Personalized messages based on social networks
final year project, Sapientia - Hungarian University of Sapientia
2010, Targu Mures, Romania
This paper proposes a new approach for the dynamic generation of personalized, context-aware messages. The existence of smart phone devices that allow to include person specific information, such as birthday, location, creates interesting opportunities for various applications ranging from marketing to birthday wishes. We recognize the fact that people generally provide different formulations of messages (e.g. words used, level of granularity) based on the social relationship between communication partners. Our main objective is to provide end-users with an intelligent and adaptive birthday wishes system, where the cell phone composes SMS messages automatically in response to movement, with specific formulations adapted to multilevel social categories for the generation of rich birthday greetings.
This paper proposes a language-independent model by exploding the message into parts that are easy recognizable and are used in all languages, such as salutation, wishes, and signature. These message parts will adjust based on the social category of the communication partner. With these services you can easily turn your mobile device into a personal assistant on the go. We show promising preliminary experimentation results. Our contributions are both in the area of person classification, personalization, language modeling, profile data extraction from social networks and also in contextawareness like weather and places retrieval. We propose (i) a theoretical language model for the decomposition of messages into components (ii) an adaptive message generation system and a (iii) proof-of-concept implementation of the personal assistant with the Google Android Framework. An earlier version of this paper was published at the Scientific Conference of Students at Sapientia - EMTE, in year 2010.
This work is the result of my own activity. I have neither given nor received unauthorized assistance on this work.
Prof. Dr. Kása Zoltán