User preference prediction between ads-supported and subscribed users
Abstract: The goal of this master’s thesis was to create a model that predicts preference towards a specific exclusive feature in a subscribed service. It investigated unsupervised and semi-supervised learning to identify customer segments that prefer an specific exclusive feature. These customers segments were then used as targets for supervised learning algorithms to predict which segment a user on the ads-supported version would belong to. Two experiments was preformed, one to investigate and identify customer segments with the help of a survey and secondly, the preference prediction. It was found that Ward’s agglomerative clustering agreed the best with the preference analysis from the survey. Nevertheless, the correlation between the preference survey and the usage clustering was weak. The random forest classifier was preformed the best on the resulting dataset from Ward’s agglomerative clustering. It was concluded that user usage segmentation for the exclusive features showed promising results as well as the over all method. Nevertheless, due to the weak correlation between the survey and the usage clustering it rather predicts usage than preference.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)