Output details
11 - Computer Science and Informatics
Robert Gordon University
Detection of News Feeds Items Appropriate for Children
Filtering inappropriate content has often been the first choice approach in terms of protecting children from harmful content online. In this work this negative stance is reversed and the opposite question is considered: can we actively find appropriate content for children. Besides the challenge of the user group, the secondary challenge addressed in this work is how to handle content in short form, primarily news feed snippets that have a short document length. The main contribution is a classifier based on a combination of SVM and classic readability filters, traditionally used for assessing the comprehensibility of literature for different levels of educational attainment.