No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Interpretable ontology meta-matching in the biomedical domain using Mamdani fuzzy inference
Abstract
Ontology meta-matching techniques have been consolidated as one of the best approaches to face the problem of discovering semantic relationships between knowledge models that belong to the same domain but have been developed independently. After more than a decade of research, the community has reached a stage of maturity characterized by increasingly better results and aspects such as the robustness and scalability of solutions have been solved. However, the resulting models remain practically intelligible to a human operator. In this work, we present a novel approach based on Mamdani fuzzy inference exploiting a model very close to natural language. This fact has a double objective: to achieve results with high degrees of accuracy but at the same time to guarantee the interpretability of the resulting models. After validating our proposal with several ontological models popular in the biomedical field, we can conclude that the results obtained are promising.
... In order to overcome this problem, there is work that relies on the use of fuzzy logics; specifically Mamdani models [18], to build solutions that can have some meaning for the users who use them. In particular, fuzzy systems based on Mamdani's postulates are quite interpretable since they operate on the basis of simple rules. ...
... However, heterogeneity is a critical issue which makes the task of finding semantic correspondences difficult. To overcome this problem, several solutions have been already proposed [19,20,21,18,22,23,24]. ...
... Using other models, such as fuzzy logics, also leads to more interpretable models, as shown in [18]. However, many issues inherent to such models must be studied: their structure, readability, the number of terms involved, the number and length of the rules that implement them, etc. ...
The problem of ontology matching consists of finding the semantic correspondences between two ontologies that, although belonging to the same domain, have been developed separately. Ontology matching methods are of great importance today since they allow us to find the pivot points from which an automatic data integration process can be established. Unlike the most recent developments based on deep learning, this study presents our research efforts on the development of novel methods for ontology matching that are accurate and interpretable at the same time. For this purpose, we rely on a symbolic regression model (implemented via genetic programming) that has been specifically trained to find the mathematical expression that can solve the ground truth provided by experts accurately. Moreover, our approach offers the possibility of being understood by a human operator and helping the processor to consume as little energy as possible. The experimental evaluation results that we have achieved using several benchmark datasets seem to show that our approach could be promising.
Ontologies provide a standardized approach to knowledge representation that can be shared across various domains. By extracting heterogeneous ontology alignment, E-businesses can efficiently exchange information and enhance communication, decision-making, and reduce data integration costs. In this study, we investigate the heterogeneous ontology alignment extraction problem for E-business, which aims to determine an optimal concept pair set with the highest f-measure value. Given the alignment extraction’s inherent complexity, we use a Genetic Algorithm (GA) to address it. In particular, we first model the HOAEP as a multi-modal problem with sparse solutions and then propose a Compact Co-Evolutionary Niching Genetic Algorithm (CCNGA) to address it. CCNGA first employs probability distribution estimation to simplify population representation, and then uses three evolutionary strategies to simultaneously search for the global optimum. The experimental testing cases include OAEI’s Conference track and three real E-business ontologies, and T-test results demonstrate that CCNGA significantly outperforms other state-of-the-art ontology alignment extraction techniques.
The electron ionisation (EI) ion source has been widely used in the development of the mass spectrometer, and the stability of emission current of EI ion source directly affects important performance such as the sensitivity of the mass spectrometer. Realizing the stable control of emission current is the key to ensuring the accurate measurement of the mass spectrometer. However, the stability of the emission current is easily affected by factors such as filament resistance and pressure of the mass spectrometer chamber. For adaptive control of emission current, two adaptive control methods based on negative feedback operational amplifier and fuzzy logic system (FLS) are proposed and applied to the self-developed quadrupole mass spectrometer (QMS). Further, three emission current control experiments when the QMS is powered on, works stably, and the pressure changes are designed to verify the effectiveness of the adaptive control method proposed. Experimental results show that the two adaptive control methods in this paper are compared to the open-loop control method of stabilizing the output of the heating power supply have faster control response rate, longer stable control time, and better adaptive control ability.
Discovering the semantic relationships among heterogeneous ontologies has been one of the core research topics in Semantic Web. As ontology matching systems inevitably adopt heuristic strategies, wrong mappings are often contained in final alignments. Most methods for mapping revision depend on dealing with logical incoherence. However, erroneous mappings that do not cause incoherence may be left out. Hence, manual validations with domain expertise are needed. Nevertheless, existing interactive methods for mapping revision still suffer from two limitations. Firstly, revision methods designed to present friendly graphical interfaces and contextual information related to mappings pay little attention to reducing manual decisions, leading to the efficiency problem. Secondly, revision methods focusing on automated reasoning to reduce manual decisions employ models of high complexity, leading to the practicality problem when dealing with alignments across large-scale ontologies. To address these problems, we propose a novel graph-based method for interactive mapping revision, aiming to reduce the manual efforts as much as possible. DL-Lite ontologies and their mappings are encoded into an integrated graph, where the mapping arcs will be judged by the experts. We define the decision space tailored for mapping revision, which can be used to improve the efficiency of manual making decisions. After a manual decision is made in each interaction, the mapping arcs will be automatically updated in the integrated graph. The whole update process modeled in our defined graph-based decision space could be accomplished in polynomial time. We further design an impact function based on the integrated graph and weights of mappings, which can display the most influential mappings to experts. In this way, the number of manual decisions can be reduced further. To cope with the practicality of our method for alignments of large-scale ontologies, we introduce the notion of “reliable” mappings as an attempt to alleviate the burden of experts for making decisions, and propose two soft principles to ensure the reliability of selected mappings. Moreover, we define influence relation and design a corresponding algorithm to enhance the method for detecting incoherence of the integrated graph, which is transformed from the ontologies beyond DL-Lite and their mappings. We implement our method and evaluate its efficiency by 16 alignments generated across real-world ontologies. Experimental results show that our proposed method can improve the efficiency by 19% on average and save more manual decisions than other interactive revision methods in most cases.
Ontology presents a field knowledge by defining the meaning of domain concepts and relations among them. To conduct the inter-operability between two heterogeneous ontologies, we need to determine a number of entity correspondences, the so-called Ontology Matching (OM). Traditional OM techniques dedicate to find the simple correspondences which cannot sufficiently express the diverse kinds of heterogeneity, and therefore, complex correspondences are required, which maps one or multiple entities from one ontology to those from the other. Due to the semantic complexity, automatically discovering large-scale simple and complex correspondences has become a challenge. To face this challenge, in this work, an Anchor-based Semantic Partitioning Technique (ASPT) is first presented to divide the large-scale ontologies into several semantic segments, whose computational complexity is lower than the existing ontology partitioning algorithms. Then, a Pattern-based Similarity Measure (PSM) is used to distinguish both simple and complex correspondences. Finally, a Confidence Matrix based Evolutionary Algorithm (CM-EA) is proposed to efficiently match the ontology segments. We compare our proposal with EA-based matching technique and state-of-the-art ontology matching systems on OAEI’s Complex track, and the experimental results show that our approach is able to efficiently match large-scale complex ontologies.
Ontology applies commonly to solve the problem of heterogeneity of data in the Semantic Web, but the heterogeneity problem between two ontologies seriously affects their communication. As an effective method, ontology matching can address the problem above, whose core technique is the similarity measure. A single similarity measure calculates the similarity value about a feature between two concepts, but none of the similarity measures can ensure their effectiveness in all context due to the diverse heterogeneous features between two ontologies. Therefore, multiple similarity measures are usually aggregated to improve the result’s confidence. The problem that how to determine the optimal aggregating weights for the different similarity measures to obtain a high-quality alignment is called the meta-matching problem of ontology, which is modeled as a nonlinear problem with many local optimal solutions. Evolutionary Algorithm (EA) can represent an efficient methodology to address the ontology meta-matching problem, but EA-based ontology matching techniques suffer from the premature convergence and the requirement of a reference alignment to evaluate the solutions. To overcome the defects mentioned above, in this work, an improved EA-based matching approach is proposed, where two approximate evaluation indicators, i.e. pseudo-recall and pseudo-precision, are presented to evaluate the solution’s quality, and an adaptive selection pressure is utilized to overcome the premature convergence. The experiment utilizes the Ontology Alignment Evaluation Initiative (OAEI)’s benchmark, and the experimental results will prove the effectiveness of our proposed method.
Ontology Matching (OM) is performed to find semantic correspondences between the entity elements of different ontologies to enable semantic integration, reuse, and interoperability. Representation learning techniques have been introduced to the field of OM with the development of deep learning. However, there still exist two limitations. Firstly, these methods only focus on the terminological-based features to learn word vectors for discovering mappings, ignoring the network structure of ontology. Secondly, the final alignment threshold is usually determined manually within these methods. It is difficult for an expert to adjust the threshold value and even more so for a non-expert user. To address these issues, we propose an alternative ontology matching framework called Deep Attentional Embedded Ontology Matching (DAEOM), which models the matching process by embedding techniques with jointly encoding ontology terminological description and network structure. We propose a novel inter-intra negative sampling skill tailored for the structural relations asserted in ontologies, and further improve our iterative final alignment method by introducing an automatic adjustment of the final alignment threshold. The preliminary result on real-world biomedical ontologies indicates that DAEOM is competitive with several OAEI top-ranked systems in terms of F-measure.
Purpose
Although ontology matchers are annually proposed to address different aspects of the semantic heterogeneity problem, finding the most suitable alignment approach is still an issue. This study aims to propose a computational solution for ontology meta-matching (OMM) and a framework designed for developers to make use of alignment techniques in their applications.
Design/methodology/approach
The framework includes some similarity functions that can be chosen by developers and then, automatically, set weights for each function to obtain better alignments. To evaluate the framework, several simulations were performed with a data set from the Ontology Alignment Evaluation Initiative. Simple similarity functions were used, rather than aligners known in the literature, to demonstrate that the results would be more influenced by the proposed meta-alignment approach than the functions used.
Findings
The results showed that the framework is able to adapt to different test cases. The approach achieved better results when compared with existing ontology meta-matchers.
Originality/value
Although approaches for OMM have been proposed, it is not easy to use them during software development. On the other hand, this work presents a framework that can be used by developers to align ontologies. New ontology matchers can be added and the framework is extensible to new methods. Moreover, this work presents a novel OMM approach modeled as a linear equation system which can be easily computed.
Background:
While representation learning techniques have shown great promise in application to a number of different NLP tasks, they have had little impact on the problem of ontology matching. Unlike past work that has focused on feature engineering, we present a novel representation learning approach that is tailored to the ontology matching task. Our approach is based on embedding ontological terms in a high-dimensional Euclidean space. This embedding is derived on the basis of a novel phrase retrofitting strategy through which semantic similarity information becomes inscribed onto fields of pre-trained word vectors. The resulting framework also incorporates a novel outlier detection mechanism based on a denoising autoencoder that is shown to improve performance.
Results:
An ontology matching system derived using the proposed framework achieved an F-score of 94% on an alignment scenario involving the Adult Mouse Anatomical Dictionary and the Foundational Model of Anatomy ontology (FMA) as targets. This compares favorably with the best performing systems on the Ontology Alignment Evaluation Initiative anatomy challenge. We performed additional experiments on aligning FMA to NCI Thesaurus and to SNOMED CT based on a reference alignment extracted from the UMLS Metathesaurus. Our system obtained overall F-scores of 93.2% and 89.2% for these experiments, thus achieving state-of-the-art results.
Conclusions:
Our proposed representation learning approach leverages terminological embeddings to capture semantic similarity. Our results provide evidence that the approach produces embeddings that are especially well tailored to the ontology matching task, demonstrating a novel pathway for the problem.
Background:
The goal of ontology matching is to identify correspondences between entities from different yet overlapping ontologies so as to facilitate semantic integration, reuse and interoperability. As a well developed mathematical model for analyzing individuals and structuring concepts, Formal Concept Analysis (FCA) has been applied to ontology matching (OM) tasks since the beginning of OM research, whereas ontological knowledge exploited in FCA-based methods is limited. This motivates the study in this paper, i.e., to empower FCA with as much as ontological knowledge as possible for identifying mappings across ontologies.
Methods:
We propose a method based on Formal Concept Analysis to identify and validate mappings across ontologies, including one-to-one mappings, complex mappings and correspondences between object properties. Our method, called FCA-Map, incrementally generates a total of five types of formal contexts and extracts mappings from the lattices derived. First, the token-based formal context describes how class names, labels and synonyms share lexical tokens, leading to lexical mappings (anchors) across ontologies. Second, the relation-based formal context describes how classes are in taxonomic, partonomic and disjoint relationships with the anchors, leading to positive and negative structural evidence for validating the lexical matching. Third, the positive relation-based context can be used to discover structural mappings. Afterwards, the property-based formal context describes how object properties are used in axioms to connect anchor classes across ontologies, leading to property mappings. Last, the restriction-based formal context describes co-occurrence of classes across ontologies in anonymous ancestors of anchors, from which extended structural mappings and complex mappings can be identified.
Results:
Evaluation on the Anatomy, the Large Biomedical Ontologies, and the Disease and Phenotype track of the 2016 Ontology Alignment Evaluation Initiative campaign demonstrates the effectiveness of FCA-Map and its competitiveness with the top-ranked systems. FCA-Map can achieve a better balance between precision and recall for large-scale domain ontologies through constructing multiple FCA structures, whereas it performs unsatisfactorily for smaller-sized ontologies with less lexical and semantic expressions.
Conclusions:
Compared with other FCA-based OM systems, the study in this paper is more comprehensive as an attempt to push the envelope of the Formal Concept Analysis formalism in ontology matching tasks. Five types of formal contexts are constructed incrementally, and their derived concept lattices are used to cluster the commonalities among classes at lexical and structural level, respectively. Experiments on large, real-world domain ontologies show promising results and reveal the power of FCA.
With the proliferation of sensors, semantic web technologies are becoming closely related to sensor network. The linking of elements from semantic web technologies with sensor networks is called semantic sensor web whose main feature is the use of sensor ontologies. However, due to the subjectivity of different sensor ontology designer, different sensor ontologies may define the same entities with different names or in different ways, raising so-called sensor ontology heterogeneity problem. There are many application scenarios where solving the problem of semantic heterogeneity may have a big impact, and it is urgent to provide techniques to enable the processing, interpretation and sharing of data from sensor web whose information is organized into different ontological schemes. Although sensor ontology heterogeneity problem can be effectively solved by Evolutionary Algorithm (EA)-based ontology meta-matching technologies, the drawbacks of traditional EA, such as premature convergence and long runtime, seriously hamper them from being applied in the practical dynamic applications. To solve this problem, we propose a novel Compact Co-Evolutionary Algorithm (CCEA) to improve the ontology alignment’s quality and reduce the runtime consumption. In particular, CCEA works with one better probability vector (PV) \(PV_{better}\) and one worse PV \(PV_{worse}\), where \(PV_{better}\) mainly focuses on the exploitation which dedicates to increase the speed of the convergence and \(PV_{worse}\) pays more attention to the exploration which aims at preventing the premature convergence. In the experiment, we use Ontology Alignment Evaluation Initiative (OAEI) test cases and two pairs of real sensor ontologies to test the performance of our approach. The experimental results show that CCEA-based ontology matching approach is both effective and efficient when matching ontologies with various scales and under different heterogeneous situations, and compared with the state-of-the-art sensor ontology matching systems, CCEA-based ontology matching approach can significantly improve the ontology alignment’s quality.
Motivation:
The amount of information available in textual format is rapidly increasing in the biomedical domain. Therefore, natural language processing (NLP) applications are becoming increasingly important to facilitate the retrieval and analysis of these data. Computing the semantic similarity between sentences is an important component in many NLP tasks including text retrieval and summarization. A number of approaches have been proposed for semantic sentence similarity estimation for generic English. However, our experiments showed that such approaches do not effectively cover biomedical knowledge and produce poor results for biomedical text.
Methods:
We propose several approaches for sentence-level semantic similarity computation in the biomedical domain, including string similarity measures and measures based on the distributed vector representations of sentences learned in an unsupervised manner from a large biomedical corpus. In addition, ontology-based approaches are presented that utilize general and domain-specific ontologies. Finally, a supervised regression based model is developed that effectively combines the different similarity computation metrics. A benchmark data set consisting of 100 sentence pairs from the biomedical literature is manually annotated by five human experts and used for evaluating the proposed methods.
Results:
The experiments showed that the supervised semantic sentence similarity computation approach obtained the best performance (0.836 correlation with gold standard human annotations) and improved over the state-of-the-art domain-independent systems up to 42.6% in terms of the Pearson correlation metric.
Availability and implementation:
A web-based system for biomedical semantic sentence similarity computation, the source code, and the annotated benchmark data set are available at: http://tabilab.cmpe.boun.edu.tr/BIOSSES/ .
Contact:
gizemsogancioglu@gmail.com or arzucan.ozgur@boun.edu.tr.
This paper presents a novel background knowledge approach
which selects and combines existing mappings from a given biomedical
ontology repository to improve ontology alignment. Current background
knowledge approaches usually select either manually or automatically a
limited number of di�erent ontologies and use them as a whole for back-
ground knowledge. Whereas in our approach, we propose to pick up only
relevant concepts and relevant existing mappings linking these concepts
all together in a speci�c and customized background knowledge graph.
Paths within this graph will help to discover new mappings. We have im-
plemented and evaluated our approach using the content of the NCBO
BioPortal repository and the Anatomy benchmark from the Ontology
Alignment Evaluation Initiative. We used the mapping gain measure to
assess how much our �nal background knowledge graph improves results
of state-of-the-art alignment systems. Furthermore, the evaluation shows
that our approach produces a high quality alignment and discovers map-
pings that have not been found by state-of-the-art systems.
Semantic similarity measurement aims to determine the likeness between two text expressions that use different lexicographies for representing the same real object or idea. There are a lot of semantic similarity measures for addressing this problem. However, the best results have been achieved when aggregating a number of simple similarity measures. This means that after the various similarity values have been calculated, the overall similarity for a pair of text expressions is computed using an aggregation function of these individual semantic similarity values. This aggregation is often computed by means of statistical functions. In this work, we present CoTO (Consensus or Trade-Off) a solution based on fuzzy logic that is able to outperform these traditional approaches.
This paper proposes a novel approach which uses a multi-objective evolutionary algorithm based on decomposition to address the ontology alignment optimization problem. Comparing with the approach based on Genetic Algorithm (GA), our method can simultaneously optimize three goals (maximizing the alignment recall, the alignment precision and the f-measure). The experimental results shows that our approach is able to provide various alignments in one execution which are less biased to one of the evaluations of the alignment quality than GA approach, thus the quality of alignments are obviously better than or equal to those given by the approach based on GA which considers precision, recall and f-measure only, and other multi-objective evolutionary approach such as NSGA-II approach. In addition, the performance of our approach outperforms NSGA-II approach with the average improvement equal to 32.79 \(\%\) . Through the comparison of the quality of the alignments obtained by our approach with those by the state of the art ontology matching systems, we draw the conclusion that our approach is more effective and efficient.
We are currently facing a proliferation of heterogeneous biomedical data sources accessible through various knowledge-based applications. These data are annotated by increasingly extensive and widely disseminated knowledge organisation systems ranging from simple terminologies and structured vocabularies to formal ontologies. In order to solve the interoperability issue, which arises due to the heterogeneity of these ontologies, an alignment task is usually performed. However, while significant effort has been made to provide tools that automatically align small ontologies containing hundreds or thousands of entities, little attention has been paid to the matching of large sized ontologies in the life sciences domain.
We have designed and implemented ServOMap, an effective method for large scale ontology matching. It is a fast and efficient high precision system able to perform matching of input ontologies containing hundreds of thousands of entities. The system, which was included in the 2012 and 2013 editions of the Ontology Alignment Evaluation Initiative campaign, performed very well. It was ranked among the top systems for the large ontologies matching.
We proposed an approach for large scale ontology matching relying on Information Retrieval (IR) techniques and the combination of lexical and machine learning contextual similarity computing for the generation of candidate mappings. It is particularly adapted to the life sciences domain as many of the ontologies in this domain benefit from synonym terms taken from the Unified Medical Language System and that can be used by our IR strategy. The ServOMap system we implemented is able to deal with hundreds of thousands entities with an efficient computation time.
Fuzzy Logic Controllers are a specific model of Fuzzy Rule Based Systems suitable for engineering applications for which classic control strategies do not achieve good results or for when it is too difficult to obtain a mathematical model. Recently, the International Electrotechnical Commission has published a standard for fuzzy control programming in part 7 of the IEC 61131 norm in order to offer a well defined common understanding of the basic means with which to integrate fuzzy control applications in control systems. In this paper, we introduce an open source Java library called jFuzzyLogic which offers a fully functional and complete implementation of a fuzzy inference system according to this standard, providing a programming interface and Eclipse plugin to easily write and test code for fuzzy control applications. A case study is given to illustrate the use of jFuzzyLogic.
Nowadays, there are a lot of techniques and tools for addressing the ontology matching problem; however, the complex nature of this problem means that the existing solutions are unsatisfactory. This work intends to shed some light on a more flexible way of matching ontologies using ontology meta-matching. This emerging technique selects appropriate algorithms and their associated weights and thresholds in scenarios where accurate ontology matching is necessary. We think that an overview of the problem and an analysis of the existing state-of-the-art solutions will help researchers and practitioners to identify the most appropriate specific features and global strategies in order to build more accurate and dynamic systems following this paradigm.
Nowadays many techniques and tools are available for addressing the ontology matching problem, however, the complex nature
of this problem causes existing solutions to be unsatisfactory. This work aims to shed some light on a more flexible way of
matching ontologies. Ontology meta-matching, which is a set of techniques to configure optimum ontology matching functions.
In this sense, we propose two approaches to automatically solve the ontology meta-matching problem. The first one is called
maximum similarity measure, which is based on a greedy strategy to compute efficiently the parameters which configure a composite
matching algorithm. The second approach is called genetics for ontology alignments and is based on a genetic algorithm which
scales better for a large number of atomic matching algorithms in the composite algorithm and is able to optimize the results
of the matching process.
KeywordsOntology meta-matching–Knowledge management–Information integration
Born primarily as means to model knowledge, ontologies have successfully been exploited to enable knowledge exchange among people, organizations and software agents. However, because of strong subjectivity of ontology modeling, a matching process is necessary in order to lead ontologies into mutual agreement and obtain the relative alignment, i.e., the set of correspondences among them. The aim of this paper is to propose a memetic algorithm to perform an automatic matching process capable of computing a suboptimal alignment between two ontologies. To achieve this aim, the ontology alignment problem has been formulated as a minimum optimization problem characterized by an objective function depending on a fuzzy similarity. As shown in the performed experiments, the memetic approach results more suitable for ontology alignment problem than other evolutionary techniques such as genetic algorithms.
In this work we present GOAL (Genetics for Ontology Align- ments) a new approach to compute the optimal ontology alignment func- tion for a given ontology input set. Although this problem could be solved by an exhaustive search when the number of similarity measures is low, our method,is expected to scale better for a high number,of measures. Our approach is a genetic algorithm which is able to work with several goals: maximizing the alignment precision, maximizing the alignment re- call, maximizing the f-measure or reducing the number of false positives. Moreover, we test it here by combining some cutting-edge similarity mea- sures over a standard benchmark, and the results obtained show several advantages in relation to other techniques. Key words: ontology alignment; genetic algorithms; semantic integra-
With a growing number of ontologies and datasets using those ontologies, ontology mappings become an essential building block of the Semantic Web. In the last years, a larger number of sophisticated ontology matching tools for generating such mappings has been developed. The quality of the mappings provided by those tools typically depends on the settings of the tools' parameters. As this is a non-trivial task for an end user, we propose the ECOMatch approach, which asks the user to provide example mappings instead of parameter settings, and automatically determines a suitable parameter setting based on those examples. We show how the preliminary result quality of ontology mappings can be improved by applying automatic, example-based configuration of ontology matching tools.
In this paper, we present LogMap--a highly scalable ontology matching system with 'built-in' reasoning and diagnosis capabilities. To the best of our knowledge, LogMap is the only matching system that can deal with semantically rich ontologies containing tens (and even hundreds) of thousands of classes. In contrast to most existing tools, LogMap also implements algorithms for 'on the fly' unsatisfiability detection and repair. Our experiments with the ontologies NCI, FMA and SNOMED CT confirm that our system can efficiently match even the largest existing bio-medical ontologies. Furthermore, LogMap is able to produce a 'clean' set of output mappings in many cases, in the sense that the ontology obtained by integrating LogMap's output mappings with the input ontologies is consistent and does not contain unsatisfiable classes.
Bayesian networks (BNs) can capture interdependencies among ontology mapping methods and thus possibly improve the way they are combined. Experiments on ontologies from the OAEI collection are shown, and the possibility of modelling explicit mapping patterns in combination with methods is discussed. Most existing systems for ontology mapping combine various methods for achiev- ing higher performance in terms of recall and precision. Our approach relies on Bayesian networks (BNs) as well-known formal technique that can capture in- terdependencies among random variables. A Bayesian network (BN) (3) is a di- rected acyclic graph with attached local probability distributions. Nodes in the graph represent random variables with mutually exclusive and exhaustive sets of values (states). Edges in the graph represents direct interdependences between two random variables. We believe that this approach can bring additional ben- efits compared to ad hoc combination of methods, mainly resulting from better adaptability (training from data within a well-established formal framework). Two approaches that use BNs for Ontology Mapping have recently been re- ported. The first is OMEN (4), which mainly serves for enhancing existing map- pings. Its input are results of another mapping tool, while its output are more precise mappings as well as and new mappings. Nodes in the BN represent pairs of concepts that can potentially be mapped. Edges follow the taxonomy given in original ontologies. The network structure thus mimics that of ontologies them- selves, though heuristics for graph pruning are employed in this transformation. For constructing conditional probability tables (CPTs) for each node meta-rules are used, such as : "if two nodes match and so do two arrows coming out of these nodes then the probability that nodes at the other end of the arrows match is increased". The second project, BayesOWL ((5)), is rather a framework for on- tology mapping than a mapping method per se. The probabilistic ontological information is assumed to be learnt (in forms of probabilistic constraints) from web data using a text-classification-based learner; this information is translated to BNs. Mappings among concepts from two dierent ontologies then can be discovered using so-called evidential reasoning across two BNs.
Despite serious research efforts, automatic ontology matching still suffers from severe problems with respect to the quality of matching results. Existing matching systems trade-off precision and recall and have their specific strengths and weaknesses. This leads to problems when the right matcher for a given task has to be selected. In this paper, we present a method for improv- ing matching results by not choosing a specific matcher but applying machine learning techniques on an ensemble of matchers. Hereby we learn rules for the correctness of a correspondence based on the output of different matchers and additional information about the nature of the elements to be matched, thus lever- aging the weaknesses of an individual matcher. We show that our method always performs significantly better than the median of the matchers used and in most cases outperforms the best matcher with an optimal threshold for a given pair of ontologies. As a side product of our experiments, we discovered that the major- ity vote is a simple but powerful heuristic for combining matchers that almost reaches the quality of our learning results.
Most recent schema matching systems assemble multiple components, each employing a particular matching technique. The domain user mustthen tune the system: select the right component to be executed and correctly adjust their numerous “knobs” (e.g., thresholds, formula coefficients). Tuning is skill and time intensive, but (as we show) without it the matching accuracy is significantly inferior. We describe eTuner, an approach to automatically tune schema matching systems. Given a schema S, we match S against synthetic schemas, for which the ground truth mapping is known, and find a tuning that demonstrably improves the performance of matching S against real schemas. To efficiently search the huge space of tuning configurations, eTuner works sequentially, starting with tuning the lowest level components. To increase the applicability of eTuner, we develop methods to tune a broad range of matching components. While the tuning process is completely automatic, eTuner can also exploit user assistance (whenever available) to further improve the tuning quality. We employed eTuner to tune four recently developed matching systems on several real-world domains. The results show that eTuner produced tuned matching systems that achieve higher accuracy than using the systems with currently possible tuning methods.
Particle swarm optimisation (PSO) is a biologically-inspired, population-based optimisation technique that has been successfully applied to various problems in science and engineering. In the context of semantic technologies, optimisation problems also occur but have rarely been considered as such. This work addresses the problem of ontology alignment, which is the identification of overlaps in heterogeneous knowledge bases backing semantic applications. To this end, the ontology alignment problem is revisited as an optimisation problem. A discrete particle swarm optimisation algorithm is designed in order to solve this optimisation problem and compute an alignment of two ontologies. A number of characteristics of traditional PSO algorithms are partially relaxed in this article, such as fixed dimensionality of particles. A complex fitness function based on similarity measures of ontological entities, as well as a tailored particle update procedure are presented. This approach brings several benefits for solving the ontology alignment problem, such as inherent parallelisation, anytime behaviour, and flexibility according to the characteristics of particular ontologies. The presented algorithm has been implemented under the name MapPSO (ontology mapping using particle swarm optimisation). Experiments demonstrate that applying PSO in the context of ontology alignment is a feasible approach. Yes Yes
The Unified Medical Language System (http://umlsks.nlm.nih.gov) is a repository of biomedical vocabularies developed by the US National Library of Medicine. The UMLS integrates over 2
million names for some 900 000 concepts from more than 60 families of biomedical vocabularies, as well as 12 million relations
among these concepts. Vocabularies integrated in the UMLS Metathesaurus include the NCBI taxonomy, Gene Ontology, the Medical
Subject Headings (MeSH), OMIM and the Digital Anatomist Symbolic Knowledge Base. UMLS concepts are not only inter‐related,
but may also be linked to external resources such as GenBank. In addition to data, the UMLS includes tools for customizing
the Metathesaurus (MetamorphoSys), for generating lexical variants of concept names (lvg) and for extracting UMLS concepts
from text (MetaMap). The UMLS knowledge sources are updated quarterly. All vocabularies are available at no fee for research
purposes within an institution, but UMLS users are required to sign a license agreement. The UMLS knowledge sources are distributed
on CD‐ROM and by FTP.
Cancer researchers need to be able to organize and report their results in a way that others can find, build upon, and relate to the specific clinical conditions of individual patients. NCI Thesaurus is a description logic terminology based on current science that helps individuals and software applications connect and organize the results of cancer research, e.g., by disease and underlying biology. Currently containing some 34,000 concepts--covering chemicals, drugs and other therapies, diseases, genes and gene products, anatomy, organisms, animal models, techniques, biologic processes, and administrative categories--NCI Thesaurus serves applications and the Web from a terminology server. As a scalable, formal terminology, the deployed Thesaurus, and associated applications and interfaces, are a model for some of the standards required for the NHII (National Health Information Infrastructure) and the Semantic Web.
We have developed an ontology to provide standardized nomenclature for anatomical terms in the postnatal mouse. The Adult Mouse Anatomical Dictionary is structured as a directed acyclic graph, and is organized hierarchically both spatially and functionally. The ontology will be used to annotate and integrate different types of data pertinent to anatomy, such as gene expression patterns and phenotype information, which will contribute to an integrated description of biological phenomena in the mouse.
This paper reports on the alignment between mouse and human anatomies, a critical resource for comparative science as diseases in mice are used as mod-els of human disease. The two ontologies under investigation are the NCI Thesaurus (human anatomy) and the Adult Mouse Anatomical Dictionary, each comprising about 2500 anatomical concepts. This study compares two approaches to aligning ontologies. One is fully automatic, based on a combination of lexical and structural similarity; the other is manual. The resulting mappings were evaluated by an expert. 715 and 781 mappings were identified by each method respectively, of which 639 are common to both and all valid. The applications of the map-ping are discussed from the perspective of biology and from that of ontology.
Ontologies have emerged to establish a well-defined meaning for information, solving problems of heterogeneity in data semantics and facilitating the process of information exchange. However, ontologies have generated a new semantic problem, since using more than one ontology can generate ambiguity in the meaning of a given data. The problem of ontology matching is to search for relationships between entities of distinct ontologies, solving the problem of semantic heterogeneity of the data. The problem is a relevant issue in the area of knowledge representation and several approaches have been proposed to solve it. This work presents a systematic mapping of the main works published in the area with an emphasis on metaheuristics-based meta-matching approaches.
Multivariate adaptive regression splines (MARS) is a statistical modeling approach with wide real-world applications. In the MARS model building process, knot positioning is a critical step that potentially affects the accuracy of the final MARS model. Identifying well-positioned knots entails assessing the quality of many knots in each model building iteration, which requires intensive computational effort. By exploring the change in the residual sum of squares (RSS) within MARS, we find that local optima from previous iterations can be very close to those of the current iteration. In our approach, the prior change in RSS information is used to “warm start” an optimal knot positioning. We propose two methods for MARS knot positioning. The first method is a hill climbing method (HCM), which ignores prior change in RSS information. The second method is a hill climbing method using prior change in RSS information (PHCM). Numerical experiments are conducted on data with up to 30 dimensions. Our results show that both versions of hill climbing methods outperform a standard MARS knot selection method on datasets with different noise levels. Further, PHCM using prior change in RSS information performs best in both accuracy and computational speed. In addition, an open source Python code will be available upon acceptance of the paper on GitHub (https://github.com/JuXinglong/MARSHC).
Interpretability has been always present in Machine Learning and Artificial Intelligence. However, it is difficult to measure it (even to define it), and quite commonly it collides with other properties as accuracy, with a clear meaning and well defined metrics. This situation has reduced its influence in the area. But due to different external reasons, interpretability is now gaining importance in Artificial Intelligence, and particularly in Machine Learning. This new situation has two effects on the field of fuzzy systems. First, considering the capability of the fuzzy formalism to describe complex phenomena in terms that are quite close to human language, fuzzy systems have gained significant presence as an interpretable modeling tool. Second, the attention paid to interpretability of fuzzy systems, that grew during the first decade of this century and then experienced a certain decay, is growing again. The present paper will consider four questions regarding interpretability: what is, why is it important, how to measure it, and how to achieve it. These questions will be first introduced in the general framework of Artificial Intelligence, to be then focused from the point of view of fuzzy systems.
Currently, the semantic analysis is used by different fields, such as information retrieval, the biomedical domain, and natural language processing. The primary focus of this research work is on using semantic methods, the cosine similarity algorithm, and fuzzy logic to improve the matching of documents. The algorithms were applied to plain texts in this case CVs (resumes) and job descriptions. Synsets of WordNet were used to enrich the semantic similarity methods such as the Wu-Palmer Similarity (WUP), Leacock-Chodorow similarity (LCH), and path similarity (hypernym/hyponym). Additionally, keyword extraction was used to create a postings list where keywords were weighted. The task of recruiting new personnel in the companies that publish job descriptions and reciprocally finding a company when workers publish their resumes is discussed in this research work. The creation of a new gold standard was required to achieve a comparison of the proposed methods. A web application was designed to match the documents manually, creating the new gold standard. Thereby the new gold standard confirming benefits of enriching the cosine algorithm semantically. Finally, the results were compared with the new gold standard to check the efficiency of the new methods proposed. The measures used for the analysis were precision, recall, and f-measure, concluding that the cosine similarity weighted semantically can be used to get better similarity scores.
Ontology alignment is a fundamental task to support information sharing and reuse in heterogeneous information systems. Optimizing the combination of matchers through evolutionary algorithms to align ontology is an effective method. However, such methods have two significant shortcomings: weights need to be set manually to combine matchers, and a reference alignment is required during the optimization process. In this paper, a meta-matching approach GSOOM for automatically configuring weights and threshold using grasshopper optimization algorithm (GOA) has been proposed. In this approach, the ontology alignment problem is modeled as optimizing individual fitness of GOA. A fitness function is proposed, which includes two goals: maximizing the number of matching and the average similarity score. Since it does not require an expert to provide a reference alignment, it is more suitable for real-world scenarios. To demonstrate the advantages of the approach, we conduct exhaustive experiments tasks on several standard datasets and compare its performance to other state-of-the-other methods. The experimental results illustrate that our approach is more efficiently and is significantly superior to other metaheuristic-based methods.
This paper proposes an ontology-based framework to support integration and visualization of data from Intelligent Transportation Systems. These activities may be technically demanding for transportation stakeholders, due to technical and human factors, and may hinder the use of visualization tools in practice. The existing ontologies do not provide the necessary semantics for integration of spatio-temporal data from such systems. Moreover, a formal representation of the components of visualization techniques and expert knowledge can leverage the development of visualization tools that facilitate data analysis. The proposed Visualization-oriented Urban Mobility Ontology (VUMO) provides a semantic foundation to knowledge-assisted visualization tools (KVTs). VUMO contains three facets that interrelate the characteristics of spatio-temporal mobility data, visualization techniques and expert knowledge. A built-in rule set leverages semantic technologies standards to infer which visualization techniques are compatible with analytical tasks, and to discover implicit relationships within integrated data. The annotation of expert knowledge encodes qualitative and quantitative feedback from domain experts that can be exploited by recommendation methods to automate part of the visualization workflow. Data from the city of Porto, Portugal were used to demonstrate practical applications of the ontology for each facet. As a foundational domain ontology, VUMO can be extended to meet the distinctiveness of a KVT.
In order to perform ontology matching with high accuracy, while at the same time retaining applicability to most diverse input ontologies, the matching process generally incorporates multiple methods. Each of these methods is aimed at a particular ontology component, such as annotations, structure, properties or instances. Adequately combining these methods is one of the greatest challenges in designing an ontology matching system. In a parallel composition of basic matchers, the ability to dynamically set the weights of the basic matchers in the final output, thus making the weights optimal for the given input, is the key breakthrough for obtaining first-rate matching performance. In this paper we present CroMatcher, an ontology matching system, introducing several novelties to the automated weight calculation process. We apply substitute values for matchers that are inapplicable for the particular case and use thresholds to eliminate low-probability alignment candidates. We compare the alignments produced by the matchers and give less weight to the matchers producing mutually similar alignments, whereas more weight is given to those matchers whose alignment is distinct and rather unique. We also present a new, iterative method for producing one-to-one final alignment of ontology structures, which is a significant enhancement of similar non-iterative methods proposed in the literature. CroMatcher has been evaluated against other state-of-the-art matching systems at the OAEI evaluation contest. In a large number of test cases it achieved the highest score, which puts it among the state-of-the-art leaders.
To bring the Life Sciences domain closer to a Semantic Web realization it is fundamental to establish meaningful relations between biomedical ontologies. The successful application of ontology matching techniques is strongly tied to an effective exploration of the complex and diverse biomedical terminology contained in biomedical ontologies. In this paper, we present an overview of the lexical components of several biomedical ontologies and investigate how different approaches for their use can impact the performance of ontology matching techniques. We propose novel approaches for exploring the different types of synonyms encoded by the ontologies and for extending them based both on internal synonym derivation and on external ontologies.
We evaluate our approaches using AgreementMaker, a successful ontology matching platform that implements several lexical matchers, and apply them to a set of four benchmark biomedical ontology matching tasks. Our results demonstrate the impact that an adequate consideration of ontology synonyms can have on matching performance, and validate our novel approach for combining internal and external synonym sources as a competitive and in many cases improved solution for biomedical ontology matching.
There are three main drawbacks of current evolutionary approaches for determining the weights of ontology matching system. The first drawback is that it is difficult to simultaneously deal with several pairs of ontologies, i.e. finding a universal weight configuration that can be used for different ontology pairs without adjustment. The second one is that a reference alignment between two ontologies to be aligned should be given in advance which could be very expensive to obtain especially when the scale of ontologies is considerably large. The last one arises from f-measure, a generally used evaluation metric of the alignment's quality, which may cause the bias improvement of the solution. To overcome these three defects, in this paper, we propose to use both MatchFmeasure, a rough evaluation metric on no reference alignment to approximate f-measure, and Unanimous Improvement Ratio (UIR), a measure that complements MatchFmeasure, in the process of optimizing the ontology alignments by Memetic Algorithm (MA). The experimental results have shown that the MA using both MatchFmeasure and UIR is effective to simultaneously align multiple pairs of ontologies and avoid the bias improvement caused by MatchFeasure. Moreover, the comparison with state-of-the-art ontology matching systems further indicates the effectiveness of the proposed method.
AgreementMaker is one of the leading ontology matching systems, thanks to its combination of a flexible and extensible framework with a comprehensive user interface. In many domains, such as the biomedical, ontologies are becoming increasingly large thus presenting new challenges. We have developed a new core framework, AgreementMakerLight, focused on computational efficiency and designed to handle very large ontologies, while preserving most of the flexibility and extensibility of the original AgreementMaker framework. We evaluated the efficiency of AgreementMakerLight in two OAEI tracks: Anatomy and Large Biomedical Ontologies, obtaining excellent run time results. In addition, for the Anatomy track, AgreementMakerLight is now the best system as measured in terms of F-measure. Also in terms of F-measure, AgreementMakerLight is competitive with the best OAEI performers in two of the three tasks of the Large Biomedical Ontologies track that match whole ontologies.
All the state of the art approaches based on evolutionary algorithm (EA) for addressing the meta-matching problem in ontology alignment require the domain expert to provide a reference alignment (RA) between two ontologies in advance. Since the RA is very expensive to obtain especially when the scale of ontology is very large, in this paper, we propose to use the Partial Reference Alignment (PRA) built by clustering-based approach to take the place of RA in the process of using evolutionary approach. Then a problem-specific Memetic Algorithm (MA) is proposed to address the meta-matching problem by optimizing the aggregation of three different basic similarity measures (Syntactic Measure, Linguistic Measure and Taxonomy based Measure) into a single similarity metric. The experimental results have shown that using PRA constructed by our approach in most cases leads to higher quality of solution than using PRA built in randomly selecting classes from ontology and the quality of solution is very close to the approach using RA where the precision value of solution is generally high. Comparing to the state of the art ontology matching systems, our approach is able to obtain more accurate results. Moreover, our approach’s performance is better than GOAL approach based on Genetic Algorithm (GA) and RA with the average improvement up to 50.61%. Therefore, the proposed approach is both effective.
A mathematical tool to build a fuzzy model of a system where fuzzy implications and reasoning are used is presented. The premise of an implication is the description of fuzzy subspace of inputs and its consequence is a linear input-output relation. The method of identification of a system using its input-output data is then shown. Two applications of the method to industrial processes are also discussed: a water cleaning process and a converter in a steel-making process.
The need for trading off interpretability and accuracy is intrinsic to the use of fuzzy systems. The obtaining of accurate but also human-comprehensible fuzzy systems played a key role in Zadeh and Mamdani’s seminal ideas and system identification methodologies. Nevertheless, before the advent of soft computing, accuracy progressively became the main concern of fuzzy model builders, making the resulting fuzzy systems get closer to black-box models such as neural networks. Fortunately, the fuzzy modeling scientific community has come back to its origins by considering design techniques dealing with the interpretability-accuracy tradeoff. In particular, the use of genetic fuzzy systems has been widely extended thanks to their inherent flexibility and their capability to jointly consider different optimization criteria. The current contribution constitutes a review on the most representative genetic fuzzy systems relying on Mamdani-type fuzzy rule-based systems to obtain interpretable linguistic fuzzy models with a good accuracy.
One of the main threads in the history of knowledge-representation formalisms is the trade-off between the expressiveness of first-order logic on the one hand and the tractability and ease-of-use of frame-based systems on the other hand. Frame-based systems provide intuitive, cognitively easy-to-understand, and scalable means for modeling a domain. However, when a domain model is particularly complex, frame-based representation may lead to complicated and sometimes awkward solutions. We have encountered such problems when developing the Digital Anatomist Foundational Model, an ontology aimed at representing comprehensively the physical organization of the human body. We show that traditional frame-based techniques such as is-a hierarchies, slots (roles) and role restrictions are not sufficient for a comprehensive model of this domain. The diverse modeling challenges and problems in this project required us to use such knowledge-representation techniques as reified relations, metaclasses and a metaclass hierarchy, different propagation patterns for template and own slots, and so on. We posit that even though the modeling structure imposed by frame-based systems may sometimes lead to complicated solutions, it is still worthwhile to use frame-based representation for very large-scale projects such as this one.
In this paper a genetic algorithm-based optimization procedure for ontology matching problem is presented as a feature-matching process. First, from a global view, we model the problem of ontology matching as an optimization problem of a mapping between two compared ontologies, and every ontology has its associated feature sets. Second, as a powerful heuristic search strategy, genetic algorithm is employed for the ontology matching problem. Given a certain mapping as optimizing object for GA, fitness function is defined as a global similarity measure function between two ontologies based on feature sets. Finally, a set of experiments are conducted to analysis and evaluate the performance of GA in solving ontology matching problem.
A clinical terminology is essential for Electronic Health records. It represents clinical information input into clinical IT systems by clinicians in a machine-readable manner. Use of a Clinical Terminology, implemented within a clinical information system, will enable the delivery of many patient health benefits including electronic clinical decision support, disease screening and enhanced patient safety. For example, it will help reduce medication-prescribing errors, which are currently known to kill or injure many citizens. It will also reduce clinical administration effort and the overall costs of healthcare.
Multi-objective evolutionary algorithms (MOEAs) that use
non-dominated sorting and sharing have been criticized mainly for: (1)
their O(MN<sup>3</sup>) computational complexity (where M is the number
of objectives and N is the population size); (2) their non-elitism
approach; and (3) the need to specify a sharing parameter. In this
paper, we suggest a non-dominated sorting-based MOEA, called NSGA-II
(Non-dominated Sorting Genetic Algorithm II), which alleviates all of
the above three difficulties. Specifically, a fast non-dominated sorting
approach with O(MN<sup>2</sup>) computational complexity is presented.
Also, a selection operator is presented that creates a mating pool by
combining the parent and offspring populations and selecting the best N
solutions (with respect to fitness and spread). Simulation results on
difficult test problems show that NSGA-II is able, for most problems, to
find a much better spread of solutions and better convergence near the
true Pareto-optimal front compared to the Pareto-archived evolution
strategy and the strength-Pareto evolutionary algorithm - two other
elitist MOEAs that pay special attention to creating a diverse
Pareto-optimal front. Moreover, we modify the definition of dominance in
order to solve constrained multi-objective problems efficiently.
Simulation results of the constrained NSGA-II on a number of test
problems, including a five-objective, seven-constraint nonlinear
problem, are compared with another constrained multi-objective
optimizer, and the much better performance of NSGA-II is observed
POMap++ results for OAEI 2019: Fully automated machine learning approach for ontology matching
- A Laadhar
- F Ghozzi
- I Megdiche
- F Ravat
- O Teste
- F Gargouri
Laadhar, A., Ghozzi, F., Megdiche, I., Ravat, F., Teste, O., & Gargouri, F. (2019).
POMap++ results for OAEI 2019: Fully automated machine learning approach for
ontology matching. In P. Shvaiko, J. Euzenat, E. Jiménez-Ruiz, O. Hassanzadeh, &
C. Trojahn (Eds.), CEUR workshop proceedings: Vol. 2536, Proceedings of the 14th
international workshop on ontology matching co-located with the 18th international
semantic web conference (pp. 169-174). CEUR-WS.org, URL http://ceur-ws.org/Vol-2536/oaei19_paper13.pdf.