Search

Multilabel classification (MLC) is a machine learning task where the goal is to learn to label an example with multiple labels simultaneously. It receives increasing interest from the machine learning community, as evidenced by the increasing number of papers and methods that appear in the literature. Hence, ensuring proper, correct, robust, and trustworthy benchmarking is of utmost importance for the further development of the field. We believe that this can be achieved by adhering to the recently emerged data management standards, such as the FAIR (Findable, Accessible, Interoperable, and Reusable) and TRUST (Transparency, Responsibility, User focus, Sustainability, and Technology) principles. We introduce an ontology-based online catalogue of MLC datasets originating from various application domains following these principles. The catalogue extensively describes many MLC datasets with comprehensible meta-features, MLC-specific semantic descriptions, and different data provenance information. The MLC data catalogue is available at: http://semantichub.ijs.si/MLCdatasets ., (© 2022. The Author(s).)

Therapies halting the progression of fibrosis are ineffective and limited. Activated myofibroblasts are emerging as important targets in the progression of fibrotic diseases. Previously, we performed a high-throughput screen on lung fibroblasts and subsequently demonstrated that the inhibition of myofibroblast activation is able to prevent lung fibrosis in bleomycin-treated mice. High-throughput screens are an ideal method of repurposing drugs, yet they contain an intrinsic limitation, which is the size of the library itself. Here, we exploited the data from our "wet" screen and used "dry" machine learning analysis to virtually screen millions of compounds, identifying novel anti-fibrotic hits which target myofibroblast differentiation, many of which were structurally related to dopamine. We synthesized and validated several compounds ex vivo ("wet") and confirmed that both dopamine and its derivative TS1 are powerful inhibitors of myofibroblast activation. We further used RNAi-mediated knock-down and demonstrated that both molecules act through the dopamine receptor 3 and exert their anti-fibrotic effect by inhibiting the canonical transforming growth factor β pathway. Furthermore, molecular modelling confirmed the capability of TS1 to bind both human and mouse dopamine receptor 3. The anti-fibrotic effect on human cells was confirmed using primary fibroblasts from idiopathic pulmonary fibrosis patients. Finally, TS1 prevented and reversed disease progression in a murine model of lung fibrosis. Both our interdisciplinary approach and our novel compound TS1 are promising tools for understanding and combating lung fibrosis., (© 2021. The Author(s).)

The reconstruction of Gene Regulatory Networks (GRNs) from gene expression data, supported by machine learning approaches, has received increasing attention in recent years. The task at hand is to identify regulatory links between genes in a network. However, existing methods often suffer when the number of labeled examples is low or when no negative examples are available. In this paper we propose a multi-task method that is able to simultaneously reconstruct the human and the mouse GRNs using the similarities between the two. This is done by exploiting, in a transfer learning approach, possible dependencies that may exist among them. Simultaneously, we solve the issues arising from the limited availability of examples of links by relying on a novel clustering-based approach, able to estimate the degree of certainty of unlabeled examples of links, so that they can be exploited during the training together with the labeled examples. Our experiments show that the proposed method can reconstruct both the human and the mouse GRNs more effectively compared to reconstructing each network separately. Moreover, it significantly outperforms three state-of-the-art transfer learning approaches that, analogously to our method, can exploit the knowledge coming from both organisms. Finally, a specific robustness analysis reveals that, even when the number of labeled examples is very low with respect to the number of unlabeled examples, the proposed method is almost always able to outperform its single-task counterpart.

Nutrient management on grazed grasslands is of critical importance to maintain productivity levels, as grass is the cheapest feed for ruminants and underpins these meat and milk production systems. Many attempts have been made to model the relationships between controllable (crop and soil fertility management) and noncontrollable influencing factors (weather, soil drainage) and nutrient/productivity levels. However, to the best of our knowledge not much research has been performed on modeling the interconnections between the influencing factors on one hand and nutrient uptake/herbage production on the other hand, by using data-driven modeling techniques. Our paper proposes to use predictive clustering trees (PCT) learned for building models on data from dairy farms in the Republic of Ireland. The PCT models show good accuracy in estimating herbage production and nutrient uptake. They are also interpretable and are found to embody knowledge that is in accordance with existing theoretical understanding of the task at hand. Moreover, if we combine more PCT into an ensemble of PCT (random forest of PCT), we can achieve improved accuracy of the estimates. In practical terms, the number of grazings, which is related proportionally with soil drainage class, is one of the most important factors that moderates the herbage production potential and nutrient uptake. Furthermore, we found the nutrient (N, P, and K) uptake and herbage nutrient concentration to be conservative in fields that had medium yield potential (11 t of dry matter per hectare on average), whereas nutrient uptake was more variable and potentially limiting in fields that had higher and lower herbage production. Our models also show that phosphorus is the most limiting nutrient for herbage production across the fields on these Irish dairy farms, followed by nitrogen and potassium., (The Authors. Published by FASS Inc. and Elsevier Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).)

Ensembles are a well established machine learning paradigm, leading to accurate and robust models, predominantly applied to predictive modeling tasks. Ensemble models comprise a finite set of diverse predictive models whose combined output is expected to yield an improved predictive performance as compared to an individual model. In this paper, we propose a new method for learning ensembles of process-based models of dynamic systems. The process-based modeling paradigm employs domain-specific knowledge to automatically learn models of dynamic systems from time-series observational data. Previous work has shown that ensembles based on sampling observational data (i.e., bagging and boosting), significantly improve predictive performance of process-based models. However, this improvement comes at the cost of a substantial increase of the computational time needed for learning. To address this problem, the paper proposes a method that aims at efficiently learning ensembles of process-based models, while maintaining their accurate long-term predictive performance. This is achieved by constructing ensembles with sampling domain-specific knowledge instead of sampling data. We apply the proposed method to and evaluate its performance on a set of problems of automated predictive modeling in three lake ecosystems using a library of process-based knowledge for modeling population dynamics. The experimental results identify the optimal design decisions regarding the learning algorithm. The results also show that the proposed ensembles yield significantly more accurate predictions of population dynamics as compared to individual process-based models. Finally, while their predictive performance is comparable to the one of ensembles obtained with the state-of-the-art methods of bagging and boosting, they are substantially more efficient.

The task of gene regulatory network reconstruction from high-throughput data is receiving increasing attention in recent years. As a consequence, many inference methods for solving this task have been proposed in the literature. It has been recently observed, however, that no single inference method performs optimally across all datasets. It has also been shown that the integration of predictions from multiple inference methods is more robust and shows high performance across diverse datasets. Inspired by this research, in this paper, we propose a machine learning solution which learns to combine predictions from multiple inference methods. While this approach adds additional complexity to the inference process, we expect it would also carry substantial benefits. These would come from the automatic adaptation to patterns on the outputs of individual inference methods, so that it is possible to identify regulatory interactions more reliably when these patterns occur. This article demonstrates the benefits (in terms of accuracy of the reconstructed networks) of the proposed method, which exploits an iterative, semi-supervised ensemble-based algorithm. The algorithm learns to combine the interactions predicted by many different inference methods in the multi-view learning setting. The empirical evaluation of the proposed algorithm on a prokaryotic model organism (E. coli) and on a eukaryotic model organism (S. cerevisiae) clearly shows improved performance over the state of the art methods. The results indicate that gene regulatory network reconstruction for the real datasets is more difficult for S. cerevisiae than for E. coli. The software, all the datasets used in the experiments and all the results are available for download at the following link: http://figshare.com/articles/Semi_supervised_Multi_View_Learning_for_Gene_Network_Reconstruction/1604827.

The regulation of protein function by external or internal signals is one of the key features of living organisms. The ability to directly control the function of a selected protein would represent a valuable tool for regulating biological processes. Here, we present a generally applicable regulation of proteins called INSRTR, based on inserting a peptide into a loop of a target protein that retains its function. We demonstrate the versatility and robustness of coiled-coil-mediated regulation, which enables designs for either inactivation or activation of selected protein functions, and implementation of two-input logic functions with rapid response in mammalian cells. The selection of insertion positions in tested proteins was facilitated by using a predictive machine learning model. We showcase the robustness of the INSRTR strategy on proteins with diverse folds and biological functions, including enzymes, signaling mediators, DNA binders, transcriptional regulators, reporters, and antibody domains implemented as chimeric antigen receptors in T cells. Our findings highlight the potential of INSRTR as a powerful tool for precise control of protein function, advancing our understanding of biological processes and developing biotechnological and therapeutic interventions. [ABSTRACT FROM AUTHOR]

Participatory exposure research, which tracks behaviour and assesses exposure to stressors like air pollution, traditionally relies on time-activity diaries. This study introduces a novel approach, employing machine learning (ML) to empower laypersons in human activity recognition (HAR), aiming to reduce dependence on manual recording by leveraging data from wearable sensors. Recognising complex activities such as smoking and cooking presents unique challenges due to specific environmental conditions. In this research, we combined wearable environment/ambient and wrist-worn activity/biometric sensors for complex activity recognition in an urban stressor exposure study, measuring parameters like particulate matter concentrations, temperature, and humidity. Two groups, Group H (88 individuals) and Group M (18 individuals), wore the devices and manually logged their activities hourly and minutely, respectively. Prioritising accessibility and inclusivity, we selected three classification algorithms: k-nearest neighbours (IBk), decision trees (J48), and random forests (RF), based on: (1) proven efficacy in existing literature, (2) understandability and transparency for laypersons, (3) availability on user-friendly platforms like WEKA, and (4) efficiency on basic devices such as office laptops or smartphones. Accuracy improved with finer temporal resolution and detailed activity categories. However, when compared to other published human activity recognition research, our accuracy rates, particularly for less complex activities, were not as competitive. Misclassifications were higher for vague activities (resting, playing), while well-defined activities (smoking, cooking, running) had few errors. Including environmental sensor data increased accuracy for all activities, especially playing, smoking, and running. Future work should consider exploring other explainable algorithms available on diverse tools and platforms. Our findings underscore ML's potential in exposure studies, emphasising its adaptability and significance for laypersons while also highlighting areas for improvement. [ABSTRACT FROM AUTHOR]

Historically valuable canvas paintings are often exposed to conditions enabling microbial deterioration. Painting materials, mainly of organic origin, in combination with high humidity and other environmental conditions, favor microbial metabolism and growth. These preconditions are often present during exhibitions or storage in old buildings, such as churches and castles, and also in museum storage depositories. The accumulated dust serves as an inoculum for both indoor and outdoor fungi. In our study, we present the results on cultivable fungi isolated from 24 canvas paintings, mainly exhibited in Slovenian sacral buildings, dating from the 16th to 21st centuries. Fungi were isolated from the front and back of damaged and undamaged surfaces of the paintings using culture media with high- and low-water activity. A total of 465 isolates were identified using current taxonomic DNA markers and assigned to 37 genera and 98 species. The most abundant genus was Aspergillus, represented by 32 species, of which 9 xerophilic species are for the first time mentioned in contaminated paintings. In addition to the most abundant xerophilic A. vitricola, A. destruens, A. tardicrescens, and A. magnivesiculatus, xerophilic Wallemia muriae and W. canadensis, xerotolerant Penicillium chrysogenum, P. brevicompactum, P. corylophilum, and xerotolerant Cladosporium species were most frequent. When machine learning methods were used to predict the relationship between fungal contamination, damage to the painting, and the type of material present, proteins were identified as one of the most important factors and cracked paint was identified as a hotspot for fungal growth. Aspergillus species colonize paintings regardless of materials, while Wallemia spp. can be associated with animal fat. Culture media with low-water activity are suggested in such inventories to isolate and obtain an overview of fungi that are actively contaminating paintings stored indoors at low relative humidity. [ABSTRACT FROM AUTHOR]

We propose an approach to symbolic regression based on a novel variational autoencoder for generating hierarchical structures, HVAE. It combines simple atomic units with shared weights to recursively encode and decode the individual nodes in the hierarchy. Encoding is performed bottom-up and decoding top-down. We empirically show that HVAE can be trained efficiently with small corpora of mathematical expressions and can accurately encode expressions into a smooth low-dimensional latent space. The latter can be efficiently explored with various optimization methods to address the task of symbolic regression. Indeed, random search through the latent space of HVAE performs better than random search through expressions generated by manually crafted probabilistic grammars for mathematical expressions. Finally, EDHiE system for symbolic regression, which applies an evolutionary algorithm to the latent space of HVAE, reconstructs equations from a standard symbolic regression benchmark better than a state-of-the-art system based on a similar combination of deep learning and evolutionary algorithms. [ABSTRACT FROM AUTHOR]

The data used for analysis are becoming increasingly complex along several directions: high dimensionality, number of examples and availability of labels for the examples. This poses a variety of challenges for the existing machine learning methods, related to analyzing datasets with a large number of examples that are described in a high-dimensional space, where not all examples have labels provided. For example, when investigating the toxicity of chemical compounds, there are many compounds available that can be described with information-rich high-dimensional representations, but not all of the compounds have information on their toxicity. To address these challenges, we propose methods for semi-supervised learning (SSL) of feature rankings. The feature rankings are learned in the context of classification and regression, as well as in the context of structured output prediction (multi-label classification, MLC, hierarchical multi-label classification, HMLC and multi-target regression, MTR) tasks. This is the first work that treats the task of feature ranking uniformly across various tasks of semi-supervised structured output prediction. To the best of our knowledge, it is also the first work on SSL of feature rankings for the tasks of HMLC and MTR. More specifically, we propose two approaches—based on predictive clustering tree ensembles and the Relief family of algorithms—and evaluate their performance across 38 benchmark datasets. The extensive evaluation reveals that rankings based on Random Forest ensembles perform the best for classification tasks (incl. MLC and HMLC tasks) and are the fastest for all tasks, while ensembles based on extremely randomized trees work best for the regression tasks. Semi-supervised feature rankings outperform their supervised counterparts across the majority of datasets for all of the different tasks, showing the benefit of using unlabeled in addition to labeled data. [ABSTRACT FROM AUTHOR]

The utilization of a non-invasive electroencephalogram (EEG) as an input sensor is a common approach in the field of the brain–computer interfaces (BCI). However, the collected EEG data pose many challenges, one of which may be the age-related variability of event-related potentials (ERPs), which are often used as primary EEG BCI signal features. To assess the potential effects of aging, a sample of 27 young and 43 older healthy individuals participated in a visual oddball study, in which they passively viewed frequent stimuli among randomly occurring rare stimuli while being recorded with a 32-channel EEG set. Two types of EEG datasets were created to train the classifiers, one consisting of amplitude and spectral features in time and another with extracted time-independent statistical ERP features. Among the nine classifiers tested, linear classifiers performed best. Furthermore, we show that classification performance differs between dataset types. When temporal features were used, maximum individuals' performance scores were higher, had lower variance, and were less affected overall by within-class differences such as age. Finally, we found that the effect of aging on classification performance depends on the classifier and its internal feature ranking. Accordingly, performance will differ if the model favors features with large within-class differences. With this in mind, care must be taken in feature extraction and selection to find the correct features and consequently avoid potential age-related performance degradation in practice. [ABSTRACT FROM AUTHOR]

Data in the form of images are now generated at an unprecedented rate. A case in point is remote sensing images (RSI), now available in large-scale RSI archives, which have attracted a considerable amount of research on image classification within the remote sensing community. The basic task of single-target multi-class image classification considers the case where each image is assigned exactly one label from a predefined finite set of class labels. Recently, however, image annotations have become increasingly complex, with images labeled with several labels (instead of just one). In other words, the goal is to assign multiple semantic categories to an image, based on its high-level context. The corresponding machine learning tasks is called multi-label classification (MLC). The classification of RSI is currently predominantly addressed by deep neural network (DNN) approaches, especially convolutional neural networks (CNNs), which can be utilized as feature extractors as well as end-to-end methods. After only considering single-target classification for a long period, DNNs have recently emerged that address the task of MLC. On the other hand, trees and tree ensembles for MLC have a long tradition and are the best-performing class of MLC methods, but need predefined feature representations to operate on. In this work, we explore different strategies for model training based on the transfer learning paradigm, where we utilize different families of (pre-trained) CNN architectures, such as VGG, EfficientNet, and ResNet. The architectures are trained in an end-to-end manner and used in two different modes of operation, namely, as standalone models that directly perform the MLC task, and as feature extractors. In the latter case, the learned representations are used with tree ensemble methods for MLC, such as random forests and extremely randomized trees. We conduct an extensive experimental analysis of methods over several publicly available RSI datasets and evaluate their effectiveness in terms of standard MLC measures. Of these, ranking-based evaluation measures are most relevant, especially ranking loss. The results show that, for addressing the RSI-MLC task, it is favorable to use lightweight network architectures, such as EfficientNet-B2, which is the best performing end-to-end approach, as well as a feature extractor. Furthermore, in the datasets with a limited number of images, using traditional tree ensembles for MLC can yield better performance compared to end-to-end deep approaches. [ABSTRACT FROM AUTHOR]

Meta learning generalizes the empirical experience with different learning tasks and holds promise for providing important empirical insight into the behavior of machine learning algorithms. In this paper, we present a comprehensive meta-learning study of data sets and methods for multilabel classification (MLC). MLC is a practically relevant machine learning task where each example is labeled with multiple labels simultaneously. Here, we analyze 40 MLC data sets by using 50 meta features describing different properties of the data. The main findings of this study are as follows. First, the most prominent meta features that describe the space of MLC data sets are the ones assessing different aspects of the label space. Second, the meta models show that the most important meta features describe the label space, and, the meta features describing the relationships among the labels tend to occur a bit more often than the meta features describing the distributions between and within the individual labels. Third, the optimization of the hyperparameters can improve the predictive performance, however, quite often the extent of the improvements does not always justify the resource utilization. [ABSTRACT FROM AUTHOR]

We present six datasets containing telemetry data of the Mars Express Spacecraft (MEX), a spacecraft orbiting Mars operated by the European Space Agency. The data consisting of context data and thermal power consumption measurements, capture the status of the spacecraft over three Martian years, sampled at six different time resolutions that range from 1 min to 60 min. From a data analysis point-of-view, these data are challenging even for the more sophisticated state-of-the-art artificial intelligence methods. In particular, given the heterogeneity, complexity, and magnitude of the data, they can be employed in a variety of scenarios and analyzed through the prism of different machine learning tasks, such as multi-target regression, learning from data streams, anomaly detection, clustering, etc. Analyzing MEX's telemetry data is critical for aiding very important decisions regarding the spacecraft's status and operation, extracting novel knowledge, and monitoring the spacecraft's health, but the data can also be used to benchmark artificial intelligence methods designed for a variety of tasks. Measurement(s) electric current Technology Type(s) current readings in spacecraft housekeeping telemetry Sample Characteristic - Environment outer space [ABSTRACT FROM AUTHOR]

Montane temperate forests in central and southern Europe host diverse small mammal assemblages, but the fluctuations in these assemblages in correlation with owl predators are still poorly explored. The key questions of our study were how coexisting owls responded to different prey fluctuations and whether any particular small mammal species governed predator–prey co‐dynamics. We conducted a long‐term study (2004–2020) in low‐elevation (300–1100 m above sea level) mixed Beech and Silver Fir forest in the northern Dinaric Alps (central Slovenia). Monitoring data on the main small mammal groups – mice Muridae, voles Cricetidae, dormice Gliridae and shrews Soricidae – and three owl species – the Ural Owl Strix uralensis, Tawny Owl Strix aluco and Boreal Owl Aegolius funereus – were collected annually. To find relationships between prey and predator populations, we used two types of supervised machine learning approaches and addressed three predictive modelling tasks of multi‐target regression. The dominant species in the small mammal assemblage, the Yellow‐necked Mouse Apodemus flavicollis, had a key role in determining predator populations and their breeding performance. We noted higher sensitivity to small mammal fluctuations in boreal zone owl species (Boreal Owl and Ural Owl), which reach their southern distribution limit in the Dinaric Alps, whereas the temperate zone species (Tawny Owl) seemed to be less affected. In years of prey shortage, the Boreal Owl was found to presumably abandon its territories, the Ural Owl suppressed breeding and the Tawny Owl sustained breeding activity by shifting prey selection. Low‐elevation forests appeared to be suboptimal habitat for the competitive subordinate Boreal Owl, which may exploit occasional outbreaks of small mammal populations in these habitats even in the presence of larger competitors. Whether low‐elevation forests can play a role in maintaining threatened and cold‐adapted Boreal Owl populations in central and southern Europe in the face of recent ecosystem changes due to climate and environmental changes remains an open scientific question. [ABSTRACT FROM AUTHOR]

Feature ranking has been widely adopted in machine learning applications such as high-throughput biology and social sciences. The approaches of the popular Relief family of algorithms assign importances to features by iteratively accounting for nearest relevant and irrelevant instances. Despite their high utility, these algorithms can be computationally expensive and not-well suited for high-dimensional sparse input spaces. In contrast, recent embedding-based methods learn compact, low-dimensional representations, potentially facilitating down-stream learning capabilities of conventional learners. This paper explores how the Relief branch of algorithms can be adapted to benefit from (Riemannian) manifold-based embeddings of instance and target spaces, where a given embedding's dimensionality is intrinsic to the dimensionality of the considered data set. The developed ReliefE algorithm is faster and can result in better feature rankings, as shown by our evaluation on 20 real-life data sets for multi-class and multi-label classification tasks. The utility of ReliefE for high-dimensional data sets is ensured by its implementation that utilizes sparse matrix algebraic operations. Finally, the relation of ReliefE to other ranking algorithms is studied via the Fuzzy Jaccard Index. [ABSTRACT FROM AUTHOR]

We address the task of multi-target regression, where we generate global models that simultaneously predict multiple continuous variables. We use ensembles of generalized decision trees, called predictive clustering trees (PCTs), in particular bagging and random forests (RF) of PCTs and extremely randomized PCTs (extra PCTs). We add another dimension of randomization to these ensemble methods by learning individual base models that consider random subsets of target variables, while leaving the input space randomizations (in RF PCTs and extra PCTs) intact. Moreover, we propose a new ensemble prediction aggregation function, where the final ensemble prediction for a given target is influenced only by those base models that considered it during learning. An extensive experimental evaluation on a range of benchmark datasets has been conducted, where the extended ensemble methods were compared to the original ensemble methods, individual multi-target regression trees, and ensembles of single-target regression trees in terms of predictive performance, running times and model sizes. The results show that the proposed ensemble extension can yield better predictive performance, reduce learning time or both, without a considerable change in model size. The newly proposed aggregation function gives best results when used with extremely randomized PCTs. We also include a comparison with three competing methods, namely random linear target combinations and two variants of random projections. [ABSTRACT FROM AUTHOR]

Machine learning (ML) is a widely unexplored field in dendroclimatology, but it is a powerful tool that might improve the accuracy of climate reconstructions. In this paper, different ML algorithms are compared to climate reconstruction from tree-ring proxies. The algorithms considered are multiple linear regression (MLR), artificial neural networks (ANN), model trees (MT), bagging of model trees (BMT), and random forests of regression trees (RF). April-May mean temperature at a Quercus robur stand in Slovenia is predicted with mean vessel area (MVA, correlation coefficient with April-May mean temperature, r = 0.70, p < 0.001) and earlywood width (EW, r = -0.28, p < 0.05). Similarly, June-August mean temperature is predicted with stable carbon isotope (δ13C, r = 0.72, p < 0.001), stable oxygen isotope (δ18O, r = 0.32, p < 0.05) and tree-ring width (TRW, r = 0.11, p > 0.05 (ns)) chronologies. The predictive performance of ML algorithms was estimated by 3-fold cross-validation repeated 100 times. In both spring and summer temperature models, BMT performed best respectively in 62% and 52% of the 100 repetitions. The second-best method was ANN. Although BMT gave the best validation results, the differences in the models' performances were minor. We therefore recommend always comparing different ML regression techniques and selecting the optimal one for applications in dendroclimatology. [ABSTRACT FROM AUTHOR]

Copyright of Informatica (03505596) is the property of Slovene Society Informatika 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.)

Copyright of Acta Silvae et Ligni is the property of Biotechnical Faculty, Slovenian Forestry Institute 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.)

Multi-label classification (MLC) tasks are encountered more and more frequently in machine learning applications. While MLC methods exist for the classical batch setting, only a few methods are available for streaming setting. In this paper, we propose a new methodology for MLC via multi-target regression in a streaming setting. Moreover, we develop a streaming multi-target regressor iSOUP-Tree that uses this approach. We experimentally compare two variants of the iSOUP-Tree method (building regression and model trees), as well as ensembles of iSOUP-Trees with state-of-the-art tree and ensemble methods for MLC on data streams. We evaluate these methods on a variety of measures of predictive performance (appropriate for the MLC task). The ensembles of iSOUP-Trees perform significantly better on some of these measures, especially the ones based on label ranking, and are not significantly worse than the competitors on any of the remaining measures. We identify the thresholding problem for the task of MLC on data streams as a key issue that needs to be addressed in order to obtain even better results in terms of predictive performance. [ABSTRACT FROM AUTHOR]

The food- and airborne fungal genus Wallemia comprises seven xerophilic and halophilic species: W. sebi, W. mellicola, W. canadensis, W. tropicalis, W. muriae, W. hederae and W. ichthyophaga. All listed species are adapted to low water activity and can contaminate food preserved with high amounts of salt or sugar. In relation to food safety, the effect of high salt and sugar concentrations on the production of secondary metabolites by this toxigenic fungus was investigated. The secondary metabolite profiles of 30 strains of the listed species were examined using general growth media, known to support the production of secondary metabolites, supplemented with different concentrations of NaCl, glucose and MgCl2. In more than two hundred extracts approximately one hundred different compounds were detected using high-performance liquid chromatography-diode array detection (HPLC-DAD). Although the genome data analysis of W. mellicola (previously W. sebi sensu lato) and W. ichthyophaga revealed a low number of secondary metabolites clusters, a substantial number of secondary metabolites were detected at different conditions. Machine learning analysis of the obtained dataset showed that NaCl has higher influence on the production of secondary metabolites than other tested solutes. Mass spectrometric analysis of selected extracts revealed that NaCl in the medium affects the production of some compounds with substantial biological activities (wallimidione, walleminol, walleminone, UCA 1064-A and UCA 1064-B). In particular an increase in NaCl concentration from 5% to 15% in the growth media increased the production of the toxic metabolites wallimidione, walleminol and walleminone. [ABSTRACT FROM AUTHOR]

We address the task of hierarchical multi-label classification (HMC). HMC is a task of structured output prediction where the classes are organized into a hierarchy and an instance may belong to multiple classes. In many problems, such as gene function prediction or prediction of ecological community structure, classes inherently follow these constraints. The potential for application of HMC was recognized by many researchers and several such methods were proposed and demonstrated to achieve good predictive performances in the past. However, there is no clear understanding when is favorable to consider such relationships (hierarchical and multi-label) among classes, and when this presents unnecessary burden for classification methods. To this end, we perform a detailed comparative study over 8 datasets that have HMC properties. We investigate two important influences in HMC: the multiple labels per example and the information about the hierarchy. More specifically, we consider four machine learning tasks: multi-label classification, hierarchical multi-label classification, single-label classification and hierarchical single-label classification. To construct the predictive models, we use predictive clustering trees (a generalized form of decision trees), which are able to tackle each of the modelling tasks listed. Moreover, we investigate whether the influence of the hierarchy and the multiple labels carries over for ensemble models. For each of the tasks, we construct a single tree and two ensembles (random forest and bagging). The results reveal that the hierarchy and the multiple labels do help to obtain a better single tree model, while this is not preserved for the ensemble models. [ABSTRACT FROM AUTHOR]

Multi-label learning (MLL) problems abound in many areas, including text categorization, protein function classification, and semantic annotation of multimedia. Issues that severely limit the applicability of many current machine learning approaches to MLL are the large-scale problem, which have a strong impact on the computational complexity of learning. These problems are especially pronounced for approaches that transform MLL problems into a set of binary classification problems for which Support Vector Machines (SVMs) are used. On the other hand, the most efficient approaches to MLL, based on decision trees, have clearly lower predictive performance. We propose a hybrid decision tree architecture, where the leaves do not give multi-label predictions directly, but rather utilize local SVM-based classifiers giving multi-label predictions. A binary relevance architecture is employed in the leaves, where a binary SVM classifier is built for each of the labels relevant to that particular leaf. We use a broad range of multi-label datasets with a variety of evaluation measures to evaluate the proposed method against related and state-of-the-art methods, both in terms of predictive performance and time complexity. Our hybrid architecture on almost every large classification problem outperforms the competing approaches in terms of the predictive performance, while its computational efficiency is significantly improved as a result of the integrated decision tree. [ABSTRACT FROM AUTHOR]

Abstract: Spatial autocorrelation is the correlation among data values which is strictly due to the relative spatial proximity of the objects that the data refer to. Inappropriate treatment of data with spatial dependencies, where spatial autocorrelation is ignored, can obfuscate important insights. In this paper, we propose a data mining method that explicitly considers spatial autocorrelation in the values of the response (target) variable when learning predictive clustering models. The method is based on the concept of predictive clustering trees (PCTs), according to which hierarchies of clusters of similar data are identified and a predictive model is associated to each cluster. In particular, our approach is able to learn predictive models for both a continuous response (regression task) and a discrete response (classification task). We evaluate our approach on several real world problems of spatial regression and spatial classification. The consideration of the autocorrelation in the models improves predictions that are consistently clustered in space and that clusters try to preserve the spatial arrangement of the data, at the same time providing a multi-level insight into the spatial autocorrelation phenomenon. The evaluation of SCLUS in several ecological domains (e.g. predicting outcrossing rates within a conventional field due to the surrounding genetically modified fields, as well as predicting pollen dispersal rates from two lines of plants) confirms its capability of building spatial aware models which capture the spatial distribution of the target variable. In general, the maps obtained by using SCLUS do not require further post-smoothing of the results if we want to use them in practice. [Copyright &y& Elsevier]

The problem of real-time extraction of meaningful patterns from time-changing data streams is of increasing importance for the machine learning and data mining communities. Regression in time-changing data streams is a relatively unexplored topic, despite the apparent applications. This paper proposes an efficient and incremental stream mining algorithm which is able to learn regression and model trees from possibly unbounded, high-speed and time-changing data streams. The algorithm is evaluated extensively in a variety of settings involving artificial and real data. To the best of our knowledge there is no other general purpose algorithm for incremental learning regression/model trees able to perform explicit change detection and informed adaptation. The algorithm performs online and in real-time, observes each example only once at the speed of arrival, and maintains at any-time a ready-to-use model tree. The tree leaves contain linear models induced online from the examples assigned to them, a process with low complexity. The algorithm has mechanisms for drift detection and model adaptation, which enable it to maintain accurate and updated regression models at any time. The drift detection mechanism exploits the structure of the tree in the process of local change detection. As a response to local drift, the algorithm is able to update the tree structure only locally. This approach improves the any-time performance and greatly reduces the costs of adaptation. [ABSTRACT FROM AUTHOR]

Abstract: The aim of this paper is to discover a model equation for predicting the concentration of the algal species Peridinium gatunense (Dinoflagellate) in Lake Kinneret. This is a rather difficult task, due to the sudden ecosystem changes that occurred in the mid-1990s. Namely, the stable ecosystem (with regular Peridinium blooms until 1993) underwent changes and has transformed into an unstable system, with cyanobacterial blooms now occurring regularly. This shift in the algal succession is expected to influence attempts to model the lake ecosystem. Namely, the model structure before and after the change is likely to be different. Our modelling experiments were directed to discover a single model equation that can simulate dinoflagellate dynamics in both periods. We apply an automated modelling tool (Lagramge), which integrates the knowledge- and the data-driven modelling approach. In addition we include an expert visual estimation of the models discovered by Lagramge to assist in the selection of the optimal model. The dataset used included time-series measurements of typical data from the periods 1988 to 1992 and 1997 to 1999. Using the data and expert knowledge coded in a modelling knowledge library, Lagramge successfully discovered several suitable mathematical models for Peridinium. After the expert’s visual estimation and validation of the models, we propose one optimal model capable of long-term predictions. [Copyright &y& Elsevier]

Abstract: High quality information on forest resources is important to forest ecosystem management. Traditional ground measurements are labor and resource intensive and at the same time expensive and time consuming. For most of the Slovenian forests, there is extensive ground-based information on forest properties of selected sample locations. However there is no continuous information of objectively measured vegetation height and canopy cover at appropriate resolution. Currently, Light Detection And Ranging (LiDAR) technology provides detailed measurements of different forest properties because of its immediate generation of 3D data, its accuracy and acquisition flexibility. However, existing LiDAR sensors have limited spatial coverage and relatively high cost of acquisition. Satellite data, on the other hand, are low-cost and offer broader spatial coverage of generalized forest structure, but are not expected to provide accurate information about vegetation height. Integration of LiDAR and satellite data promises to improve the measurement, mapping, and monitoring of forest properties. The primary objective of this study is to model the vegetation height and canopy cover in Slovenia by integrating LiDAR data, Landsat satellite data, and the use of machine learning techniques. This kind of integration uses the accuracy and precision of LiDAR data and the wide coverage of satellite data in order to generate cost-effective realistic estimates of the vegetation height and canopy cover, and consequently generate continuous forest vegetation map products to be used in forest management and monitoring. Several machine learning techniques are applied to this task: they are evaluated and their performance is compared by using statistical significance tests. Ensemble methods perform significantly better than single- and multi-target regression trees and are further used for the generation of forest maps. Such maps are used for land-cover and land-use classification, as well as for monitoring and managing ongoing forest processes (like spontaneous afforestation, forest reduction and forest fires) that affect the stability of forest ecosystems. [Copyright &y& Elsevier]

An important consideration in conservation and biodiversity planning is an appreciation of the condition or integrity of ecosystems. In this study, we have applied various machine learning methods to the problem of predicting the condition or quality of the remnant indigenous vegetation across an extensive area of south-eastern Australia—the state of Victoria. The field data were obtained using the ‘habitat hectares’ approach. This rapid assessment technique produces multiple scores that describe the condition of various attributes of the vegetation at a given site. Multiple sites were assessed and subsequently circumscribed with GIS and remote-sensed data. We explore and compare two approaches for modelling this type of data: to learn a model for each score separately (single-target approach, a regression tree), or to learn one model for all scores simultaneously (multi-target approach, a multi-target regression tree). In order to lift the predictive performance, we also employ ensembles (bagging and random forests) of regression trees and multi-target regression trees. Our results demonstrate the advantages of a multi-target over a single-target modelling approach. While there is no statistically significant difference between the multi-target and single-target models in terms of model performance, the multi-target models are smaller and faster to learn than the single-target ones. Ensembles of multi-target models, also, improve the spatial prediction of condition. The usefulness of models of vegetation condition is twofold. First, they provide an enhanced knowledge and understanding of the condition of different indigenous vegetation types, and identify possible biophysical and landscape attributes that may contribute to vegetation decline. Second, these models may be used to map the condition of indigenous vegetation, in support of biodiversity planning, management and investment decisions. [Copyright &y& Elsevier]

Reconstructing biological networks, such as metabolic and signaling networks, is at the heart of systems biology. Although many approaches exist for reconstructing network structure, few approaches recover the full dynamic behavior of a network. We survey such approaches that originate from computational scientific discovery, a subfield of machine learning. These take as input measured time course data, as well as existing domain knowledge, such as partial knowledge of the network structure. We demonstrate the use of these approaches on illustrative tasks of finding the complete dynamics of biological networks, which include examples of rediscovering known networks and their dynamics, as well as examples of proposing models for unknown networks. [Copyright &y& Elsevier]

Abstract: In this paper, we present a framework for modeling dynamic systems that integrates the knowledge-based theoretical approach to modeling with the data-driven empirical modeling. The framework allows for integration of modeling knowledge specific to the domain of interest in the process of model induction from measured data. The knowledge is organized around the central notion of basic processes in the domain and it includes models thereof as well as guidelines for combining models of individual processes into a model of the entire observed system. The presented framework is applied to three tasks of modeling dynamic environmental systems from noisy measurement data in the domains of population and hydro dynamics. In all applications, the models induced with the framework can be used both to accurately predict and explain the behavior of the observed dynamic systems. [Copyright &y& Elsevier]

Abstract: A palaeoecological study was conducted to investigate past environmental conditions and vegetation dynamics around the southwestern Ljubljana Moor. In order to find potential regularities and/or dependencies among co-existent plant species through time, different machine learning methods were applied to pollen records from the cores taken at Bistra and Hočevarica. The data comprised relative pollen frequencies of the most common plant genera/families at particular core depths that correspond to particular ages in the Early and Mid Holocene periods. The applied methods include equation discovery and hierarchical clustering. Both methods have found plausible and explainable relationships among identified plant genera/families. [Copyright &y& Elsevier]