The k-means algorithm is one of the most often used method for data clustering. However, the standard k-means can only be applied in the original feature space. The kernel k-means, which extends k-means into the kerne...
详细信息
ISBN:
(纸本)9781577357384
The k-means algorithm is one of the most often used method for data clustering. However, the standard k-means can only be applied in the original feature space. The kernel k-means, which extends k-means into the kernel space, can be used to capture the non-linear structure and identify arbitrarily shaped clusters. Since both the standard k-means and kernel k-means apply the squared error to measure the distances between data points and cluster centers, a few outliers will cause large errors and dominate the objection function. Besides, the performance of kernel method is largely determined by the choice of kernel. Unfortunately, the most suitable kernel for a particular task is often unknown in advance. In this paper, we first present a robust k-means using 2,1-norm in the feature space and then extend it to the kernel space. To recap the powerfulness of kernel methods, we further propose a novel robust multiple kernel k-means (RMKKM) algorithm that simultaneously finds the best clustering label, the cluster membership and the optimal combination of multiple kernels. An alternating iterative schema is developed to find the optimal value. Extensive experiments well demonstrate the effectiveness of the proposed algorithms.
The mining of workflow process aims at finding valuable objective information from log data. It leads useful implications for new business processes and analysis. Unfortunately most of business process data is incompl...
详细信息
This paper presents a discrete-time single-server finite-buffer queue with Markovian arrival process and generally distributed batch-size-dependent service time. Given that infinite service time is not commonly encoun...
详细信息
This paper presents a discrete-time single-server finite-buffer queue with Markovian arrival process and generally distributed batch-size-dependent service time. Given that infinite service time is not commonly encountered in practical situations, we suppose that the distribution of the service time has a finite support. Recently, a similar continuous-time system with Poisson input process was discussed by Banerjee and Gupta (2012). But unfortunately, their method is hard to apply in the analysis of discrete-time case with versatile Markovian point process due to the fact that the difference equation governing the boundary state probabilities is more complex than the continuous one. If we follow their ideas, we will eventually find that some important joint queue length distributions cannot be computed and thus some key performance measures cannot be derived. In this paper, replacing the finite support renewal distribution with an appropriate phase-type distribution, the joint state probabilities at various time epochs (arbitrary, pre-arrival and departure) have been obtained by using matrix analytic method and embedded Markov chain technique. Furthermore, UL-type RG-factorization is employed in numerical computation of block-structured Markov chains with finitely-many levels. Some numerical examples are presented to demonstrate the feasibility of the proposed algorithm for several service time distributions. Moreover, the impact of the correlation factor on loss probability and mean sojourn time is also investigated. (C) 2015 Elsevier BY. All rights reserved.
Process control algorithms constitute a specific software model, as they have well-defined and constant sets of inputs and outputs, and perform calculations cyclically in real time. Therefore, the algorithms rarely ar...
详细信息
This paper proposes the automatic generation of the middle frame and the middle frame automatic coloring method of two-dimensional animation process, users simply given starting key frames and end key frames, Accordin...
详细信息
This paper presents a discrete-time single-server finite-buffer queue with Markovian arrival process and generally distributed batch-size-dependent service time. Given that infinite service time is not commonly encoun...
详细信息
This paper presents a discrete-time single-server finite-buffer queue with Markovian arrival process and generally distributed batch-size-dependent service time. Given that infinite service time is not commonly encountered in practical situations, we suppose that the distribution of the service time has a finite support. Recently, a similar continuous-time system with Poisson input process was discussed by Banerjee and Gupta (2012). But unfortunately, their method is hard to apply in the analysis of discrete-time case with versatile Markovian point process due to the fact that the difference equation governing the boundary state probabilities is more complex than the continuous one. If we follow their ideas, we will eventually find that some important joint queue length distributions cannot be computed and thus some key performance measures cannot be derived. In this paper, replacing the finite support renewal distribution with an appropriate phase-type distribution, the joint state probabilities at various time epochs (arbitrary, pre-arrival and departure) have been obtained by using matrix analytic method and embedded Markov chain technique. Furthermore, UL-type RG-factorization is employed in numerical computation of block-structured Markov chains with finitely-many levels. Some numerical examples are presented to demonstrate the feasibility of the proposed algorithm for several service time distributions. Moreover, the impact of the correlation factor on loss probability and mean sojourn time is also investigated. (C) 2015 Elsevier BY. All rights reserved.
Identifying optimal management policies for systems made up by similar components is a challenging task, due to dependence in the components' behavior. In this setting, observations collected on one component are ...
详细信息
ISBN:
(纸本)9780888652454
Identifying optimal management policies for systems made up by similar components is a challenging task, due to dependence in the components' behavior. In this setting, observations collected on one component are also relevant for learning the behavior of others. Probabilistic graphical models allow for consistent inference using all available data, taking dependence among components into account, while optimizing system operation. In this paper we propose a framework for management of systems made by similar components based on hierarchical Bayesian modeling, called Multiple Uncertain Partially Observable Markov Decision Processes (MU-POMDP), that overcomes some limitations of a previously proposed approaches. We describe a detailed numerical algorithm to learn the system parameters within this framework and we investigate its performance with an example of management of a wind farm (i.e., the system) made up by turbines of the same type (i.e., the components).
Recent progress in micro-/nanotechnologies, related to the manufacturing and control strategies has enabled the micro-electro-mechanical systems (MEMS) actuators and sensors. A new developing field has recently appear...
详细信息
In order to ensure recognition accuracy, intelligent traffic video tracking system usually requires various types of information. Therefore, multi-features fusion becomes a good choice. In this paper, a new recognitio...
详细信息
In order to detect and count pedestrians in different kinds of scenes, this paper put forward a method of solving the problem on video sequences captured from a fixed camera. After preprocessing operations on the orig...
详细信息
暂无评论