BAMLSS: Bayesian Additive Models for Location,Scale, and Shape (and Beyond)

Bayesian analysis provides a convenient setting for the estimation of complex generalized additive regression models (GAMs). Since computational power has tremendously increased in the past decade, it is now possible to tackle complicated inferential problems, for example, with Markov chain Monte Carlo simulation, on virtually any modern computer. This is one of the reasons why Bayesian methods have become increasingly popular, leading to a number of highly specialized and optimized estimation engines and with attention shifting from conditional mean models to probabilistic distributional models capturing location, scale, shape (and other aspects) of the response distribution. To embed many different approaches suggested in literature and software, a unified modeling architecture for distributional GAMs is established that exploits distributions, estimation techniques (posterior mode or posterior mean), and model terms (fixed, random, smooth, spatial,…). It is shown that within this framework implementing algorithms for complex regression problems, as well as the integration of already existing software, is relatively straightforward. The usefulness is emphasized with two complex and computationally demanding application case studies: a large daily precipitation climatology, as well as a Cox model for continuous time with space-time interactions. Supplementary material for this article is available online.     

Combined prostate diffusion tensor imaging and dynamic contrast enhanced MRI at 3T — quantitative correlation with biopsy

The purpose of this work was to compare diagnostic accuracy of Diffusion Tensor Imaging (DTI), dynamic contrast-enhanced magnetic resonance imaging (DCE MRI) and their combination in diagnosing prostate cancer. Twenty-five patients with clinical suspicion of prostate cancer underwent MRI, prior to transrectal ultrasound-guided biopsies. MRI data were correlated to biopsy results. Logistic regression models were constructed for the DTI parameters, DCE MRI parameters, and their combination. The areas under the receiver operator characteristic curves (AUC) were compared between the models. The nonparametric Wilcoxon signed rank test was used for statistical analysis. The sensitivity and specificity values were respectively 81% (74–87%) and 85% (79–90%) for DTI and 63% (55–70%) and 90% (85–94%) for DCE. The combination “DTI or DCE MRI” had 100% (97–100%) sensitivity and 77% (69–83%) specificity, while “DTI and DCE MRI” had 44% (37–52%) sensitivity and 98% (94–100%) specificity. The AUC for DTI+DCE parameters was significantly higher than that for either DTI (0.96 vs. 0.92, P=.0143) or DCE MRI parameters (0.96 vs. 0.87, P=.00187) alone. In conclusion, the combination of DTI and DCE MRI has significantly better accuracy in prostate cancer diagnosis than either technique alone.     

Functional connectivity between task-positive and task-negative brain areas and its relation to working memory performance

Functional brain imaging studies have identified a set of brain areas typically activated during cognitive tasks (task-positive brain areas) and another set of brain areas typically deactivated during cognitive tasks (task-negative brain areas). Negative correlations, or anticorrelations, between task-positive and task-negative brain areas have been reported at rest. Furthermore, the strength of these anticorrelations appears to be related to cognitive function. However, studies examining anticorrelations have typically employed global regression or similar analysis steps that force anticorrelated relationships to exist between brain areas. Therefore the validity of these findings has been questioned. Here we examine anticorrelations between a task-negative region in the medial frontal gyrus/anterior cingulate cortex and dorsolateral prefrontal cortex, a classic task-positive area, using an analysis that does not include global regression. Instead, we control for whole-brain correlations in the group-level analysis. Using this approach, we demonstrate that the strength of the functional connection between the medial frontal cortex and the dorsolateral prefrontal cortex is related to cognitive function and that this relationship is not an artifact of global regression.     

A strategy for multivariate calibration based on modified single-index signal regression: Capturing explicit non-linearity and improving prediction accuracy

In this paper, a modified single-index signal regression (mSISR) method is proposed to construct a nonlinear and practical model with high-accuracy. The mSISR method defines the optimal penalty tuning parameter in P-spline signal regression (PSR) as initial tuning parameter and chooses the number of cycles based on minimizing root mean squared error of cross-validation (RMSECV). mSISR is superior to single-index signal regression (SISR) in terms of accuracy, computation time and convergency. And it can provide the character of the non-linearity between spectra and responses in a more precise manner than SISR. Two spectra data sets from basic research experiments, including plant chlorophyll nondestructive measurement and human blood glucose noninvasive measurement, are employed to illustrate the advantages of mSISR. The results indicate that the mSISR method (i) obtains the smooth and helpful regression coefficient vector, (ii) explicitly exhibits the type and amount of the non-linearity, (iii) can take advantage of nonlinear features of the signals to improve prediction performance and (iv) has distinct adaptability for the complex spectra model by comparing with other calibration methods. It is validated that mSISR is a promising nonlinear modeling strategy for multivariate calibration.     

Uniform limit laws of the logarithm for nonparametric estimators of the regression function in presence of censored data

In this paper, we establish uniform-in-bandwidth limit laws of the logarithm for nonparametric Inverse Probability of Censoring Weighted (I.P.C.W.) estimators of the multivariate regression function under random censorship. A similar result is deduced for estimators of the conditional distribution function. The uniform-in-bandwidth consistency for estimators of the conditional density and the conditional hazard rate functions are also derived from our main result. Moreover, the logarithm laws we establish are shown to yield almost sure simultaneous asymptotic confidence bands for the functions we consider. Examples of confidence bands obtained from simulated data are displayed.      

An improved ensemble learning machine for biological activity prediction of tyrosine kinase inhibitors

Boosting is one of the most important strategies in ensemble learning because of its ability to improve the stability and performance of weak learners. It is nonparametric, multivariate, fast and interpretable but is not robust against outliers. To enhance its prediction accuracy as well as immunize it against outliers, a modified version of a boosting algorithm (AdaBoost R2) was developed and called AdaBoost R3. In the sampling step, extremum samples were added to the boosting set. In the robustness step, a modified Huber loss function was applied to overcome the outlier problem. In the output step, a deterministic threshold was used to guarantee that bad predictions do not participate in the final output. The performance of the modified algorithm was investigated with two anticancer data sets of tyrosine kinase inhibitors, and the mechanism of inhibition was studied using the relative weighted variable importance procedure. Investigating the effect of base learner's strength reveals that boosting is only successful using the classification and regression tree method (a weak to moderate learner) and does not have a significant effect using the radial basis functions partial least square method (a strong base learners). Copyright © 2015 John Wiley & Sons, Ltd.     

Prediction of High Performance Fibers Strength Using Back Propagation Neural Network

Nowadays, quantification of the effects of basic parameters such as precursor, temperature oxidation, residence time, low temperature carbonization (LTC) and high temperature carbonization (HTC) on production process polyacrylonitrile based carbon fibers is not completely understood. In this way, there is not a completely theoretical model that accomplishes to quantitatively describe production process carbon fibers very accurately which needs to be used by engineers in design, simulation and operation of that process. This paper presents the development of a back propagation neural network model for the prediction of carbon fibers produced from PAN fibers. The model is based on experimental data. The precursors, temperature oxidation, residence time, LTC and HTC have been considered as the input parameters and the strength as output parameter to develop the model. The developed model is then compared with experimental results and it is found that the results obtained from the neural network model are accurate in predicting the strength of carbon fibers.     

Qsar Models for Predicting the Acute Toxicity of Selected Organic Chemicals with Diverse Structures to Aquatic Non-Vertebrates and Humans

Abstract

The linear and non-linear relationships of acute toxicity (as determined on five aquatic non-vertebrates and humans) to molecular structure have been investigated on 38 structurally-diverse chemicals. The compounds selected are the organic chemicals from the 50 priority chemicals prescribed by the Multicentre Evaluation of In Vitro Cytotoxicity (MEIC) programme. The models used for the evaluations are the best combination of physico-chemical properties that could be obtained so far for each organism, using the partial least squares projection to latent structures (PLS) regression method and backpropagated neural networks (BPN). Non-linear models, whether derived from PLS regression or backpropagated neural networks, appear to be better than linear models for describing the relationship between acute toxicity and molecular structure. BPN models, in turn, outperform non-linear models obtained from PLS regression. The predictive power of BPN models for the crustacean test species are better than the model for humans (based on human lethal concentration). The physico-chemical properties found to be important to predict both human acute toxicity and the toxicity to aquatic non-vertebrates are the n?octanol water partition coefficient (Pow) and heat of formation (HF). Aside from the two former properties, the contribution of parameters that reflect size and electronic properties of the molecule to the model is also high, but the type of physico-chemical properties differs from one model to another. In all of the best BPN models, some of the principal component analysis (PCA) scores of the ¹³C-NMR spectrum, with electron withdrawing/accepting capacity (LUMO, HOMO and IP) are molecular size/volume (VDW or MS1) parameters are relevant. The chemical deviating from the QSAR models include non-pesticides as well as some of the pesticides tested. The latter type of chemical fits in a number of the QSAR models. Outliers for one species may be different from those of other test organisms.