As a newly developed technique for data mining, Support Vector Machine proposed by Vaphik is suitable for the data processing based on finite number of training samples, with special technique to restrict over-fitting...
详细信息
ISBN:
(纸本)1574982656
As a newly developed technique for data mining, Support Vector Machine proposed by Vaphik is suitable for the data processing based on finite number of training samples, with special technique to restrict over-fitting that contributes to powerful prediction ability. In this work, Support Vector Classification and Regression techniques are used to make modeling on the relationship between the sintered cold modulus of rupture and processing parameters of sialon-corundum castable, compared with the modeling of Partial Least Squares and BackPropagation Artificial Neutral Network. It is proved that Support Vector Machine can give the best modeling results and this new method of computation appears to be a more useful tool for optimizing processing parameters for material design.
The I/O routine in network storage has longer transferring time and overhead than local I/O. To well make use of the storage network bandwidth and meet demand of concurrent I/O requests, we need to find the appropriat...
详细信息
The I/O routine in network storage has longer transferring time and overhead than local I/O. To well make use of the storage network bandwidth and meet demand of concurrent I/O requests, we need to find the appropriate schedule policy for I/O requests at the level of the network. The focus in the paper is on how to schedule concurrent I/O data transferring in share storage network, and to improve performance such as the response time affected by scheduling method. Our objective is to analysis the response time of concurrent requests as function their schedule order such as the sequenced, parallel and hybrid policies. We then verify our results by both the analytical method and experiment
While demand for large-scale storage services is growing very rapidly, RAID is still today's standard solution for enterprise class storage systems. The software and hardware designing of a RAID controller based o...
详细信息
It obviously gets limited advantage in massive storage system that all the management works are done by the servers, such as file server and metadata server of traditional storage area network (SAN). Moreover, storage...
详细信息
It obviously gets limited advantage in massive storage system that all the management works are done by the servers, such as file server and metadata server of traditional storage area network (SAN). Moreover, storage device with deficient intelligence leads to lack of supporting local policy and does not scale well in performance and functionality when the number of storage device is added. At the same time, servers, e.g. metadata servers, are easy turned into bottlenecks when load balancing policies only depend on servers to scheme. This paper studied object storage system (OSS) which is a novel solution for massive storage. And the whole object storage system is divided into some domains. Each domain has its own metadata server, object storage controllers and clients (or application servers). Load balancing policies, involved three facets: metadata server (based on globe), object storage controller (based on local) and object, are presented.
Most of traditional load balancing strategies are based on servers or applications architecture. Storage subsystem, as the back tier, provides file or block interface to server or application tier, is often ignored wh...
详细信息
ISBN:
(纸本)076952432X
Most of traditional load balancing strategies are based on servers or applications architecture. Storage subsystem, as the back tier, provides file or block interface to server or application tier, is often ignored when the bottleneck of service is studied. Our load balancing strategy is based on storage object, which provides object interface for the up tier. We concentrate on storage objects with massive dynamic processing. The characteristics of storage object in object storage system (OSS) enable us to apply our load balancing strategy at the back end of servers or applications in order to deal with bottlenecks at the storage tier in data-intensive applications
This paper analyses the mathematical properties of four famous wavelets and evaluates their coding performance for different kinds of images in EZW image coders. It can be found that although the 9/7 biorthogonal wave...
详细信息
This paper analyses the mathematical properties of four famous wavelets and evaluates their coding performance for different kinds of images in EZW image coders. It can be found that although the 9/7 biorthogonal wavelets, which obtain better tradeoff among incompatible mathematical properties, give the best performance. But it still can't code edges and textures efficiently due to existing ringing and smearing artifacts at low bit rates. The harmonic decomposition must follow the uncertainty principle that prevents any atom perfectly, which is well localized in the space domain to retain frequency localization. It gives an opportunity to develop new paradigms to efficiently representing and coding structures very well localized in the space domain, but with large frequency band such as edges and textures.
In the typical streaming media system, the streaming media server is system bottleneck with the expansion of Internet subscribers. This paper proposes an innovational high performance streaming media system architectu...
详细信息
In the typical streaming media system, the streaming media server is system bottleneck with the expansion of Internet subscribers. This paper proposes an innovational high performance streaming media system architecture (HPSMS) based on the logical separation of streaming media transport protocol. The system avoids expensive store-and-forward data copies between streaming media server and storage devices, improves the system performance greatly. The system bandwidth continuously increases with the expansion of storage system capacity is the highlight. The performance of the proposed HPSMS is evaluated through a practical prototype implementation.
Online learning, where feature spaces can change over time, offers a flexible learning paradigm that has attracted considerable attention. However, it still faces three significant challenges. First, the heterogeneity...
详细信息
Online learning, where feature spaces can change over time, offers a flexible learning paradigm that has attracted considerable attention. However, it still faces three significant challenges. First, the heterogeneity of real-world data streams with mixed feature types presents challenges for traditional parametric modeling. Second, data stream distributions can shift over time, causing an abrupt and substantial decline in model performance. Additionally, the time and cost constraints make it infeasible to label every data instance in a supervised setting. To overcome these challenges, we propose a new algorithm Online Learning from Mix-typed, Drifted, and Incomplete Streaming Features (OL-MDISF), which aims to relax restrictions on both feature types, data distribution, and supervision information. Our approach involves utilizing copula models to create a comprehensive latent space, employing an adaptive sliding window for detecting drift points to ensure model stability, and establishing label proximity information based on geometric structural relationships. To demonstrate the model’s efficiency and effectiveness, we provide theoretical analysis and comprehensive experimental results.
Graph Pattern Matching (GPM) entails the identification of subgraphs within a larger graph structure that either precisely mirror or closely parallel a predefined pattern graph. Despite the fact that research on GPM i...
详细信息
Graph Pattern Matching (GPM) entails the identification of subgraphs within a larger graph structure that either precisely mirror or closely parallel a predefined pattern graph. Despite the fact that research on GPM in large-scale graph data has been largely centered on social network analysis or enhancing the precision and efficiency of matching algorithms for expeditious subgraph retrieval, there is a noticeable absence of studies committed to probing GPM in medical domains. To rectify this shortcoming and probe the potential of GPM in clinical contexts, particularly in aiding patients with the selection of optimal tumor treatment plans, this paper introduces the concept of probabilistic graph pattern matching specifically modified for the Tumor Knowledge Graph (TKG). We propose a multi-constraint graph pattern matching algorithm, hereinafter designated as TKG-McGPM, customized for the Tumor Knowledge Graph. Through experimental verification, we establish that TKG-McGPM can facilitate more efficient and informed decision-making in tumor treatment planning.
暂无评论