Exploring Social Patterns in Mobile Data

• Published in: 2010 Second International Conference on Advances in Databases, Knowledge, and Data Applications pp. 62 - 68
• Main Authors: Kasarapu, Parthan, Saravanan, M, Garigipati, Prasad
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.04.2010
• Subjects: Advertising; Algorithm design and analysis; Clustering algorithms; Customer segmentation; Gibbs' sampling; Internet; Mobile handsets; Performance analysis; Sampling methods; Semantic Community Discovery; Social Patterns; Telecommunications; Web pages
• ISBN: 9781424460816; 1424460816
• DOI: 10.1109/DBKDA.2010.37

To compete with other telecom providers, it is important to understand the behavior of the customers and predict their needs. In order to realize this, it is required to explore the customers details based on their mobile usage behavior into social patterns (segments) and target the suitable segments for advertising. In our approach, the usage data of the customers in association with their browsing behavior is used to form the segments considered to be an important addition. From the analysis of their usage rates with respect to a certain domain, the operator can drill down to the sub domain level interests and target them with specific customized services. This can be done by performing latent semantic analysis using Gibbs sampling algorithm and K-Means clustering on the description of their accessed web pages with their usage and spend data. The traditional method involves forming web communities using link based approach. Our method based on identifying social communities could produce an alternative approach for the mobile operators. The usage rates within a certain cluster, and the customers' interest towards a specific domain can help to determine their extent of willingness to spend in specific areas. Our approach produces better results than the traditional methods by enabling the telecom providers to target a specific group of consumers.