Search

In the information age, smart data modelling and data management can be carried out to address the wealth of data produced in scientific experiments. In this paper, we propose a semantic model for the statistical analysis of datasets by linear mixed models. We tie together disparate statistical concepts in an interdisciplinary context through the application of ontologies, in particular the Statistics Ontology (STATO), to produce FAIR data summaries. We hope to improve the general understanding of statistical modelling and thus contribute to a better description of the statistical conclusions from data analysis, allowing their efficient exploration and automated processing. [ABSTRACT FROM AUTHOR]

This paper develops a methodology for the estimation of the global effect of a treatment from multiple surrogates collected using a longitudinal design. The global effect is modelled as a latent function of time which is allowed to vary for different treatment groups. The latent variable is modelled using low-rank thin plate splines, allowing the shape of the latent profiles to be estimated from data. Identifiability issues are addressed and Bayesian Linear Mixed Model formulation is used which allows for the implementation of WinBUGS. The method is validated through a simulation study and has been implemented on a motivating dataset to measure the effect of exposure to polluted air on lung function. [ABSTRACT FROM AUTHOR]

It is common practice to compare the fit of non‐nested models using the Akaike (AIC) or Bayesian (BIC) information criteria. The basis of these criteria is the log‐likelihood evaluated at the maximum likelihood estimates of the unknown parameters. For the general linear model (and the linear mixed model, which is a special case), estimation is usually carried out using residual or restricted maximum likelihood (REML). However, for models with different fixed effects, the residual likelihoods are not comparable and hence information criteria based on the residual likelihood cannot be used. For model selection, it is often suggested that the models are refitted using maximum likelihood to enable the criteria to be used. The first aim of this paper is to highlight that both the AIC and BIC can be used for the general linear model by using the full log‐likelihood evaluated at the REML estimates. The second aim is to provide a derivation of the criteria under REML estimation. This aim is achieved by noting that the full likelihood can be decomposed into a marginal (residual) and conditional likelihood and this decomposition then incorporates aspects of both the fixed effects and variance parameters. Using this decomposition, the appropriate information criteria for model selection of models which differ in their fixed effects specification can be derived. An example is presented to illustrate the results and code is available for analyses using the ASReml‐R package. Residual maximum likelihood is a standard method of analysis for general linear models that arise, for example, from linear mixed models. The residual likelihood is determined by the model for the fixed effects. If model selection between models with differing fixed effects is required, using the Akaike or Bayesian information criterions based solely on the residual likelihood is inappropriate. However, it is shown that the two criteria can be used based on the full likelihood evaluated at the residual maximum likelihood estimates of the parameters. [ABSTRACT FROM AUTHOR]

Abstract Linearizability is a well-known consistency condition for shared objects in concurrent systems. We focus on the problem of implementing linearizable objects of arbitrary data types in message-passing systems with bounded, but uncertain, message delay and bounded, but non-zero, clock skew. We present an algorithm that exploits axiomatic properties of different operations to reduce the running time of each operation below that obtainable with previously known algorithms. We also prove lower bounds on the time complexity of various kinds of operations, specified by the axioms they satisfy, resulting in reduced gaps in some cases and tight bounds in others. [ABSTRACT FROM AUTHOR]

Model selection can be defined as the task of estimating the performance of different models in order to choose the most parsimonious one, among a potentially very large set of candidate statistical models. We propose a graphical representation to be considered as an extension to the class of mixed models of the deviance plot proposed in the literature within the framework of classical and generalized linear models. This graphical representation allows, once a reduced number of models have been selected, to identify important covariates focusing only on the fixed effects component, assuming the random part properly specified. Nevertheless, we suggest also a standalone figure representing the residual random variance ratio: a cross-evaluation of the two graphical representations will allow to derive some conclusions on the random part specification of the model and a more accurate selection of the final model. [ABSTRACT FROM AUTHOR]

Nonnegative matrix factorization (NMF) has become a prominent signal processing and data analysis technique. To address streaming data, online methods for NMF have been introduced recently, mainly restricted to the linear model. In this paper, we propose a framework for online nonlinear NMF, where the factorization is conducted in a kernel-induced feature space. By exploring recent advances in the stochastic gradient descent and the mini-batch strategies, the proposed algorithms have a controlled computational complexity. We derive several general update rules, in additive and multiplicative strategies, and detail the case of the Gaussian kernel. The performance of the proposed framework is validated on unmixing synthetic and real hyperspectral images, comparing to state-of-the-art techniques. [ABSTRACT FROM AUTHOR]

Humpback whales are among the most common whale species occurring in Skjálfandi Bay, North Iceland. Since 2001, 309 photographed humpback whales have been identified and catalogued. The percentage of whales re-sighted from year to year has been always less than 27% and only 49 out of 309 identified individuals were sighted in the Bay in two or more years, indicating that the humpback whales occurring in Skjálfandi Bay have various feeding areas. Kernel Density Estimations indicated distribution changes within the bay and throughout the season. In early summer, humpback whale densities were higher in the southern and western part of Skjálfandi Bay, whereas as the season progressed whale concentration increased further north in the direction of the open sea. Generalised Linear Models (GLMs) were used to analyse a 10 year time series of monthly humpback whale Sightings Per Unit of Effort (SPUE; sightings per minute) during the summer season (May-October). Whalewatching boats were used as research platforms. The total effort was 136,503 minutes. 1,401 sightings of humpback whales were recorded. GLMs show significantly higher SPUE in 2006 (0.0132), 2011 (0.0111) and 2012 (0.0246) when compared to the start of the time series. September showed a significantly lower SPUE (0.0024) when compared to the baseline month (May). An independent dataset derived from logbook data provided by commercial whalewatching operations was used to derive the percentage of humpback whale positive surveys (surveys where the species was seen) over a longer time series (1995-2012). These data were analysed using Generalised Additive Models (GAMs) and show an increase on the percentage of positive surveys in the period 2006-09, a slight decrease in 2010 and a new increase in 2011-2012. Annual trends in positive surveys for two additional species of locally abundant cetaceans, the harbour porpoise and the white-beaked dolphin, were explored for comparison purposes. No trend was found for porpoises while the percentage of positive surveys for white-beaked dolphins showed a general decrease after 2002. Fluctuations in SPUE and percentage of positive surveys may be due to varying environmental conditions within the bay, with potential implications for whalewatching operations. Given the scientific value of these conclusions, this study also supports the value of integrating whalewatching vessels as opportunistic platforms for cetacean research. [ABSTRACT FROM AUTHOR]

Copyright of Ilkogretim Online is the property of Ilkogretim Online and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper investigates a semi-parametric model for functional data, based on partial linear ideas. A methodology is developped for testing the linear component of such a functional partial linear model. The behavior of the test is studied through some finite simulated samples before being applied to some chemometrical curves dataset. One interesting feature of the methodology is that it works with not necessarily independent data. [ABSTRACT FROM AUTHOR]

Abstract: Using data from a realistic experiment conducted using recorded traffic sounds, the relationship between indicators of magnitude, spectrum, and temporal evolution of the sonic environment and the reported annoyance are analyzed. The exposure is characterized by the binaurally recorded overall indoor sound. It is found that a series of parameters derived from multiple linear models related to the temporal and spectral structure of the sound pressure level produces better models than the overall indoor A-weighted equivalent noise level. The descriptors thus complement this indicator, at least when exposure is based on overall indoor sound. Principle components amongst the exposure indicators studied relate to the detectability of the sound indoors and to the typical temporal difference between road and rail traffic. Linear models based on these indicators also outperform models based on source related façade. [ABSTRACT FROM AUTHOR]

Advances in semiconductor technology enable larger processor design spaces, leading to increasingly complex systems. At an initial stage, designers must evaluate many architecture design points to achieve a suitable design. Currently, most architecture exploration is performed using cycle accurate simulators. Although accurate, these tools are slow, thus limiting a comprehensive design search. The vast design space of today's complex processors and time to market economic pressures motivate the need for faster architectural evaluation methods. This paper presents a superscalar processor performance model that enables rapid exploration of the architecture design space for superscalar processors. It supplements current design tools by quickly identifying promising areas for more thorough and time consuming exploration with traditional tools. The model estimates the instruction throughput of a superscalar processor based on early architectural design parameters and application properties. It has been validated with the SimpleScalar out-of-order simulator. The core of the model, which executes 1.6 million times faster, produces instruction throughput estimates that are with within 5.5 percent of the corresponding SimpleScalar values. [ABSTRACT FROM AUTHOR]