Final Ensemble Research Articles

Determination of 0–0-Transition Frequencies from Diffuse Vibronic Spectra

The equality of the probabilities of direct and inverse vibronic transitions between the initial and final elementary quantum Franck–Condon states in the adiabatic approximation is taken as the equality of the probabilities of one electronic transition in which only the initial and final elementary vibrational ensembles exchange places. The transition frequencies will have mirror symmetry with respect to the 0–0-transition frequency ν0. Accounting for this and selecting hot molecules in the anti-Stokes regions with the initial state in thermal equilibrium leads to a restoration of the known relation for the transition cross sections σ(ν) in the form [σ(ν)/ν] exp (±hν/2kT) = 𝜑[(ν – ν0)2], where ″+″ denotes emission; ″–″, absorption. Examples of using the relation to determine ν0 from absorption, fluorescence, and phosphorescence spectra are given and compared with ν0 obtained by jet-cooling.

Using a Guided Machine Learning Ensemble Model to Predict Discharge Disposition following Meningioma Resection.

Objective Machine learning (ML) algorithms are powerful tools for predicting patient outcomes. This study pilots a novel approach to algorithm selection and model creation using prediction of discharge disposition following meningioma resection as a proof of concept. Materials and Methods A diversity of ML algorithms were trained on a single-institution database of meningioma patients to predict discharge disposition. Algorithms were ranked by predictive power and top performers were combined to create an ensemble model. The final ensemble was internally validated on never-before-seen data to demonstrate generalizability. The predictive power of the ensemble was compared with a logistic regression. Further analyses were performed to identify how important variables impact the ensemble. Results Our ensemble model predicted disposition significantly better than a logistic regression (area under the curve of 0.78 and 0.71, respectively, p = 0.01). Tumor size, presentation at the emergency department, body mass index, convexity location, and preoperative motor deficit most strongly influence the model, though the independent impact of individual variables is nuanced. Conclusion Using a novel ML technique, we built a guided ML ensemble model that predicts discharge destination following meningioma resection with greater predictive power than a logistic regression, and that provides greater clinical insight than a univariate analysis. These techniques can be extended to predict many other patient outcomes of interest.

Optimizing multi-sensor deployment via ensemble pruning for wearable activity recognition

With the rapid development of sensor types and processers, most wearable activity recognition systems tend to making use of multiple homogeneous or heterogeneous sensors to obtain plethora information. However, in a realistic environment, it is difficult to configure an appropriate multi-sensor deployment to gain a tradeoff among computational complexity, accuracy and subject personality. In this paper, a multi-sensor fusion with ensemble pruning system (MSF-EP) is designed to connect with multi-sensor based wearable activity recognition system. As a result, the multi-sensor configuration problem is transformed to multiple ensemble classifier pruning problem. With respect to ensemble pruning for MSF-EP system, two popular order-based ensemble pruning approaches are utilized firstly: reduce-error pruning (RE) and complementarily pruning (Comp). Then, in light of the proposed MSF-EP system, two new ensemble pruning criteria are proposed: mRMR pruning and discriminative pruning (Disc). The mRMR pruning measure is based on mutual information and is a composite criterion of classifier redundancy and relevance with regard to the selected classifier set. Taking into account the discriminations of misclassified instances and correctly classified instances in classifier subset selected respectively, another ensemble pruning measure Disc, which combining RE pruning and Comp pruning, is presented in the form of mutual information also. Finally, the proposed pruning learner with lower error is selected as final ensemble classifier. Through the algorithm, the number and type of multi-sensor are appropriately decided to optimize the multi-sensor fusion without regrading the accuracy performance. The system conducts experimental studies on two real-world activity recognition data sets and the results show the superiority of system over other ensemble algorithms.

Margin & diversity based ordering ensemble pruning

Abstract Ensemble pruning is a technique used to improve ensemble performance and reduce the ensemble size by selecting an optimal or sub-optimal subset as the final ensemble for prediction. In this research, using example margin and ensemble diversity, we prove that the ensemble pruning method should focus more on the following two factors: (1) examples with small absolute margin and (2) classifiers that correctly classify more examples and contribute larger diversity. Based on this principle, we propose a novel metric called the margin & diversity based measure (MDM) to explicitly evaluate the importance of individual classifiers. By incorporating ensemble members in a decreasing order based on the MDM, sub-ensembles are formed such that users can select the top T ensemble members for predictions. Compared to the original ensemble and other state-of-the-art ensemble pruning methods, the proposed method shows better performance in terms of accuracy.

Coronal Mass Ejections detection using multiple features based ensemble learning

Coronal Mass Ejection (CME) is a major solar activity that affects the earth, thus CMEs detection is of great importance for space weather forecast, disaster prevention and reduction. We model the detection of Coronal Mass Ejections (CMEs) as the classification of the brightest block in the current running difference image. Because CMEs usually correspond to the areas with high gray values or complex texture features, multiple features including gray features and texture features are extracted to represent the brightest block. And classifier is designed based on these features. Our method includes four steps: first, because the CMEs spread along the radial direction of the sun, in order to facilitate the analysis, the original coordinate is transformed into the polar coordinate; Secondly, because the typical appearance of the CMEs is bright or complex texture enhancement, we use the brightest block to represent the whole image; Thirdly, we extract the gray, texture and HOG features of the brightest gray blocks. Finally, we use the extracted features to design decision trees as the base classifiers, and AdaBoost is used to obtain the final ensemble classifier. As far as we know, this is the first time that the learning based classification framework is presented in the CMEs detection. Moreover, multiple feature fusion is first used to model the various CMEs. Experimental results show that the integration of multi-feature based detection algorithm proposed can achieve better detection results.

Metallacycle-cored supramolecular assemblies with tunable fluorescence including white-light emission

Control over the fluorescence of supramolecular assemblies is crucial for the development of chemosensors and light-emitting materials. Consequently, the postsynthetic modification of supramolecular structures via host-guest interactions has emerged as an efficient strategy in recent years that allows the facile tuning of the photophysical properties without requiring a tedious chemical synthesis. Herein, we used a phenanthrene-21-crown-7 (P21C7)-based 60° diplatinum(II) acceptor 8 in the construction of three exohedral P21C7 functionalized rhomboidal metallacycles 1-3 which display orange, cyan, and green emission colors, respectively. Although these colors originate from the dipyridyl precursors 10-12, containing triphenylamine-, tetraphenylethene-, and pyrene-based fluorophores, respectively, the metal-ligand coordination strongly influences their emission properties. The metallacycles were further linked into emissive supramolecular oligomers by the addition of a fluorescent bis-ammonium linker 4 that forms complementary host-guest interactions with the pendant P21C7 units. Notably, the final ensemble derived from a 1:1 mixture of 1 and 4 displays a concentration-dependent emission. At low concentration, i.e., <25 µM, it emits a blue color, whereas an orange emission was observed when the concentration exceeds >5 mM. Moreover, white-light emission was observed from the same sample at a concentration of 29 µM, representing a pathway to construct supramolecular assemblies with tunable fluorescence properties.

Improving Classification Performance through an Advanced Ensemble Based Heterogeneous Extreme Learning Machines.

Extreme Learning Machine (ELM) is a fast-learning algorithm for a single-hidden layer feedforward neural network (SLFN). It often has good generalization performance. However, there are chances that it might overfit the training data due to having more hidden nodes than needed. To address the generalization performance, we use a heterogeneous ensemble approach. We propose an Advanced ELM Ensemble (AELME) for classification, which includes Regularized-ELM, L2-norm-optimized ELM (ELML2), and Kernel-ELM. The ensemble is constructed by training a randomly chosen ELM classifier on a subset of training data selected through random resampling. The proposed AELM-Ensemble is evolved by employing an objective function of increasing diversity and accuracy among the final ensemble. Finally, the class label of unseen data is predicted using majority vote approach. Splitting the training data into subsets and incorporation of heterogeneous ELM classifiers result in higher prediction accuracy, better generalization, and a lower number of base classifiers, as compared to other models (Adaboost, Bagging, Dynamic ELM ensemble, data splitting ELM ensemble, and ELM ensemble). The validity of AELME is confirmed through classification on several real-world benchmark datasets.

GA-Based Membrane Evolutionary Algorithm for Ensemble Clustering.

Ensemble clustering can improve the generalization ability of a single clustering algorithm and generate a more robust clustering result by integrating multiple base clusterings, so it becomes the focus of current clustering research. Ensemble clustering aims at finding a consensus partition which agrees as much as possible with base clusterings. Genetic algorithm is a highly parallel, stochastic, and adaptive search algorithm developed from the natural selection and evolutionary mechanism of biology. In this paper, an improved genetic algorithm is designed by improving the coding of chromosome. A new membrane evolutionary algorithm is constructed by using genetic mechanisms as evolution rules and combines with the communication mechanism of cell-like P system. The proposed algorithm is used to optimize the base clusterings and find the optimal chromosome as the final ensemble clustering result. The global optimization ability of the genetic algorithm and the rapid convergence of the membrane system make membrane evolutionary algorithm perform better than several state-of-the-art techniques on six real-world UCI data sets.

Subagging for the improvement of predictive stability of extreme learning machine for spectral quantitative analysis of complex samples

Extreme learning machine (ELM) has been attracted increasing attentions for its fast learning speed and excellent generalization performance. However, the prediction result of a single ELM regression model is usually unstable due to the randomly generating of the input weights and hidden layer bias. To overcome this drawback, an ensemble form of ELM, termed as subagging ELM, was proposed and used for spectral quantitative analysis of complex samples. In the approach, a series of ELM sub-models was built by randomly selecting a certain number of samples from the original training set without replacement, and then the predictions of these sub-models were combined by a simple averaging way to give the final ensemble prediction. The performance of the method was tested with fuel oil and blood samples. Compared to a single ELM model, the results confirm that subagging ELM can achieve much better stability and higher accuracy than ELM.

Geographical and altitudinal distribution of Brachycephalus (Anura: Brachycephalidae) endemic to the Brazilian Atlantic Rainforest.

Mountains of the Brazilian Atlantic Forest can act as islands of cold and wet climate, leading to the isolation and speciation of species with low dispersal capacity, such as the toadlet species of the genus Brachycephalus. This genus is composed primarily by diurnal species, with miniaturized body sizes (<2.5 cm), inhabiting microhabitats in the leaf litter of montane forests. Still, little is known about the geographical distribution, altitudinal range, and ecological limits of most Brachycephalus species. In this study, we review the available data on the geographical and altitudinal distribution of Brachycephalus based on occurrence records compiled from literature and museums, both for the genus as a whole and separately for the three recently proposed groups of species (ephippium, didactylus, and pernix). The final ensemble dataset comprised 333 records, 120 localities, 28 described species, and six undescribed ones. Species were recorded in six relief units, the richest of which being the Serra do Mar, with 30 species. When the Serra do Mar is subdivided into three subunits, Northern, Central and Southern Serra do Mar, the number of species increase from north to the south, with records of six, nine, and 16 species, respectively. We were able to estimate the extent of occurrence of nearly half of the described species, and the resulting estimates indicate that many of them show remarkably small ranges, some of which less than 50 ha. Brachycephalus species are present from sea level to roughly 1,900 m a.s.l., with the highest richness being found between 751 and 1,000 m a.s.l. (21 spp.). The species with the broadest altitudinal range were B. didactylus (1,075 m) and Brachycephalus sp. 1 (1,035 m), both in the didactylus group, and B. ephippium (1,050 m), of the ephippium group. The broadest altitudinal amplitude for species of the pernix group was recorded for B. brunneus (535 m). The lowest altitudinal records for the pernix group were at 845 m a.s.l. in the state of Paraná and at 455 m a.s.l. in the state of Santa Catarina. The altitudinal occurrence in the pernixspecies group seems to decrease southward. Syntopy between species is also reviewed.

Exploring Psycho-Physiological Correlates to Trust

This paper describes a methodology for exploring trust using psychological (subjective) and physiological (objective) correlates to trust. The aim was to explore trust using natural dialogs of real-world scenarios that embed fifteen subjective measures. The goal was to apply the method in modeling human-robot-human interaction, involving three types of androids and to predict trust. Two forms of dialogs were employed: a guided script and a predetermined dialog representing three social scenarios. Objective features included facial expressions, voice and heart rate. Subjective trust measures comprised ability, benevolence and integrity. A repeated measures experimental design was employed. Forty-two subjects participated in the study. The data was analyzed using exploratory factor analysis and correlation. Multiple neuro-fuzzy models were trained using the data set and combined as an ensemble using evolutionary algorithms. The final ensemble estimated trust with 67% accuracy. The implications of the findings and limitations of the method are discussed.

Self-assembled growth of GaN nanowires on amorphous AlxOy: from nucleation to the formation of dense nanowire ensembles

We present a comprehensive description of the self-assembled nucleation and growth of GaN nanowires (NWs) by plasma-assisted molecular beam epitaxy on amorphous AlxOy buffers (a-AlxOy) prepared by atomic layer deposition. The results are compared with those obtained on nitridated Si(111). Using line-of-sight quadrupole mass spectrometry, we analyze in situ the incorporation of Ga starting from the incubation and nucleation stages till the formation of the final nanowire ensemble and observe qualitatively the same time dependence for the two types of substrates. However, on a-AlxOy the incubation time is shorter and the nucleation faster than on nitridated Si. Moreover, on a-AlxOy we observe a novel effect of decrease in incorporated Ga flux for long growth durations which we explain by coalescence of NWs leading to reduction of the GaN surface area where Ga may reside. Dedicated samples are used to analyze the evolution of surface morphology. In particular, no GaN nuclei are detected when growth is interrupted during the incubation stage. Moreover, for a-AlxOy, the same shape transition from spherical cap-shaped GaN crystallites to the NW-like geometry is found as it is known for nitridated Si. However, while the critical radius for this transition is only slightly larger for a-AlxOy than for nitridated Si, the critical height is more than six times larger for a-AlxOy. Finally, we observe that in fully developed NW ensembles, the substrate no longer influences growth kinetics and the same N-limited axial growth rate is measured on both substrates. We conclude that the same nucleation and growth processes take place on a-AlxOy as on nitridated Si and that these processes are of a general nature. Quantitatively, nucleation proceeds somewhat differently, which indicates the influence of the substrate, but once shadowing limits growth processes to the upper part of the NW ensemble, they are not affected anymore by the type of substrate.

Metamodel selection based on stepwise regression

Metamodels are often used to replace expensive simulations of engineering problems. When a training set is given, a series of metamodels can be constructed, and then there are two strategies to deal with these metamodels: (1) picking out the best one with the highest accuracy as an approximation of the computationally intensive simulation; and (2) combining all of them into an ensemble model. However, since the choice of approximate model depends on design of experiments (DOEs), employing of the first strategy thus increases the risk of adopting an inappropriate model. Nevertheless, the second strategy also seems not to be a good choice, since adding redundant metamodels may lead to loss of accuracy. Therefore, it is a necessary step to eliminate the redundant metamodels from the set of the candidates before constructing the final ensemble. Illuminated by the method of variable selection widely used in polynomial regression, a metamodel selection method based on stepwise regression is proposed. In our method, just a subset of n ones (n ≤ p, where p is the number of all of the candidate metamodels) is used. In addition, a new ensemble technique is proposed from the view of polynomial regression in this work. This new ensemble technique, combined with metamodel selection method, has been evaluated using six benchmark problems. The results show that eliminating the redundant metamodels before constructing the ensemble can provide more ideal prediction accuracy than directly constructing the ensemble by utilizing all of the candidates.

Protein Complex Detection via Effective Integration of Base Clustering Solutions and Co-Complex Affinity Scores.

With the increasing availability of protein interaction data, various computational methods have been developed to predict protein complexes. However, different computational methods may have their own advantages and limitations. Ensemble clustering has thus been studied to minimize the potential bias and risk of individual methods and generate prediction results with better coverage and accuracy. In this paper, we extend the traditional ensemble clustering by taking into account the co-complex affinity scores and present an Ensemble H ierarchical Clustering framework (EnsemHC) to detect protein complexes. First, we construct co-cluster matrices by integrating the clustering results with the co-complex evidences. Second, we sum up the constructed co-cluster matrices to derive a final ensemble matrix via a novel iterative weighting scheme. Finally, we apply the hierarchical clustering to generate protein complexes from the final ensemble matrix. Experimental results demonstrate that our EnsemHC performs better than its base clustering methods and various existing integrative methods. In addition, we also observed that integrating the clusters and co-complex affinity scores from different data sources will improve the prediction performance, e.g., integrating the clusters from TAP data and co-complex affinities from binary PPI data achieved the best performance in our experiments.

A subspace ensemble framework for classification with high dimensional missing data

Real world classification tasks may involve high dimensional missing data. The traditional approach to handling the missing data is to impute the data first, and then apply the traditional classification algorithms on the imputed data. This method first assumes that there exist a distribution or feature relations among the data, and then estimates missing items with existing observed values. A reasonable assumption is a necessary guarantee for accurate imputation. The distribution or feature relations of data, however, is often complex or even impossible to be captured in high dimensional data sets, leading to inaccurate imputation. In this paper, we propose a complete-case projection subspace ensemble framework, where two alternative partition strategies, namely bootstrap subspace partition and missing pattern-sensitive subspace partition, are developed for incomplete datasets with even missing patterns and uneven missing patterns, respectively. Multiple component classifiers are then separately trained in these subspaces. After that, a final ensemble classifier is constructed by a weighted majority vote of component classifiers. In the experiments, we demonstrate the effectiveness of the proposed framework over eight high dimensional UCI datasets. Meanwhile, we apply the two proposed partition strategies over data sets with different missing patterns. As indicated, the proposed algorithm significantly outperforms existing imputation methods in most cases.

Random subspace method with class separability weighting

AbstractThe random subspace method (RSM) is one of the ensemble learning algorithms widely used in pattern classification applications. RSM has the advantages of small error rate and improved noise insensitivity due to ensemble construction of the base‐learners. However, randomness may cause a reduction of the final ensemble decision performance because of contributions of classifiers trained by subsets with low class separability. In this study, we present a new and improved version of the RSM by introducing a weighting factor into the combination phase. One of the class separability criteria, J3, is used as a weighting factor to improve the classification performance and eliminate the drawbacks of the standard RSM algorithm. The randomly selected subsets are quantified by computing their J3 measure to determine voting weights in the model combination phase, assigning lower voting weight to classifiers trained by subsets with poor class separability. Two models are presented including J3‐weighted RSM and optimized J3 weighted RSM. In J3 weighted RSM, computed weighting values are directly multiplied by class assignment posteriors, whereas in optimized J3 weighted RSM, computed weighting values are optimized by a pattern search algorithm before multiplying by posteriors. Both models are shown to provide better error rates at lower subset dimensionality.

Projections of extreme storm surge levels along Europe

Storm surges are an important coastal hazard component and it is unknown how they will evolve along Europe’s coastline in view of climate change. In the present contribution, the hydrodynamic model Delft3D-Flow was forced by surface wind and atmospheric pressure fields from a 8-member climate model ensemble in order to evaluate dynamics in storm surge levels (SSL) along the European coastline (1) for the baseline period 1970–2000; and (2) during this century under the Representative Concentration Pathways RCP4.5 and RCP8.5. Validation simulations, spanning from 2008 to 2014 and driven by ERA-Interim atmospheric forcing, indicated good predictive skill (0.06 m < RMSE < 0.29 m and 10 % < RMSE < 29 % for 110 tidal gauge stations across Europe). Peak-over-threshold extreme value analysis was applied to estimate SSL values for different return periods, and changes of future SSL were obtained from all models to obtain the final ensemble. Values for most scenarios and return periods indicate a projected increase in SSL at several locations along the North European coastline, which is more prominent for RCP8.5 and shows an increasing tendency towards the end of the century for both RCP4.5 and RCP8.5. Projected SSL changes along the European coastal areas south of 50°N show minimal change or even a small decrease, with the exception of RCP8.5 under which a moderate increase is projected towards the end of the century. The present findings indicate that the anticipated increase in extreme total water levels due to relative sea level rise (RSLR), can be further enforced by an increase of the extreme SSL, which can exceed 30 % of the RSLR, especially for the high return periods and pathway RCP8.5. This implies that the combined effect could increase even further anticipated impacts of climate change for certain European areas and highlights the necessity for timely coastal adaptation and protection measures. The dataset is publicly available under this link: http://data.jrc.ec.europa.eu/collection/LISCOAST .

Texture Descriptors Ensembles Enable Image-Based Classification of Maturation of Human Stem Cell-Derived Retinal Pigmented Epithelium.

AimsA fast, non-invasive and observer-independent method to analyze the homogeneity and maturity of human pluripotent stem cell (hPSC) derived retinal pigment epithelial (RPE) cells is warranted to assess the suitability of hPSC-RPE cells for implantation or in vitro use. The aim of this work was to develop and validate methods to create ensembles of state-of-the-art texture descriptors and to provide a robust classification tool to separate three different maturation stages of RPE cells by using phase contrast microscopy images. The same methods were also validated on a wide variety of biological image classification problems, such as histological or virus image classification.MethodsFor image classification we used different texture descriptors, descriptor ensembles and preprocessing techniques. Also, three new methods were tested. The first approach was an ensemble of preprocessing methods, to create an additional set of images. The second was the region-based approach, where saliency detection and wavelet decomposition divide each image in two different regions, from which features were extracted through different descriptors. The third method was an ensemble of Binarized Statistical Image Features, based on different sizes and thresholds. A Support Vector Machine (SVM) was trained for each descriptor histogram and the set of SVMs combined by sum rule. The accuracy of the computer vision tool was verified in classifying the hPSC-RPE cell maturation level.Dataset and ResultsThe RPE dataset contains 1862 subwindows from 195 phase contrast images. The final descriptor ensemble outperformed the most recent stand-alone texture descriptors, obtaining, for the RPE dataset, an area under ROC curve (AUC) of 86.49% with the 10-fold cross validation and 91.98% with the leave-one-image-out protocol. The generality of the three proposed approaches was ascertained with 10 more biological image datasets, obtaining an average AUC greater than 97%.ConclusionsHere we showed that the developed ensembles of texture descriptors are able to classify the RPE cell maturation stage. Moreover, we proved that preprocessing and region-based decomposition improves many descriptors’ accuracy in biological dataset classification. Finally, we built the first public dataset of stem cell-derived RPE cells, which is publicly available to the scientific community for classification studies. The proposed tool is available at https://www.dei.unipd.it/node/2357 and the RPE dataset at http://www.biomeditech.fi/data/RPE_dataset/. Both are available at https://figshare.com/s/d6fb591f1beb4f8efa6f.

Solution structure of the calmodulin-like C-terminal domain of Entamoeba α-actinin2.

Cell motility is dependent on a dynamic meshwork of actin filaments that is remodelled continuously. A large number of associated proteins that are severs, cross-links, or caps the filament ends have been identified and the actin cross-linker α-actinin has been implied in several important cellular processes. In Entamoeba histolytica, the etiological agent of human amoebiasis, α-actinin is believed to be required for infection. To better understand the role of α-actinin in the infectious process we have determined the solution structure of the C-terminal calmodulin-like domain using NMR. The final structure ensemble of the apo form shows two lobes, that both resemble other pairs of calcium-binding EF-hand motifs, connected with a mobile linker.

Structural determinants of ligand binding in the ternary complex of human ileal bile acid binding protein with glycocholate and glycochenodeoxycholate obtained from solution NMR.

The coordinates of the 10 lowest energy structures of the human I-BABP : GCDA : GCA complex as well as the distance restraints used to calculate the final ensemble have been deposited in the Brookhaven Protein Data Bank with accession number 2MM3.