Noisy Type Assertion Detection in Semantic Datasets
published: Dec. 19, 2014, recorded: October 2014, views: 1652
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Semantic datasets provide support to automate many tasks such as decision-making and question answering. However, their performance is always decreased by the noises in the datasets, among which, noisy type assertions play an important role. This problem has been mainly studied in the domain of data mining but not in the semantic web community. In this paper, we study the problem of noisy type assertion detection in semantic web datasets by making use of concept disjointness relationships hidden in the datasets. We transform noisy type assertion detection into multiclass classification of pairs of type assertions which type an individual to two potential disjoint concepts. The multiclass classification is solved by Adaboost with C4.5 as the base classifier. Furthermore, we propose instance-concept compatability metrics based on instance-instance relationships and instance-concept assertions. We evaluate the approach on both synthetic datasets and DBpedia. Our approach effectively detect noisy type assertions in DBpedia with a high precision of 95%.
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !