Abstract Clustering as a fundamental unsupervised learning is considered an important method of data analysis, and K -means demonstrably the most popular clustering algorithm. In this paper, we consider on feature space to solve low efficiency caused in Big Data by -means. Different from traditional methods, algorithm guaranteed consistency accuracy before after descending dimension, accelerated when centeres distance functions satisfy certain conditions, completely matched preprocessing...

Database access logs are the starting point for many forms of database administration, from performance tuning, to security auditing, benchmark design, and more. Unfortunately, query also large unwieldy, it can be difficult an analyst extract broad patterns set queries found therein. Clustering is a natural first step towards understanding massive logs. However, clustering methods rely on notion pairwise similarity, which challenging compute SQL queries, especially when underlying data...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Insider threats to databases in the financial sector have become a very serious and pervasive security problem. This paper proposes framework analyze access patterns by clustering SQL queries issued database. Our system Ettu works grouping with other similarly structured queries. The small number of intent groups that result can then be efficiently labeled human operators. We show how our is designed components work. preliminary results accurately models user intent.

Abstract Driven by deep learning techniques in recent years, single target recognition and tracking have developed significantly, but face challenges of real-time accuracy. In this study, an improved IPSO-BP network is formed optimizing three critical aspects the IPSO algorithm: adjusting inertia weight calculation formula, improving factor, creating a new iterative formula for particle updating, which turn combined with BP neural network. After training, paper constructs algorithm higher...

Remote sensing (RS) images are considered to be reflections of the real world. However, RS often suffered from low resolution, making further research difficult follow. Although super resolution (SR) techniques based on deep learning have achieved considerable breakthroughs, they show limited performance when dealing with low-quality complicated backgrounds; for instance, SR results tend loss details and undesired structural distortion. Thus, this paper proposes an innovative dual-branch...

It is essential for a data center to maintain server security and stability. Long-time overload operation or high room temperature may cause service disruption even crash, which would result in great economic loss business. Currently, the methods avoid outages are monitoring forecasting. Thermal camera can provide fine texture information intelligent thermal management large center. This paper presents an efficient method fault diagnosis based on infrared image. Initially distribution of...

Text topic mining and visualization are the basis for clustering topics， distinguishing front topics hot topics. This paper constructs LDA model based on Python language researches mining， dynamic visualization，taking metrology of Library information science in 2017 as an example. In this model，parameter parameter estimated by Gibbs sampling，and best number was determined coherence scores. The can well simulate semantic large corpus，and make corpus not limited to key words. bubble bar graph...

We consider two competing manufacturers who are unreliable and exert effort endogenously to improve their reliability within a dynamic decision framework. The first decide the optimal level of then input quantities after observing improvement outcomes. explore relationship between quantity realised reliability, find that balance effects – price reduction effect cost plays an important role. When market potential is low, dominates effect, resulting in increases reliability. opposite situation...

As a powerful unsupervised learning technique, clustering is the fundamental task of big data analysis. However, many traditional algorithms for that collection high dimension, sparse and noise do not perform well both in terms computational efficiency accuracy. To alleviate these problems, this paper presents Feature K-means model on feature space introduces its fast algorithm based Alternating Direction Multiplier Method (ADMM). We show equivalence original prove convergence iterative...

Digital pathology has aroused widespread interest in modern pathology. The key to digitalization is scan the whole slide image (WSI) at high magnification. file size of each WSI 40 times magnification (40×) may range from 1 gigabyte (GB) 5 GB depending on specimen, which leads huge storage capacity, very slow scanning and network exchange, seriously increasing time costs for digital pathology.We design a strategy slides with low resolution (LR) (5×), superresolution (SR) method proposed...

Convolutional network models have been widely used in image segmentation. However, there are many types of boundary contour features medical images which seriously affect the stability and accuracy segmentation models, such as ambiguity tumors, variability lesions, weak boundaries fine blood vessels. In this paper, order to solve these problems we first introduce dual-tree complex wavelet scattering transform module, then innovatively propose a learning model. addition, new improved active...

Analyzing database access logs is a key part of performance tuning, intrusion detection, benchmark development, and many other administration tasks. Unfortunately, it common for production databases to deal with millions or more queries each day, so these must be summarized before they can used. Designing an appropriate summary encoding requires trading off between conciseness information content. For example: simple workload sampling may miss rare, but high impact queries. In this paper, we...

Augmented reality (AR) technology has been applied in various areas, such as large-scale manufacturing, national defense, healthcare, movie and mass media so on. An important way to realize AR display is using computer-generated hologram (CGH), which accompanied by low image quality heavy computing defects. Meanwhile, the diffraction of Liquid Crystal on Silicon (LCoS) a negative effect quality. In this paper, modified algorithm based traditional Gerchberg-Saxton (GS) was proposed improve...

Level set method has been widely applied in the field of image segmentation. However, level formulation is inevitably affected by regularization function, in-homogeneity and weak edge process evolution, which often leads to instability inaccuracy segmentation results. To solve these problems, a new distance term defined double-well potential function proposed satisfy more ideal characteristics signed property. In addition, novel indicator introduced segment images with uneven intensity or...

SOC storage (SOCS) plays a vital role in global climate change. Understanding the spatial pattern and features of soil organic carbon (SOC) its influencing factors is important for increasing fixation. However, few studies exist on reserves farmland regional scale. This study revealed SOCD SOCS values distribution using Hubei Province as sampling region. The results demonstrated that distributions system density were uneven, heterogeneity was related to geography, cultivated area, type....