generating subgraph of G. Every branch is a single point or a single side .

基于PCA及属性距离和的孤立点检测算法

Algorithm for outlier detection based on principal component analysis and sum of attributes distance

基于改进微粒群算法的K-MEANS聚类和孤立点查找

K-Means Clustering and Outlier Detection Based on PSO

针对经典K–means算法易受噪声和孤立点影响这一缺点，对算法做了进一步改进，以减少噪声和孤立点对聚类效果的影响。

Point to classical K-means algorithm vulnerable to noise and the impact of isolated point defects I have improve the algorithm to reduce the noise and isolated points on the cluster effect .

K均值算法的聚类个数K需指定，聚类结果与数据输入顺序相关，而且易受孤立点影响。

K-means algorithm has some deficiencies . The number K must be pointed and its effectiveness liable to be effected by isolated data and the input sequence of data .

基于CD-Tree与SOD，设计了新的孤立点检测算法。

Finally , the CD-Tree-based algorithm is designed for outlier detection based on CD-Tree and SOD .

比如初始聚类数K要事先指定，初始聚类中心选择存在随机性，算法容易生成局部最优解，受孤立点的影响很大等。

For example we must choose the initial clustering number . The choice of initial clustering centre has randomness . The algorithm receives locally optimal solution easily , the effect of isolated point is serious .

在此基础上提出一种改进的K-Means算法，主要是改进原来的算法对孤立点比较敏感的缺点。

Based on it , an improved algorithm of K-Means is proposed , it can conquer disadvantage that customary algorithm is effected by the isolated point .

然而单纯的Hausdorff距离对噪声和孤立点均比较敏感，导致误匹配率较高。

Hausdorff distance , however simple on the noise and isolated points are more sensitive , resulting in a higher rate of false matches .

限制边割将连通图分离成不含孤立点的不连通图，如果最小限制边割只能分离孤立边，则称图G是超级限制边连通的。

Restricted edge cut separates a connected graph into a disconnected one without isolated vertex . Graph G is super restricted edge connected if no subgraph but an isolated edge can be separated by any minimum restricted edge cut .

该方法在去噪前，先用定位精度高的小尺度LOG算子检测图像的边缘，对检测出的边缘进行均值平滑滤波，以减少边缘图像中的孤立点噪声；

Before denoising , the edge of a noised image was detected with small-scale LOG operator which had higher orientation precision , and the image edge was smoothed with average filter to reduce a great lot of isolated point .

此外，我们还将提出的算法与现有的一些算法在KDD常用数据集上进行实验比较，并得到了更好的孤立点的去除性能。

Furthermore , we also compared our algorithm against some existing methods on the top of the KDD dataset and got better outlier removal performance .

介绍了文档聚类中基于划分的k-means算法，k-means算法适合于海量文档集的处理，但它对孤立点很敏感。

This paper first introduces the partitioning-based k-means algorithms for documents clustering . The k-means algorithm adapts to processing the vast amount of documents , but it is sensitive to outliers .

为了克服K-Means方法对孤立点敏感性的缺点，并进一步提高聚类的质量和时间效率，本文将基于密度的聚类算法应用于文本对象之上。

To eliminate the sensitivity to outliers in K-Means and to improve the clustering efficiency and performance further more , density-based clustering algorithm is applied to document clustering in this thesis .

其次在对聚类算法进行研究总结的基础上，提出了将粒子群优化算法和k-means算法相结合的PSO+孤立点+k-means的新型聚类算法。

Second the subject of particle swarm optimization is put forward to work within the clustering algorithm and k-means algorithm combined together , which is efficient and convenient in multidimensional clustering method .

考虑极小极大定理中所获临界点集是否含有鞍点的问题，在不假设临界值为R1中孤立点的情况下获得了鞍点的存在性定理。

The existence of the saddle points generated by Mini-max theorem is studied under the case that the critical value need not be isolated point in R.

为了消除普通FCM算法中随机初始化和孤立点对算法聚类效果的影响，本文提出了改进FCM算法。

In order to eliminate the effect of the random initialization and isolated point on the clustering result in the fuzzy C means ( FCM ) algorithm , an improved FCM algorithm is presented .

该算法是对基于局部稀疏系数（LSC）孤立点挖掘论文中局部稀疏率和局部稀疏系数计算的一种改进。

This algorithm is an improvement of local sparsity ratio and local sparsity coefficient computation for Local Sparsity Coefficient-Based ( LSC ) Mining of Outliers paper .

有限混合（FM）模型已经广泛地应用于图像分割，但是由于没有考虑空间信息，导致分割的结果对噪声很敏感，分割出的区域存在很多杂散的孤立点。

The conventional finite mixture model ( FM ), being widely used in image segmentation , does not take the spatial information into account , which leads this model to work only on well defined images .

再考虑到现实中的数据质量问题，鉴于树算法对孤立点有免疫力和自动处理缺失数据的优点，所以选择CART树算法作为主要建模工具。

When data quality issue in reality is also taken into account , we choose CART as the modeling tool , given that tree algorithm is immune to outliers and can deal with missing values automatically .

相对传统的k-means算法，本文算法不仅具有能有效地处理孤立点，有较好的抗噪声能力，而且不需要设置簇（聚类中心）数目的特点。

Contrast to the traditional k-means algorithm , the algorithm has the features of dealing with the isolated dot effectively , having better anti-noise capability , and not needing to set the number of cluster .