Tesi di laurea magistrale
Data Mining Classification Models to Predict Premium Users of an Online Social Network
Corso di studi
INFORMATICA PER L'ECONOMIA E PER L'AZIENDA (BUSINESS INFORMATICS)
relatore Prof. Ruggieri, Salvatore
- Machine Learning
- Support Vector Machine
- Decision Tree
- Data mining
- Social Network
Data inizio appello
Data di rilascio
The introduction of social media platforms has been fast and ubiquitous. The issues of different cultures in our society requires a new online platform to answer expatriates’ need of communication and their common questions in their life while they are living abroad. On the other hand, analyzing user’s behavior and interactions on online social network would help to address better their demands and requests. By utilizing classification and clustering models on users, we are able to scrutinize data with purpose of assessing their inclination to upgrade to a premium account. We predict membership of a social network company by ensemble algorithms (stacking and bagging algorithms). The classification models consist of K-Nearest Neighbors (K-NN), Support Vector Machine (SVM) and Decision Tree to take advantageous of combining models for ensuring maximum performance. The models detect high potential users who might want to upgrade their account to premium (paid) membership.
1 file non consultabili su richiesta dell'autore.