Comparing Boosting and Bagging for Decision Trees of Rankings

Decision tree learning is among the most popular and most traditional families of machine learning algorithms. While these techniques excel in being quite intuitive and interpretable, they also suffer from instability: small perturbations in the training data may result in big changes in the predictions. The so-called ensemble methods combine the output of multiple trees, which makes the decision more reliable and stable. They have been primarily applied to numeric prediction problems and to classification tasks. In the last years, some attempts to extend the ensemble methods to ordinal data can be found in the literature, but no concrete methodology has been provided for preference data. In this paper, we extend decision trees, and in the following also ensemble methods to ranking data. In particular, we propose a theoretical and computational definition of bagging and boosting, two of the best known ensemble methods. In an experimental study using simulated data and real-world datasets, our results confirm that known results from classification, such as that boosting outperforms bagging, could be successfully carried over to the ranking case.

Identification of enhancer regulatory elements that direct epicardial gene expression during zebrafish heart regeneration

The epicardium is a mesothelial tissue layer that envelops the heart. Cardiac injury activates dynamic gene expression programs in epicardial tissue, which in the case of zebrafish enables subsequent regeneration through paracrine and vascularizing effects. To identify tissue regeneration enhancer elements (TREEs) that control injury-induced epicardial gene expression during heart regeneration, we profiled transcriptomes and chromatin accessibility in epicardial cells purified from regenerating zebrafish hearts. We identified hundreds of candidate TREEs, defined by increased chromatin accessibility of non-coding elements near genes with increased expression during regeneration. Several of these candidate TREEs were incorporated into stable transgenic lines, with 5 of 6 elements directing injury-induced epicardial expression but not ontogenetic epicardial expression in hearts of larval animals. Whereas two independent TREEs linked to the gene gnai3 showed similar functional features of gene regulation in transgenic lines, two independent ncam1a-linked TREEs directed distinct spatiotemporal domains of epicardial gene expression. Thus, multiple TREEs linked to a regeneration gene can possess either matching or complementary regulatory controls. Our study provides a new resource and principles for understanding the regulation of epicardial genetic programs during heart regeneration.

Permian–Triassic phylogenetic and morphologic evolution of rhynchonellide brachiopods

The Rhynchonellida is a major group of brachiopods that survived the “big five” mass extinctions and flourished after the Permian/Triassic (P/Tr) crisis. However, phylogenetic and character evolution in the Rhynchonellida across the P/Tr transition is poorly understood. In view of the widespread homoplasy across this order, we employ a tip-dated Bayesian analysis to reconstruct phylogenetic relationships for late Permian–Triassic rhynchonellides. The same data were also analyzed using three other methods: undated Bayesian, equal-weighting, and implied-weighting parsimony. Compared with trees generated by other methods, those constructed by tip-dating best account for the homoplasy in this group and are closer to previous assumptions on the evolution of this order. Based on the analyses of multiple trees, the major increase in lineage richness occurred in the Early and early Middle Triassic. Also, richness in the Anisian almost reached the highest level seen in the Triassic. According to fossil records, a pronounced reduction in shell size and in the development of ornamentation occurred after the P/Tr extinction, which is largely due to the loss of large and highly sculptured genera and the diversification of small-sized and weakly ornamented genera. Ancestral-state estimation of shell size and development of ornamentation, coupled with comparisons of other characters, indicate that the Early–Middle Triassic mature “small-sized taxa” may have characters displayed by juveniles of their ancestors. This suggests that for these genera, paedomorphosis was possibly a strategy to survive and diversify in the harsh environment after the P/Tr extinction.

Role of vegetation placement for temperature moderation in an urban microclimate

Through optimal planning and site design, strategic selection and placement of vegetation is one approach to prevent warming in the urban core. To test this hypothesis, a paired sampling design using temperature loggers, was conducted in the City of Toronto to assess the overall effect that shading through vegetation had on moderating temperatures in the microclimate proximate to built structures. The role of vines, a single mature tree, and multiple trees growing at one site, was investigated to compare their temperature moderating benefits. Tree placement on the west facing aspect of built structures delivered the greatest overall benefits when compared to south and east facing building walls. Temperature differences between loggers reached a maximum of 11.7°C during the month of August. A mixed model evaluated the longitudinal study data and revealed that temperatures were significantly cooler (p < 0.05) in the shade of both trees and vines compared to those recorded in the sun for all aspects throughout peak solar access periods.

Role of vegetation placement for temperature moderation in an urban microclimate

Through optimal planning and site design, strategic selection and placement of vegetation is one approach to prevent warming in the urban core. To test this hypothesis, a paired sampling design using temperature loggers, was conducted in the City of Toronto to assess the overall effect that shading through vegetation had on moderating temperatures in the microclimate proximate to built structures. The role of vines, a single mature tree, and multiple trees growing at one site, was investigated to compare their temperature moderating benefits. Tree placement on the west facing aspect of built structures delivered the greatest overall benefits when compared to south and east facing building walls. Temperature differences between loggers reached a maximum of 11.7°C during the month of August. A mixed model evaluated the longitudinal study data and revealed that temperatures were significantly cooler (p < 0.05) in the shade of both trees and vines compared to those recorded in the sun for all aspects throughout peak solar access periods.

Novel metric for hyperbolic phylogenetic tree embeddings

Advances in experimental technologies, such as DNA sequencing, have opened up new avenues for the applications of phylogenetic methods to various fields beyond their traditional application in evolutionary investigations, extending to the fields of development, differentiation, cancer genomics, and immunogenomics. Thus, the importance of phylogenetic methods is increasingly being recognized, and the development of a novel phylogenetic approach can contribute to several areas of research. Recently, the use of hyperbolic geometry has attracted attention in artificial intelligence research. Hyperbolic space can better represent a hierarchical structure compared to Euclidean space, and can therefore be useful for describing and analyzing a phylogenetic tree. In this study, we developed a novel metric that considers the characteristics of a phylogenetic tree for representation in hyperbolic space. We compared the performance of the proposed hyperbolic embeddings, general hyperbolic embeddings, and Euclidean embeddings, and confirmed that our method could be used to more precisely reconstruct evolutionary distance. We also demonstrate that our approach is useful for predicting the nearest-neighbor node in a partial phylogenetic tree with missing nodes. Furthermore, we proposed a novel approach based on our metric to integrate multiple trees for analyzing tree nodes or imputing missing distances. This study highlights the utility of adopting a geometric approach for further advancing the applications of phylogenetic methods.

Preliminary Investigation of the Integrated Evaluation Model based on Random Forest Method

With the deepening of reforms, as the closing year of the “Modern Vocational Education System Construction Plan (2014-2020)”, it has witnessed the development of vocational education. Its development trend is “only specialization” and “light basic courses”. Currently, the random forest method is very popular. This method forms a “forest” by building many decision trees, and votes on multiple trees to make decisions. This method can improve the classification accuracy of new samples effectively and it is more suitable for the research proposed. Random forest evaluation model has three functions: formative, diagnostic and summative. This paper establishes a comprehensive evaluation model for basic courses to make the evaluation more standardized, comprehensive and operability. And this evaluation system can supervise all stages of teaching.

Multi-tree Genetic Programming with A New Fitness Function for Melanoma Detection

© 2019 IEEE. The occurrence of malignant melanoma had enormously increased since past decades. For accurate detection and classification, not only discriminative features are required but a properly designed model to combine these features effectively is also needed. In this study, the multi-tree representation of genetic programming (GP) has been utilised to effectively combine different types of features and evolve a classification model for the task of melanoma detection. Local binary patterns have been used to extract pixel-level informative features. For incorporating the properties of ABCD (asymmetrical property, border shape, color variation and geometrical characteristics) rule of dermoscopy, various features have been used to include local and global information of the skin lesions. To meet the requirements of the proposed multi-tree GP representation, genetic operators such as crossover and mutation are designed accordingly. Moreover, a new weighted fitness function is designed to evolve better GP individuals having multiple trees influencing each other's performance during the evolution, in order to get overall performance gains. The performance of the new method is checked on two benchmark skin image datasets, and compared with six widely used classification algorithms and the single tree GP method. The experimental results have shown that the proposed method has significantly outperformed all these classification methods.

Finnish National Phenological Network 1997–2017: from observations to trend detection

Plant phenological dataset collected at 42 sites across the mainland of Finland and covering the years 1997–2017 is presented and analysed for temporal trends. The dataset of n = 16,257 observations represents eleven plant species and fifteen phenological stages and results in forty different variables, i.e. phenophases. Trend analysis was carried out for n = 808 phenological time-series that contained at least 10 observations over the 21-year study period. A clear signal of advancing spring and early-summer phenology was detected, 3.4 days decade−1, demonstrated by a high proportion of negative trends for phenophases occurring in April through June. Latitudinal correlation indicated stronger signal of spring and early-summer phenology towards the northern part of the study region. The autumn signal was less consistent and showed larger within-site variations than those observed in other seasons. More than 60% of the dates based on single tree/monitoring square were exactly the same as the averages from multiple trees/monitoring squares within the site. In particular, the reliability of data on autumn phenology was increased by multiple observations per site. The network is no longer active.