Jessie+-+Annotated+Bibliography

Hey guys, this is my annotated bibliography which I already sent to our group leader, Dany :) Anyway, this is what I've got, I used MLA format to cite my sources... I don't know if we should do some intro paragraphs to "preface our bibliography", but we can definitely create one via wiki so we don't //have// to meet up... Hope y'all have a swell weekend!

CCT225 - Annotated Bibliography Jessie Liang March 11, 2010

The authors propose use of Semantic Web technologies in the AVATAR tool to decrease the weaknesses of automatic recommendation systems in Digital Television that seem to only have generated user disapproval. By explaining how existing recommendation systems function, they determine the errors in the software and what strategy they believe should be implemented in place. The paper details the logic behind the software modification through mathematical formulas and restructuring of information hierarchies. There is emphasis on the relationship between television programming structures and individual user profiles. The authors' success is measured by testing the new features with real DTV users. Results reveal their new strategy yields higher recommendation precision.
 * 1. Blanco-Fernández, Y., Pazos-Arias, J. J., Gil-Solla, A., Ramos-Cabrer, M., & Lópex-Nores, M. (2008). Personalization strategies and semantic reasoning: Working in tandem in advanced recommender systems. In Uchyigit, Gulden and Ma, Matthew Y. (Ed.), Personalization techniques and recommender systems (pp. 191). Singapore: World Scientific Publishing Co. Pte. Ltd.**

This paper starts by explaining the origins and the importance of recommendation systems. Authors also give an overview of existing technologies and their levels of user acceptance. The focus of the paper is to determine the factors affecting these systems' approval and actual use. Their results of their research show that understanding the available technology - like strengths and limitations – can influence their success in increased usage. The authors illustrate how information should be structured within the technology's database. In order to accurately calculate what to recommend, the inputted data must be specific and well-define to produce consistent suggestions to the users. The authors' experiment with their technology in the financial services sector. The paper concludes that newer technologies are more successful, but comprehending consumers thoroughly increase recommendation system usage.
 * 2. Felfernig, A., Teppan, E., & Gula, B. (2008). User acceptance of knowledge-based recommenders. In Uchyigit, Gulden and Ma, Matthew Y. (Ed.), Personalization techniques and recommender systems (pp. 249). Singapore: World Scientific Publishing Co. Pte. Ltd.**

Primarily, the authors explain the catalyst for creating a prototype for recommendation systems user interface. Basing themselves on Amazon.com's system, they identify the uses and benefits both customers and sellers attain from suggestions the web site makes based on the products being viewed. The paper proposes creating recommendation systems that are based on user behavior observations instead of opinions, surveys, and questionnaires. From this, the authors also illustrate how algorithms are influential in designing more accurate recommender systems. Thus, from the analysis of existing recommendation systems and algorithms, they suggest an improved method that should either replace or upgrade and improve already existing systems that are utilized in library organizational software.
 * 3. Franke, M., & Geyer-Schulz, A. (2008). Using restricted random walks for library recommendations and knowledge space exploration. In Uchyigit, Gulden and Ma, Matthew Y. (Ed.), Personalization techniques and recommender systems (pp. 277). Singapore: World Scientific Publishing Co. Pte. Ltd.**

In attempt to improve accuracy and save time in producing results within a recommendation system, the authors focus on the “feature selection” approach. They create a new method called GU metric, which through tests, show that their new form is as good or better than existing ones. Their purpose in integrating the GU metric is to classify information so to present the user with research results that not only have same words in the text, but also similar content. Their feature selection approach is used in TV recommender systems that function with individual profiling as well as content-based profiling. Their systems recommendation systems uses smaller “feature” words in the research; combined with past user habits to create the GU metric's more precise algorithm.
 * 4. Uchyigit, G., & Clark, K. (2008). An experimental study of feature selection methods for text classification. In Uchyigit, Gulden and Ma, Matthew Y. (Ed.), Personalization techniques and recommender systems (pp. 303). Singapore: World Scientific Publishing Co. Pte. Ltd.**

The authors explore the text classification technique to create their “electronic programming guide (EPG)” that filters and eliminates irrelevant and unrelated information in search results. The article explains how 3G services and Digital Television make more information available to people, however, simultaneously bombard users with triviality. Their proposal, then, is to combine the Internet, home media (like Digital TV), and mobile devices to form an integrated network of information. This database is compiled of different sources that store a user's patterns of behavior. The EPG would in turn employ the collected information to yield highly accurate and personal research results. Their recommender system is tested for actual user interest and produces favorable outcomes.
 * 5. Zhu, J., Ma, M. Y., Guo, J. K., & Wang, Z. (2008). Content classification and recommendation techniques for viewing electronic programming guide on a portable device. In Uchyigit, Gulden and Ma, Matthew Y. (Ed.), Personalization techniques and recommender systems (pp. 223). Singapore: World Scientific Publishing Co. Pte. Ltd.**