Alvin Chin
I'm researching into mobile social networking, online communities, human-computer interfaces to support social networking, context-aware computing and pervasive computing.
Address: No.5 Donghuan Zhonglu,
Economic and Technological Development Area
Beijing, 100176, China
Address: No.5 Donghuan Zhonglu,
Economic and Technological Development Area
Beijing, 100176, China
less
InterestsView All (25)
Uploads
Papers by Alvin Chin
networks such as Facebook and LinkedIn. These friend
recommendations are based usually on common friends or similar
profile such as having the same interest or coming from the same
company, a trait known as homophily. However, many times
people do not know why they should add this friend. Should I add
this friend because we met from a conference and if so, what
conference? Existing friend recommendation systems cannot
answer this question easily. In this paper, we create a friend
recommendation system using proximity and homophily, that we
conduct in the workplace and conference. Besides common
friends and common interests (homophily features), we also
include encounters and meetings (proximity features) and
messages sent and question and answer posts (social interaction
features) as reasons for adding this person as a friend. We conduct
a user study to examine whether our friend recommendation is
better than common friends. Results show that on average, our
algorithm recommends more friends to participants that they add
and more recommendations are ranked as good, compared with
the common friend algorithm. In addition, people add friends due
to having encountered them before in real life. The results can be
used to help design context-aware recommendations in physical
environments and in online social networks.
based mobile application that recommends the nearby coupons and deals
for users, by which users can also follow the shops they are interested
in. If the potential followers of a shop can be discovered, the merchant’s
targeted advertising can be more effective and the recommendations
for users will also be improved. In this paper, we propose to predict the
link relations between users and shops based on the following behavior.
In order to better model the characteristics of the shops, we first adopt
Topic Modeling to analyze the semantics of their descriptions and then
propose a novel approach, named INtent Induced Topic Search (INITS)
to update the hidden topics of the shops with and without a description.
In addition, we leverage the user logs and search engine results to get
the similarity between users and shops. Then we adopt the latent factor
model to calculate the similarity between users and shops, in which we
use the multiple information sources to regularize the factorization. The
experimental results demonstrate that the proposed approach is effective
for detecting followers of the shops and the INITS model is useful
for shop topic inference.
world is an important way for us to make new friends
and build social networks. We aim to explore the role of
event size and interactivity in affecting social
networking behaviors. In this paper, we obtained data
from an event-based social network site and conducted
a quantitative analysis that reveals a relationship
between online following behavior and characteristics of
real-world events. We also employ behavior setting
theory, social role theory, and user interview data to
help us understand the quantitative results. Our finding
that small events on average promote more new
connections between individuals than large events has
important implications for event organizers, event
participants, and social media designers.
social networks, which are temporarily built based on events such as conferences. From the data distribution and social theory perspectives, we
found several interesting patterns. For example, the duration of two random
persons staying at the same place and at the same time obeys a
two-stage power-law distribution. Ephemeral social networks represent
more elite social activities: elite users tend to meet together and ordinary
users are also inclined to meet elite users. We develop a framework
to infer the likelihood of two users to meet together, and we apply the
framework to two mobile social networks: UbiComp and Reality. The
former is formed by researchers attending UbiComp 2011 and the latter
is a network of students published by MIT. On both networks, we validate
the proposed predictive framework, which significantly improve the
accuracy for predicting geographic coincidence by comparing with two
baseline methods.
fitted exponents for user activities distribution are slowly becoming larger over time, which appears to be contrary to the famous “rich get richer” assertion in the preferential attachment model because users in Friend View regard the
reciprocity as important during the interaction and (2) both undirected friend network and directed comment network in Friend View are small-world and scale-free networks over time with slowly decreasing clustering coefficient.
However, compared to online social networks where users have a large number of friends but loose weakly-tied subgroups, users in Friend View tend to have close strongly-tied cohesive subgroups. The results can help us understand users’ social activities and interactions over time in mobile social networks.