Print Email Facebook Twitter Prediction of liking of functions based on the properties of features on social networking sites Title Prediction of liking of functions based on the properties of features on social networking sites Author Schüsler, O.M. Contributor Houben, G.J.P.M. (mentor) Faculty Electrical Engineering, Mathematics and Computer Science Department Computer Science Programme Web Information Systems Date 2012-08-30 Abstract When developing a website, feedback from the community is imperative. This is not different for Social Networking Sites (SNS). There is no way however to measure how well the functionality is liked. During this research, a new method- ology is developed that will be able to predict the liking of functions based on features of the site. This will hopefully enable the creators of SNSs to better their sites to increase the number of visitors. For this, first the functions have been de- fined for the 3 current leading SNSs - Facebook, Google+ and LinkedIn. For this list of functions, lists of features were generated by users. With a question- naire, a dataset is generated, from which the relation between the features and functions can be derived by machine learning. In the end, three achievements are gained: better understanding of functions, features and the relationship between each other, that are common to these three SNSs, predictors, which takes ratings from a questionnaire as input, that predict how much a user likes a function, and a dataset containing ratings from 125 users. Subject Social Networking SitesLikingFunctionsPredictor To reference this document use: http://resolver.tudelft.nl/uuid:44087018-fbe5-4614-beb2-df5ed4482a18 Embargo date 2012-07-31 Part of collection Student theses Document type master thesis Rights (c) 2012 Schüsler, O.M. Files PDF Scriptie_Olaf_Schusler.pdf 400.62 KB Close viewer /islandora/object/uuid:44087018-fbe5-4614-beb2-df5ed4482a18/datastream/OBJ/view