On the upper bound of the number of modes of a multivariate normal mixture
نویسندگان
چکیده
The main result of this article states that one can get as many as D + 1 modes from just a two component normal mixture in D dimensions. Multivariate mixture models are widely used for modeling homogeneous populations and for cluster analysis. Either the components directly or modes arising from these components are often used to extract individual clusters. Although in lower dimensions these strategies work well, our results show that high dimensional mixtures are often very complex and researchers should take extra precautions when using these for cluster analysis. Even in the simplest case of mixing only two normal components in D dimensions one can generate D + 1 modes. When the components are non-normal or if we have more than two components the number of modes are bound to be even larger, which might lead us to incorrect inference on the number of clusters. Further analysis shows that the number of modes depends on the component means and eigenvalues of the ratio of the two component covariance matrices, which in turn provides a clear guideline as to when one can use mixture analysis for clustering high dimensional data.
منابع مشابه
On trees attaining an upper bound on the total domination number
A total dominating set of a graph $G$ is a set $D$ of vertices of $G$ such that every vertex of $G$ has a neighbor in $D$. The total domination number of a graph $G$, denoted by $gamma_t(G)$, is~the minimum cardinality of a total dominating set of $G$. Chellali and Haynes [Total and paired-domination numbers of a tree, AKCE International ournal of Graphs and Combinatorics 1 (2004), 6...
متن کاملNew results on upper domatic number of graphs
For a graph $G = (V, E)$, a partition $pi = {V_1,$ $V_2,$ $ldots,$ $V_k}$ of the vertex set $V$ is an textit{upper domatic partition} if $V_i$ dominates $V_j$ or $V_j$ dominates $V_i$ or both for every $V_i, V_j in pi$, whenever $i neq j$. The textit{upper domatic number} $D(G)$ is the maximum order of an upper domatic partition. We study the properties of upper domatic number and propose an up...
متن کاملFailure Mode and Effects Analysis Using Generalized Mixture Operators
Failure mode and effects analysis (FMEA) is a method based on teamwork to identify potential failures and problems in a system, design, process and service in order to remove them. The important part of this method is determining the risk priorities of failure modes using the risk priority number (RPN). However, this traditional RPN method has several shortcomings. Therefore, in this paper we p...
متن کاملDetermination of the number of components in finite mixture distribution with Skew-t-Normal components
Abstract One of the main goal in the mixture distributions is to determine the number of components. There are different methods for determination the number of components, for example, Greedy-EM algorithm which is based on adding a new component to the model until satisfied the best number of components. The second method is based on maximum entropy and finally the third method is based on non...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- J. Multivariate Analysis
دوره 108 شماره
صفحات -
تاریخ انتشار 2012