Analyzing Social Activity using Number of Agents
Prof. Dr. Kenichi Yoshida (Tsukuba University, Japan)
Abstract: Rapid evolution of network and computer technology realizes year of big data. High speed networks help in collecting information of people behavior and social activities. Huge
computing resources in cloud services analyze collected information. In such era, it is possible to analyze various social activities by simply counting number of agents, i.e., peoples, who are got involved in the activities. In this talk, we’d like to introduce examples of studies which show the efficacy of simple counting to analyze social activities. First example is analysis of TV viewers. In , we have proposed a method for creating the “Attention Graph” which depicts the amount of viewers’ attention generated by TV drama. The Attention Graph, is generated by counting the number of dialogues described in Internet communities concerning TV drama. It assists in specifying noteworthy zones from complete TV programs. In other words, it can be used to analyze viewers reaction on TV drama.
Since the analysis of people reactions toward TV and other form of advertisements, such as WWW advertisements, is important for marketing, related researches are studied
enthusiastically. As an example, we have extended  so that the method has tolerance against stealth marketing . A simple counting of DNS access log realizes the tolerance
against stealth marketing. Above examples are clear examples of analysis on social
activities. We also found an aspect of analyzing social activities in finding spams and internet viruses. Since most of spams and viruses have economic purpose, simple counting agents,
i.e., people, who have relation with them is a good way to find them. Mass unsolicited electronic mail, often known as spam, is a serious threat not only to the Internet but also to
the society. We proposed a spam detection method which just count number of similar mails . Since most spammers wish to spread advertisements, number of receivers who receive similar mail is a good index to find spams. Internet viruses and DDoS also have economic purpose, and have aspect of social activity. Thus we can find them by simply counting number of computers which send/receive packets . Although spams and viruses are studied in computer science area as engineering problems,  and  show the fact that the analysis of spam and virus as social activities has promising way to find these network threats. Last example of this talk is prediction of short term stock returns. Although efficient market hypothesis is widely accepted in financial market studies and entails the unpredictability of future stock prices,  shows that a simple analysis can classify short-term stock price changes with an 82.9% accuracy. This method essentially uses number of buyers and sellers. Although all these studies only rely on simple counting, they show the efficacy of simple counting of agents to analyze social activities.
 H. Uehara and K. Yoshida, “Annotating tv drama based on viewer dialogue-analysis of viewers’ attention generated on an internet bulletin board,” in The 2005 Symposium on Applications and the Internet. IEEE, 2005, pp. 334–340.
 T. Mitamura and K. Yoshida, “Viewers’ side analysis of social interests,” in 2012 IEEE 12th International Conference on Data Mining Workshops. IEEE, 2012, pp. 301–308.
 K. Yoshida, F. Adachi, T. Washio, H. Motoda, T. Homma, A. Nakashima, H. Fujikawa, and K. Yamazaki, “Density-based spam detector,” IEICE transactions on information and systems, vol. 87, no. 12, pp. 2678–2688, 2004.
 S. Mori and K. Sato, Akira annd Yoshida, “Enhancing performance of cardinality analysis for faster network,” in IEICE technical report, vol. 115, no. 307. IEICE, 2015, pp. 81–85.
 K. Yoshida and A. Sakurai, “Short-term stock price analysis based on order book information,” Information and Media Technologies, vol. 10, no. 4, pp. 521–530, 2015.
Biography: Kenichi YOSHIDA received his Ph.D. from Osaka University
in 1992. In 1980, he joined Hitachi Ltd., and is working for
University of Tsukuba from 2002. His current research interest
includes application of Internet and application of machine
learning techniques. Most works introduced in this talk are
Ph.D. studies of his students.
Concept of discard-after-learn approach and its side effects
Prof. Dr. Chidchanok Lursinsap (Chulalongkorn University, Thailand)
Abstract: In learning and clustering problems, several side effects possibly reducing the accuracy of the results cannot be ignored. Among these side effects are imbalance data, overflow of data in streaming situation, uncontrollable space and time complexities of learning and clustering processes. However, this side effects can be rather effectively resolved by a concept of discard-after-learn with the help of recursive functions.
Biography: Professor Chidchanok Lursinsap received the B.Eng. degree (honors) in computer engineering from Chulalongkorn University, Bangkok, Patumwan, Thailand, in 1978 and the M.S. and Ph.D. degrees in computer science from the University of Illinois at Urbana-Champaign, Urbana, in 1982 and 1986, respectively. He was a Lecturer at the Department of Computer Engineering, Chulalongkorn University, in 1979. In 1986, he was a Visiting Assistant Professor at the Department of Computer Science, University of Illinois at Urbana-Champaign. From 1987 to 1996, he worked at The Center for Advanced Computer Studies, University of Louisiana at Lafayette, as an Assistant and Associate Professor. After that, he came back to Thailand to establish Ph.D. program in computer science at Chulalongkorn University and became a Full Professor. His major research interests include neural learning and its applications to other science and engineering areas.