Scaling up data mining techniques to large datasets using parallel and distributed processing

Advances in hardware and software technology enable us to collect, store and distribute large quantities of data on a very large scale. Automatically discovering and extracting hidden knowledge in the form of patterns from these large data volumes is known as data mining. Data mining technology is not only a part of business intelligence, but is also used in many other application areas such as research, marketing and financial analytics. For example medical scientists can use patterns extracted from historic patient data in order to determine if a new patient is likely to respond positively to a particular treatment or not; marketing analysts can use extracted patterns from customer data for future advertisement campaigns; finance experts have an interest in patterns that forecast the development of certain stock market shares for investment recommendations. However, extracting knowledge in the form of patterns from massive data volumes imposes a number of computational challenges in terms of processing time, memory, bandwidth and power consumption. These challenges have led to the development of parallel and distributed data analysis approaches and the utilisation of Grid and Cloud computing. This chapter gives an overview of parallel and distributed computing approaches and how they can be used to scale up data mining to large datasets.

Precipitation scaling with temperature in warm and cold climates: an analysis of CMIP5 simulations

We investigate the scaling between precipitation and temperature changes in warm and cold climates using six models that have simulated the response to both increased CO2 and Last Glacial Maximum (LGM) boundary conditions. Globally, precipitation increases in warm climates and decreases in cold climates by between 1.5%/°C and 3%/°C. Precipitation sensitivity to temperature changes is lower over the land than over the ocean and lower over the tropical land than over the extratropical land, reflecting the constraint of water availability. The wet tropics get wetter in warm climates and drier in cold climates, but the changes in dry areas differ among models. Seasonal changes of tropical precipitation in a warmer world also reflect this “rich get richer” syndrome. Precipitation seasonality is decreased in the cold-climate state. The simulated changes in precipitation per degree temperature change are comparable to the observed changes in both the historical period and the LGM.

Evidence of a universal scaling relationship for leaf CO2 drawdown along an aridity gradient

The leaf carbon isotope ratio (δ13C) of C3 plants is inversely related to the drawdown of CO2 concentration during photosynthesis, which increases towards drier environments. We aimed to discriminate between the hypothesis of universal scaling, which predicts between-species responses of δ13C to aridity similar to within-species responses, and biotic homoeostasis, which predicts offsets in the δ13C of species occupying adjacent ranges. The Northeast China Transect spans 130–900 mm annual precipitation within a narrow latitude and temperature range. Leaves of 171 species were sampled at 33 sites along the transect (18 at ≥ 5 sites) for dry matter, carbon (C) and nitrogen (N) content, specific leaf area (SLA) and δ13C. The δ13C of species generally followed a common relationship with the climatic moisture index (MI). Offsets between adjacent species were not observed. Trees and forbs diverged slightly at high MI. In C3 plants, δ13C predicted N per unit leaf area (Narea) better than MI. The δ13C of C4 plants was invariant with MI. SLA declined and Narea increased towards low MI in both C3 and C4 plants. The data are consistent with optimal stomatal regulation with respect to atmospheric dryness. They provide evidence for universal scaling of CO2 drawdown with aridity in C3 plants.

Towards a parallel computationally efficient approach to scaling up data stream classification

Advances in hardware technologies allow to capture and process data in real-time and the resulting high throughput data streams require novel data mining approaches. The research area of Data Stream Mining (DSM) is developing data mining algorithms that allow us to analyse these continuous streams of data in real-time. The creation and real-time adaption of classification models from data streams is one of the most challenging DSM tasks. Current classifiers for streaming data address this problem by using incremental learning algorithms. However, even so these algorithms are fast, they are challenged by high velocity data streams, where data instances are incoming at a fast rate. This is problematic if the applications desire that there is no or only a very little delay between changes in the patterns of the stream and absorption of these patterns by the classifier. Problems of scalability to Big Data of traditional data mining algorithms for static (non streaming) datasets have been addressed through the development of parallel classifiers. However, there is very little work on the parallelisation of data stream classification techniques. In this paper we investigate K-Nearest Neighbours (KNN) as the basis for a real-time adaptive and parallel methodology for scalable data stream classification tasks.

Turbulent structure and scaling of the inertial subrange in a stratocumulus-topped boundary layer observed by a Doppler lidar

The turbulent structure of a stratocumulus-topped marine boundary layer over a 2-day period is observed with a Doppler lidar at Mace Head in Ireland. Using profiles of vertical velocity statistics, the bulk of the mixing is identified as cloud driven. This is supported by the pertinent feature of negative vertical velocity skewness in the sub-cloud layer which extends, on occasion, almost to the surface. Both coupled and decoupled turbulence characteristics are observed. The length and timescales related to the cloud-driven mixing are investigated and shown to provide additional information about the structure and the source of the mixing inside the boundary layer. They are also shown to place constraints on the length of the sampling periods used to derive products, such as the turbulent dissipation rate, from lidar measurements. For this, the maximum wavelengths that belong to the inertial subrange are studied through spectral analysis of the vertical velocity. The maximum wavelength of the inertial subrange in the cloud-driven layer scales relatively well with the corresponding layer depth during pronounced decoupled structure identified from the vertical velocity skewness. However, on many occasions, combining the analysis of the inertial subrange and vertical velocity statistics suggests higher decoupling height than expected from the skewness profiles. Our results show that investigation of the length scales related to the inertial subrange significantly complements the analysis of the vertical velocity statistics and enables a more confident interpretation of complex boundary layer structures using measurements from a Doppler lidar.

A scanning electron microscopy study of diseased root surfaces conditioned with EDTA gel plus Cetavlon after scaling and root planing

In the present investigation, a scanning electron microscopy analysis was performed to evaluate the effects of the topical application of ethylenediaminetetraacetic acid (EDTA) gel associated with Cetavlon (EDTAC) in removing the smear layer and exposing collagen fibers following root surface instrumentation. Twenty-eight teeth from adult humans, single rooted and scheduled for extraction due to periodontal reasons, were selected. Each tooth was submitted to manual (scaling and root planing) instrumentation alone or combined with ultrasonic instruments, with or without etching using a 24% EDTAC gel. Following extraction, specimens were processed and examined under a scanning electron microscope. A comparative morphological semi-quantitative analysis was performed; the intensity of the smear layer and the decalcification of cementum and dentinal surfaces were graded in 12 sets using an arbitrary scale ranging from 1 (area covered by a smear layer) to 4 (no smear layer). Root debridement with hand instruments alone or combined with ultrasonic instruments resulted in a similar smear layer covering the root surfaces. The smear layer was successfully removed from the surfaces treated with EDTAC, which exhibited numerous exposed dentinal tubules and collagen fibers. This study supports the hypothesis that manual instrumentation alone or instrumentation combined with ultrasonic instrumentation is unable to remove the smear layer, whereas the subsequent topical application of EDTAC gel effectively removes the smear layer, uncovers dentinal openings and exposes collagen fibers.

A new scale-invariant ratio and finite-size scaling for the stochastic susceptible-infected-recovered model

The critical behavior of the stochastic susceptible-infected-recovered model on a square lattice is obtained by numerical simulations and finite-size scaling. The order parameter as well as the distribution in the number of recovered individuals is determined as a function of the infection rate for several values of the system size. The analysis around criticality is obtained by exploring the close relationship between the present model and standard percolation theory. The quantity UP, equal to the ratio U between the second moment and the squared first moment of the size distribution multiplied by the order parameter P, is shown to have, for a square system, a universal value 1.0167(1) that is the same for site and bond percolation, confirming further that the SIR model is also in the percolation class.

Longitudinal scaling property of the charge balance function in Au plus Au collisions at root s(NN)=200 GeV

We present measurements of the charge balance function, from the charged particles, for diverse pseudorapidity and transverse momentum ranges in Au + Au collisions at root S(NN) = 200 GeV using the STAR detector at RHIC. We observe that the balance function is boost-invariant within the pseudorapidity coverage vertical bar-1.3, 1.3 vertical bar. The balance function properly scaled by the width of the observed pseudorapidity window does not depend on the position or size of the pseudorapidity window. This scaling property also holds for particles in different transverse momentum ranges. In addition, we find that the width of the balance function decreases monotonically with increasing transverse momentum for all centrality classes. (c) 2010 Elsevier B.V. All rights reserved.

A note on a practical relationship between filter coefficients and scaling and wavelet functions of Discrete Wavelet Transforms

In this paper, the relationship between the filter coefficients and the scaling and wavelet functions of the Discrete Wavelet Transform is presented and exemplified from a practical point-of-view. The explanations complement the wavelet theory, that is well documented in the literature, being important for researchers who work with this tool for time-frequency analysis. (c) 2011 Elsevier Ltd. All rights reserved.

SCALING LIMIT FOR A DRAINAGE NETWORK MODEL

We consider the two-dimensional version of a drainage network model introduced ill Gangopadhyay, Roy and Sarkar (2004), and show that the appropriately rescaled family of its paths converges in distribution to the Brownian web. We do so by verifying the convergence criteria proposed in Fontes, Isopi, Newman and Ravishankar (2002).

Scaling-Up of Dental Pulp Stem Cells Isolated from Multiple Niches

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Short-term clinical and immunologic effects of scaling and root planing with Er : YAG laser in chronic periodontitis

Background: Recently, the erbium-doped:yttrium, aluminum, and garnet (Er:YAG) laser has been used for periodontal therapy. This study compared Er:YAG laser irradiation (100 mJ/pulse, 10 Hz, 12.9 J/cm(2)) with or without conventional scaling and root planing (SRP) to SRP only for the treatment of periodontal pockets affected with chronic periodontitis.Methods: Twenty-one subjects with pockets from 5 to 9 mm in non-adjacent sites were studied. In a split-mouth design, each site was randomly allocated to a treatment group: SRP and laser (SRPL), laser only (L), SRP only (SRP), or no treatment (C). The plaque index (PI), gingival index (GI), bleeding on probing (BOP), and interleukin (IL)-1 beta levels in crevicular fluid were evaluated at baseline and at 12 and 30 days postoperatively, whereas probing depth (PD), gingival recession (GR), and clinical attachment level (CAL) were evaluated at baseline and 30 days after treatment. A statistical analysis was conducted (P<0.05).Results: Twelve days postoperatively, the PI decreased for SRPL and SRP groups (P<0.05); the GI increased for L, SRP, and C groups but decreased for the SRPL group (P<0.05); and BOP decreased for SRPL, L, and SRP groups (P<0.01). Thirty days postoperatively, BOP decreased for treated groups and was lower than the C group (P<0.05). PD decreased in treated groups (P<0.001), and differences were found between SRPL and C groups (P<0.05). CAL gain was significant only for the SRP group (P<0.01). GR increased for SRPL and L groups (P<0.05). No difference in IL-1 beta was detected among groups and periods.Conclusion: Er:YAG laser irradiation may be used as an adjunctive aid for the treatment of periodontal pockets, although a significant CAL gain was observed with SRP alone and not with laser treatment.