Source Critique 3

Detailed Literature Review

Source 3, QuestionHolic: Hot topic discovery and trend analysis in community question answering systems

http://www.sciencedirect.com/science/article/pii/S0957417410014156 (Source) (PDF available upon request)

 

Trend Analysis

 

Purpose of the Source

The purpose of this source is to demonstrate how the use of trend analysis is utilized; in this particular case it has to do with online question forums and finding the trends of answers within them. It defines trend analysis as finding patterns over a period of time from collected historical data. This source also brings up a method they refer to as “hot topic detection” which is essentially trend spotting. In this method they are looking for consistencies and topics that continually surface. They then use this system to observe the changes in the trends at hand over time and present them utilizing trend graphs.

Strengths of the Technique

The purpose of the following list is to identify the strong suits of the method of trend analysis and why it is effective in predicting future events.

·      Data is already collected – you are left with combining it all to find commonalties among the data

·      Repetitive terms or factors are deemed as consistencies and thus trends

Weaknesses of the Technique

The purpose of the following list is to identify the weaknesses of the method of trend analysis.

·      Must cross wide range of sectors in order to be effective

·      When presenting statistically it is time consuming and hard to figure out.

·      Gradual trends are harder to detect by eye (when presenting via graphically)

Steps

-        Collecting data set characteristics for each category

-        Compiling the “hot terms” of the categories

-        Running data through algorithm

-        Putting data into visual presentation, in this case graph visuals

-        Determining what the trend is telling you, what are the commonalities

-        Considering parameters that could effect your trend

Purpose of the Technique

Trend Analysis is a good method to use when trying to predict something. This is done by compiling historical data on what is being followed and looking for trends and consistencies within that data. Hot Topic Detection (trend spotting) is actually finding those consistencies within the topic. From there one must compile and analyze the data at hand.

Comparison

This source is different from my other critiques in that it is an independent study utilizing the method of trend analysis. The authors utilize a formula that takes in account frequency of a term and then tracking the life cycle of that term thus coming up with the trend.

Most Informative

Although this source is talking about using trend analysis in monitoring trends in online forums, this technique can be applied to my analytical question (what are common trends associated with people engaging in espionage?). A key part of trend analysis is making connections within trends, for instance over time what is a common reason why people engage in espionage activities? This article goes on to provide their own system of trend spotting, which is essentially connecting commonalities found within a topic. In this study they refer to trend spotting as “Hot Topic Detection”.

About the Source Author

The authors of this article were Zhongfeng Zhang and Qiudan Li. Both work for the Key Laboratory of Complex System and Intelligence Science, Institute of Automation, Chinese Academy of Sciences.

Source Reliability

Your source reliability rating according to the Trust Evaluation Worksheet for online sources: http://daxrnorman.googlepages.com/WebSitesYouCanTrust-Oct08.pdf

Source Reliability: 43.3 9 (High Credibility)

Source Critiqued By

Bryan T. Eisenberg, Beisenbe5@gmail.com, Mercyhurst College, Erie PA, Advanced Analytic Techniques Course,” July 1^st  2011 .

Sources Cited by This Source

Dey, L., Mahajan, A., & Haque, S.M. (2009). Document clustering for event identification and trend analysis in market news. In Proceedings of ICAPR (pp. 103–106).

Duan, H., Cao, Y., Lin C., & Yu, Y. (2008). Searching questions by identifying question topic and question focus. In Proceedings of ACL (pp. 156–164).

Glance, N., Hurst, M., & Tomokiyo, T. (2004). BlogPulse: Automated trend discovery for Weblogs. In The workshop on the Weblogging ecosystem at the 13th international world wide Web conference, WWW.

Google Trends, 2006. http://www.google.com/trends. Jeon, J., Croft, W. B., & Lee, J. H. (2005). Finding similar questions in large question

and answer archives. In Proceedings of CIKM (pp. 84–90). Jurczyk, P., & Agichtein, E. (2007). Discovering authorities in question answer

communities using link analysis. In Proceedings of CIKM (pp. 919–922). Kleinberg, J. (2002). Bursty and hierarchical structure in streams. In Proceedings of

SIGKDD (pp. 91–101). Liu, Y., Bian, J., & Agichtein, E. (2008). Predicting information seeker satisfaction in

community question answering. In Proceedings of SIGIR (pp. 483–490).

Microplaza: (2009) <http://microplaza.com/public>. Nagano, S., Inaba, M., Mizoguchi, Y., & Iida, T. (2008). Takahiro Kawamura:

Ontology-based topic extraction service from Weblogs. In IEEE international

conference on semantic computing, (ICSC) (pp. 468–475). Osin ́ski, S., & Weiss, D. (2005). A concept-driven algorithm for clustering search

results. IEEE Intelligent Systems, 20, 48–54. May/June, 3. Rajaraman, K., Tan, A. (2001). Topic detection, tracking, and trend analysis using

self-organizing neural networks. In Proceedings of PAKDD (pp. 102–107). Sekiguchi, Y., Kawashima, H., Okuda, H., & Oku, M. (2006). Topic detection from blog documents using users’ interests. In The 7th international conference on mobile

data management, (MDM) (p. 108). The 2004 Topic detection and tracking. (TDT2004). Task definition and evaluation

plan, <http://www.nist.gov/speech/tests/tdt/>. Twopular: (2008) <http://twopular.com/>. Zhang, C., & Zhang, Q. (2008). Topic navigation generation using topic extraction

and clustering. In International symposium on knowledge acquisition and modeling (pp. 333–339).