On Estimation Problems of Network Sampling Methods

• Activity Number: 136
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Ran Wei*+ and Tao Shi
• Companies: Ohio State University and Ohio State University
• Keywords: network sampling ; social network ; network topology ; attribute estimation ; estimation bias

Abstract

With the popularity of online social networks such as Facebook, Twitter, Weibo and LinkedIn, the scale of network data has become enormous and is growing all the time. Taking samples which are good representatives of the full network has become one major concern. Network sampling methods are also proved powerful in touching hard-to-reach populations, such as in the survey of HIV patients and drug users.

Different network sampling methods are studied in this paper, including random walk with or without replacement, snowball sampling and forest fire sampling with different burning probabilities. We compare these methods with simple random sampling in terms of sampling mechanism, estimation bias and variances in estimating parameters and attributes of networks. Reasons of bias and large variances are explored in this paper. We investigate the interplay among network structure, sampling methods and attribute distribution for both simulated data and real world social network data. We propose new estimation methods to correct bias and improve estimation performance under different circumstances.