Social vote recommendation: building party models using the probability to vote feedback of VAA users

Abstract

Voting Advice Applications (VAAs) are online tools that match the policy preferences of voters' with the policy positions of political parties or candidates. A recent, innovative extension of VAAs has been to draw on the field of computer science to introduce a social vote recommendation borrowing the basic principles of collaborative filtering. The latter takes advantage of the community of VAA users to provide a vote recommendation. This paper presents a comparative study of social vote recommendation approaches that are based on machine learning. We build party models by utilizing both categorical variables, i.e., Voting intention and ordinal variables, i.e., Probability to vote for each one of the competing parties. The latter were first introduced in a practical VAA during the federal election in Germany in September 2013. The dataset from this election, consisting of more than 150.000 users, was used in our experiments.