Provided by: RWTH Aachen University
Topic: Data Management
In this paper, the authors describe their contribution to the 2015 Linked Data Mining Challenge. The proposed task is concerned with the prediction of review of movies as "Good" or "Bad", as does Meta-critic website based on critics' reviews. First they describe the sources used to build the training data. Although, several sources provide data about movies on the Web in different formats including RDF, data from HTML pages had to be gathered to fulfill some of their features. They then describe their experiment training a decision tree model on 241 features derived from their RDF (Resource Description Framework) knowledge base, achieving an accuracy of 0.94.