Ontology Alignment Using Machine Learning Techniques

Abstract

In the semantic web, ontology plays an important role to provide formal definitions of concepts and relationships. Therefore, communicating similar ontologies becomes essential to provide ontologies interpretability and extendibility. Thus, it is inevitable to have similar but not the same ontologies in a particular domain since there might be several definitions for a given concept. This paper presents a method to combine similarity measures of different categories without having ontology instances or any user feedback in regard with alignment of two given ontologies. To align different ontologies efficiently, K Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Tree (DT) and AdaBoost classifiers are investigated. Each classifier is optimized based on the lower cost and better classification rate. Experimental results demonstrate that the F-measure criterion improves up to 99% using feature selection and combination of AdaBoost and DT classifiers, which is highly comparable, and outperforms the previous reported F-measures.

The Impact of Imbalanced Training Data on Local Matching Learning of Ontologies

“…There has been some relevant work dealing with supervised matching learning [10,19,20,6]. Machine learning approaches for ontology alignment usually follow two phases [7]:…”

“…Malform-SVM [10] constructed a matching learning classifier from the reference alignments through a set of element level and structural level features. Nezhadi et al [19] presented a machine learning approach to aggregate different types of string similarity measures. The latter approach is evaluated through a relatively small bibliographic matching track provided by the OAEI benchmark.…”

A Machine Learning Approach to Multilingual and Cross-Lingual Ontology Matching

“…In contrast to the predominant view of ontology matching as a classification problem (as e.g. in [4,9]), we understand it as a ranking problem, similar to relevance ranking in information retrieval. In particular, we describe a novel approach that trains a ranking support vector machine (see [5]) on relative preference constraints between a concept in a source ontology and all possible concepts in a target ontology, with the goal of ranking good matches higher than bad matches.…”

“…There have been a number of machine learning approaches to ontology matching, such as the ones by Ichise [4] and Nezhadi et al [9]. In particular, Ichise also follows an svm-based approach, and Nezhadi et al evaluate different learned classifiers.…”

“…Therefore, we assign a target value of 3 to exact matches, 2 to broad and narrow matches, and 1 to all other combinations. 9 Given the features and similarity vectors as presented above, we can apply the ranking svm algorithm described by Joachims [5], which produces an svm model predicting scores for each input similarity vector. The matches can then be ranked such that those for which the model predicts higher scores are ranked above those with lower scores.…”