[Show abstract][Hide abstract] ABSTRACT: Discovering changes in the data distribution of streams and discovering recurrent data distributions are challenging problems in data mining and machine learning. Both have received a lot of attention in the context of classification. With the ever increasing growth of data, however, there is a high demand of compact and universal representations of data streams that enable the user to analyze current as well as historic data without having access to the raw data. To make a first step towards this direction, we propose a condensed representation that captures the various – possibly recurrent – data distributions of the stream by extending the notion of possible worlds. The representation enables queries concerning the whole stream and can, hence, serve as a tool for supporting decision-making processes or serve as a basis for implementing data mining and machine learning algorithms on top of it. We evaluate this condensed representation on synthetic and real-world data.
[Show abstract][Hide abstract] ABSTRACT: Data mining and machine learning algorithms usually operate directly on the data. However, if the data is not available at once or consists of billions of instances, these algorithms easily become infeasible with respect to memory and run-time concerns. As a solution to this problem, we propose a framework, called MiDEO (Mining Density Estimates inferred Online), in which algorithms are designed to operate on a condensed representation of the data. In particular, we propose to use density estimates, which are able to represent billions of instances in a compact form and can be updated when new instances arrive. As an example for an algorithm that operates on density estimates, we consider the task of mining association rules, which we consider as a form of simple statements about the data. The algorithm, called POEt (Pattern mining on Online density esTimates), is evaluated on synthetic and real-world data and is compared to state-of-the-art algorithms.
[Show abstract][Hide abstract] ABSTRACT: Background
Sound statistical validation is important to evaluate and compare the overall performance of (Q)SAR models. However, classical validation does not support the user in better understanding the properties of the model or the underlying data. Even though, a number of visualization tools for analyzing (Q)SAR information in small molecule datasets exist, integrated visualization methods that allow the investigation of model validation results are still lacking.
We propose visual validation, as an approach for the graphical inspection of (Q)SAR model validation results. The approach applies the 3D viewer CheS-Mapper, an open-source application for the exploration of small molecules in virtual 3D space. The present work describes the new functionalities in CheS-Mapper 2.0, that facilitate the analysis of (Q)SAR information and allows the visual validation of (Q)SAR models. The tool enables the comparison of model predictions to the actual activity in feature space. The approach is generic: It is model-independent and can handle physico-chemical and structural input features as well as quantitative and qualitative endpoints.
Visual validation with CheS-Mapper enables analyzing (Q)SAR information in the data and indicates how this information is employed by the (Q)SAR model. It reveals, if the endpoint is modeled too specific or too generic and highlights common properties of misclassified compounds. Moreover, the researcher can use CheS-Mapper to inspect how the (Q)SAR model predicts activity cliffs. The CheS-Mapper software is freely available at http://ches-mapper.org.
Comparing actual and predicted activity values with CheS-Mapper.
Preview · Article · Sep 2014 · Journal of Cheminformatics
[Show abstract][Hide abstract] ABSTRACT: Statistical machine learning algorithms building on patterns found by pattern mining algorithms have to cope with large solution sets and thus the high dimensionality of the feature space. Vice versa, pattern mining algorithms are frequently applied to irrelevant instances, thus causing noise in the output. Solution sets of pattern mining algorithms also typically grow with increasing input datasets. The paper proposes an approach to overcome these limitations. The approach extracts information from trained support vector machines, in particular their support vectors and their relevance according to their coefficients. It uses the support vectors along with their coefficients as input to pattern mining algorithms able to handle weighted instances. Our experiments in the domain of graph mining and molecular graphs show that the resulting models are not significantly less accurate than models trained on the full datasets, yet require only a fraction of the time using much smaller sets of patterns.
[Show abstract][Hide abstract] ABSTRACT: We address the problem of estimating a discrete joint density online, that is, the algorithm is only provided the current example and its current estimate. The proposed online estimator of discrete densities, EDDO (Estimation of Discrete Densities Online), uses classifier chains to model dependencies among features. Each classifier in the chain estimates the probability of one particular feature. Because a single chain may not provide a reliable estimate, we also consider ensembles of classifier chains and ensembles of weighted classifier chains. For all density estimators, we provide consistency proofs and propose algorithms to perform certain inference tasks. The empirical evaluation of the estimators is conducted in several experiments and on data sets of up to several million instances: We compare them to density estimates computed from Bayesian structure learners, evaluate them under the influence of noise, measure their ability to deal with concept drift, and measure the run-time performance. Our experiments demonstrate that, even though designed to work online, EDDO delivers estimators of competitive accuracy compared to batch Bayesian structure learners and batch variants of EDDO.
[Show abstract][Hide abstract] ABSTRACT: (Q)SAR model validation is essential to ensure the quality of inferred models and to indicate future model predictivity on unseen compounds. Proper validation is also one of the requirements of regulatory authorities in order to accept the (Q)SAR model, and to approve its use in real world scenarios as alternative testing method. However, at the same time, the question of how to validate a (Q)SAR model, in particular whether to employ variants of cross-validation or external test set validation, is still under discussion. In this paper, we empirically compare a k-fold cross-validation with external test set validation. To this end we introduce a workflow allowing to realistically simulate the common problem setting of building predictive models for relatively small datasets. The workflow allows to apply the built and validated models on large amounts of unseen data, and to compare the performance of the different validation approaches. The experimental results indicate that cross-validation produces higher performant (Q)SAR models than external test set validation, reduces the variance of the results, while at the same time underestimates the performance on unseen compounds. The experimental results reported in this paper suggest that, contrary to current conception in the community, cross-validation may play a significant role in evaluating the predictivity of (Q)SAR models.
[Show abstract][Hide abstract] ABSTRACT: In recent years, graph kernels have received considerable interest within the machine learning and data mining community. Here, we introduce a novel approach enabling kernel methods to utilize additional information hidden in the structural neighborhood of the graphs under consideration. Our novel structural cluster kernel (SCK) incorporates similarities induced by a structural clustering algorithm to improve state-of-the-art graph kernels. The approach taken is based on the idea that graph similarity can not only be described by the similarity between the graphs themselves, but also by the similarity they possess with respect to their structural neighborhood. We applied our novel kernel in a supervised and a semi-supervised setting to regression and classification problems on a number of real-world datasets of molecular graphs. Our results show that the structural cluster similarity information can indeed leverage the prediction performance of the base kernel, particularly when the dataset is structurally sparse and consequently structurally diverse. By additionally taking into account a large number of unlabeled instances the performance of the structural cluster kernel can further be improved.
[Show abstract][Hide abstract] ABSTRACT: Analyzing chemical datasets is a challenging task for scientific researchers in the field of chemoinformatics. It is important, yet difficult to understand the relationship between the structure of chemical compounds, their physico-chemical properties, and biological or toxic effects. To that respect, visualization tools can help to better comprehend the underlying correlations. Our recently developed 3D molecular viewer CheS-Mapper (Chemical Space Mapper) divides large datasets into clusters of similar compounds and consequently arranges them in 3D space, such that their spatial proximity reflects their similarity. The user can indirectly determine similarity, by selecting which features to employ in the process. The tool can use and calculate different kind of features, like structural fragments as well as quantitative chemical descriptors. These features can be highlighted within CheS-Mapper, which aids the chemist to better understand patterns and regularities and relate the observations to established scientific knowledge. As a final function, the tool can also be used to select and export specific subsets of a given dataset for further analysis.
Preview · Article · Mar 2012 · Journal of Cheminformatics
[Show abstract][Hide abstract] ABSTRACT: We present a novel approach to incrementally determine the trajectory of a person in 3-D based on its motions and activities in real time. In our algorithm, we estimate the motions and activities of the user given the data that are obtained from a motion capture suit equipped with several inertial measurement units. These activities include walking up and down staircases, as well as opening and closing doors. We interpret the first two types of activities as motion constraints and door-handling events as landmark detections in a graph-based simultaneous localization and mapping (SLAM) framework. Since we cannot distinguish between individual doors, we employ a multihypothesis tracking approach on top of the SLAM procedure to deal with the high data-association uncertainty. As a result, we are able to accurately and robustly recover the trajectory of the person. Additionally, we present an algorithm to build approximate geometrical and topological maps based on the estimated trajectory and detected activities. We evaluate our approach in practical experiments that are carried out with different subjects and in various environments.
Full-text · Article · Feb 2012 · IEEE Transactions on Robotics
[Show abstract][Hide abstract] ABSTRACT: OpenTox provides an interoperable, standards-based Framework for the support of predictive toxicology data management, algorithms, modelling, validation and reporting. It is relevant to satisfying the chemical safety assessment requirements of the REACH legislation as it supports access to experimental data, (Quantitative) Structure-Activity Relationship models, and toxicological information through an integrating platform that adheres to regulatory requirements and OECD validation principles. Initial research defined the essential components of the Framework including the approach to data access, schema and management, use of controlled vocabularies and ontologies, architecture, web service and communications protocols, and selection and integration of algorithms for predictive modelling. OpenTox provides end-user oriented tools to non-computational specialists, risk assessors, and toxicological experts in addition to Application Programming Interfaces (APIs) for developers of new applications. OpenTox actively supports public standards for data representation, interfaces, vocabularies and ontologies, Open Source approaches to core platform components, and community-based collaboration approaches, so as to progress system interoperability goals.
The OpenTox Framework includes APIs and services for compounds, datasets, features, algorithms, models, ontologies, tasks, validation, and reporting which may be combined into multiple applications satisfying a variety of different user needs. OpenTox applications are based on a set of distributed, interoperable OpenTox API-compliant REST web services. The OpenTox approach to ontology allows for efficient mapping of complementary data coming from different datasets into a unifying structure having a shared terminology and representation.
Two initial OpenTox applications are presented as an illustration of the potential impact of OpenTox for high-quality and consistent structure-activity relationship modelling of REACH-relevant endpoints: ToxPredict which predicts and reports on toxicities for endpoints for an input chemical structure, and ToxCreate which builds and validates a predictive toxicity model based on an input toxicology dataset. Because of the extensible nature of the standardised Framework design, barriers of interoperability between applications and content are removed, as the user may combine data, models and validation from multiple sources in a dependable and time-effective way.
Full-text · Article · Aug 2010 · Journal of Cheminformatics
[Show abstract][Hide abstract] ABSTRACT: We present a novel approach to build approximate maps of structured environments utilizing human motion and activity. Our approach uses data recorded with a data suit which is equipped with several IMUs to detect movements of a person and door opening and closing events. In our approach we interpret the movements as motion constraints and door handling events as landmark detections in a graph-based SLAM framework. As we cannot distinguish between individual doors, we employ a multi-hypothesis approach on top of the SLAM system to deal with the high data-association uncertainty. As a result, our approach is able to accurately and robustly recover the trajectory of the person. We additionally take advantage of the fact that people traverse free space and that doors separate rooms to recover the geometric structure of the environment after the graph optimization. We evaluate our approach in several experiments carried out with different users and in environments of different types.
[Show abstract][Hide abstract] ABSTRACT: Relational data is complex. This complexity makes one of the basic steps
of ILP difficult: understanding the data and results. If the user cannot
easily understand it, he draws incomplete conclusions. The situation is
very much as in the parable of the blind men and the elephant that
appears in many cultures. In this tale the blind work independently and
with quite different pieces of information, thereby drawing very
different conclusions about the nature of the beast. In contrast, visual
representations make it easy to shift from one perspective to another
while exploring and analyzing data. This paper describes a method for
embedding interpretations and queries into a single, common Euclidean
space based on their co-proven statistics. We demonstrate our method on
real-world datasets showing that ILP results can indeed be captured at a
[Show abstract][Hide abstract] ABSTRACT: Scheme-specific attribute selection with the wrapper and variants of forward selection is a popular attribute selection technique for classification that yields good results. However, it can run the risk of overfitting because of the extent of the search and the extensive use of internal cross-validation. Moreover, although wrapper evaluators tend to achieve superior accuracy compared to filters, they face a high computational cost. The problems of overfitting and high runtime occur in particular on high-dimensional datasets, like microarray data. We investigate Linear Forward Selection, a technique to reduce the number of attributes expansions in each forward selection step. Our experiments demonstrate that this approach is faster, finds smaller subsets and can even increase the accuracy compared to standard forward selection. We also investigate a variant that applies explicit subset size determination in forward selection to combat overfitting, where the search is forced to stop at a precomputed ldquooptimalrdquo subset size. We show that this technique reduces subset size while maintaining comparable accuracy.
[Show abstract][Hide abstract] ABSTRACT: The task of aligning sequences arises in many applications. Classical dynamic programming approaches require the explicit state enumeration in the reward model. This is often impractical: the number of states grows very quickly with the number of domain objects and relations among these objects. Relational sequence alignment aims at exploiting symbolic structure to avoid the full enumeration. This comes at the expense of a more complex reward model selection problem: virtually infinitely many abstraction levels have to be explored. In this paper, we apply gradient-based boosting to leverage this problem. Specifically, we show how to reduce the learning problem to a series of relational regressions problems. The main benefit of this is that interactions between states variables are introduced only as needed, so that the potentially infinite search space is not explicitly considered. As our experimental results show, this boosting approach can significantly improve upon established results in challenging applications.
[Show abstract][Hide abstract] ABSTRACT: Sequential behavior and sequence learning are essential to intelligence. Often the elements of sequences exhibit an internal
structure that can elegantly be represented using relational atoms. Applying traditional sequential learning techniques to
such relational sequences requires one either to ignore the internal structure or to live with a combinatorial explosion of
the model complexity. This chapter briefly reviews relational sequence learning and describes several techniques tailored
towards realizing this, such as local pattern mining techniques, (hidden) Markov models, conditional random fields, dynamic
programming and reinforcement learning.
[Show abstract][Hide abstract] ABSTRACT: The need to measure sequence similarity arises in informa- tion extraction, music mining, biological sequence analysis, and other domains, and often coincides with sequence alignment: the more similar two sequences are, the better they can be aligned. Aligning sequences not only shows how similar sequences are, it also shows where there are dierences and correspondences between the sequences. Traditionally, the alignment has been considered for sequences of flat symbols only. Many real world sequences such as protein secondary struc- tures, however, exhibit a rich internal structures. This is akin to the problem of dealing with structured examples studied in the field of in- ductive logic programming (ILP). In this paper, we propose to use well- established ILP distance measures within alignment methods. Although straight-forward, our initial experimental results show that this approach performs well in practice and is worth to be explored.
[Show abstract][Hide abstract] ABSTRACT: Most approaches to structure-activity-relationship (SAR) prediction proceed in two steps. In the first step, a typically large set of fingerprints, or fragments of interest, is constructed (either by hand or by some recent data mining techniques). In the second step, machine learning techniques are applied to obtain a predictive model. The result is often not only a highly accurate but also hard to interpret model. In this paper, we demonstrate the capabilities of a novel SAR algorithm, SMIREP, which tightly integrates the fragment and model generation steps and which yields simple models in the form of a small set of IF-THEN rules. These rules contain SMILES fragments, which are easy to understand to the computational chemist. SMIREP combines ideas from the well-known IREP rule learner with a novel fragmentation algorithm for SMILES strings. SMIREP has been evaluated on three problems: the prediction of binding activities for the estrogen receptor (Environmental Protection Agency's (EPA's) Distributed Structure-Searchable Toxicity (DSSTox) National Center for Toxicological Research estrogen receptor (NCTRER) Database), the prediction of mutagenicity using the carcinogenic potency database (CPDB), and the prediction of biodegradability on a subset of the Environmental Fate Database (EFDB). In these applications, SMIREP has the advantage of producing easily interpretable rules while having predictive accuracies that are comparable to those of alternative state-of-the-art techniques.
Preview · Article · Feb 2007 · Journal of Chemical Information and Modeling
[Show abstract][Hide abstract] ABSTRACT: This paper is a manifesto aimed at computer scientists interested in developing and applying scientific discovery methods.
It argues that: science is experiencing an unprecedented “explosion” in the amount of available data; traditional data analysis
methods cannot deal with this increased quantity of data; there is an urgent need to automate the process of refining scientific
data into scientific knowledge; inductive logic programming (ILP) is a data analysis framework well suited for this task;
and exciting new scientific discoveries can be achieved using ILP scientific discovery methods. We describe an example of
using ILP to analyse a large and complex bioinformatic database that has produced unexpected and interesting scientific results
in functional genomics. We then point a possible way forward to integrating machine learning with scientific databases to
form intelligent databases.