Please use this identifier to cite or link to this item:
|Title:||On the quantification of missing value impact on voting advice applications||Authors:||Agathokleous, Marilena
|Keywords:||Missing values;Classifiers;Recommender systems;Voting advice applications||Category:||Electrical Engineering - Electronic Engineering - Information Engineering||Field:||Engineering and Technology||Issue Date:||2013||Publisher:||Springer Berlin Heidelberg||Source:||14th International Conference Engineering Applications of Neural Networks, 2013, Halkidiki, Greece, 13-16 September||Abstract:||Voting Advice Application (VAA) is a web application that recommends a candidate or a party to a voter. From an online questionnaire, which voters and candidates are called to answer, the VAA proposes to each individual voter the candidate who replied like him/her. It is very important the voters to reply in all questions of the questionnaire, because every question has its meaning and is responding to the political position of a each party. Missing values might mislead the VAA and impede it to have complete knowledge about the voter, as a result to offer him/her the wrong candidate. In this paper we quantitatively investigate the effect of missing values in VAAs by examining the impact of the number of missing values to different methods of voting prediction. For our experiment we have used the data obtained from the May parliamentary elections in Greece in 2012. The corresponding dataset is made freely available to other researchers working in the areas of VAA and recommender systems through the Web.||URI:||http://ktisis.cut.ac.cy/handle/10488/4038||DOI:||10.1007/978-3-642-41013-0_51||Rights:||Springer-Verlag Berlin Heidelberg||Type:||Conference Papers|
|Appears in Collections:||Δημοσιεύσεις σε συνέδρια/Conference papers|
Show full item record
checked on Nov 19, 2017
checked on Nov 23, 2017
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.