Internet Protocol-based Television(IPTV)被稱為是新殺手級網路服務[1][16]，其具有互動性(Interactivity)、低頻寬(Low Bandwidth Requirement)、個人化(Personalization)、時間平移(Time shifting)、多接取裝置(Multiple Access Devices)等特性。IPTV服務包含了遞送群播電視、隨選視訊、三合一整合服務(Triple Play)、網路電視、網頁服務、以及電子郵件收發。以上所提及的服務項目，皆是在一般傳統電視所無法提供之互動式服務。
使用者觀看時的感受(Quality of Experience, QoE)，是衡量一項服務是否滿足使用者需求的指標。在傳統電視中，由於電視頻道是以廣播的方式進行傳播，因此當使用者有了轉台的行為，則僅需要進行接收頻率的切換，就可完成轉台的動作。但在IPTV中，若被要求的頻道尚未被傳送至網路上，則當使用者切換到該頻道，則需要等待一段串流抵達時間，因此造成一段黑屏的現象。因此頻道切換時間(Channel zapping Time)被視作是衡量使用者QoE的一項重要指標。
本論文中，著重在藉由使用者行為預測來減少頻道切換時間。目前有些已被提出的方法嘗試著預測使用者的行為來事先要求頻道，並且當使用者的操作行為滿足其預測時，則位於機上盒的頻道要求命中率就會提升。但是，在我們的觀察之中，當使用者的操作行為有了暫時性的變化，若沒有適時的作使用者操作行為的評估，則會使得預收的頻道不但無法加速頻道切換的時間，更會導致頻道資源受到耗用。為了以上所提及的問題，在本論文中，嘗試著在使用者行為變化時記錄並修正使用者的操作行為機率分佈，並根據使用者的行為機率分佈，我們所提出的機制可在使用者的轉台操作行為發生變動時，預測使用者下一步的轉台操作，以期降低頻道切換的時間。藉由模擬以及分析的結果來證明本論文所提得出的方法可以有效的增加頻道要求的命中率以及降低頻道切換時間。

Internet Protocol-based Television (IPTV) is called a new killer network service[1][16]; it includes a number of features such as interactivity, low bandwidth requirement, personalization, time shifting, and multiple access devices. IPTV services combine a lot of existing network services such as multicast TV, Video on Demand(VoD), triple play (Video、Voice、Data), web service and e-mail. These services we have mentioned are the interactive services that we cannot experience in traditional TV.
Quality of Experience (QoE) is one of the measures that whether the services meet user's needs. In traditional TV, these TV channels are broadcasting to air and end user just need to switch the read frequency by remote controller, and then they can finish a channel zapping. But in IPTV service, if the requested channel is not transmitted to core network or access network, there is a very long channel zapping time and the TV screen will be black. Therefore the channel zapping time becomes an important concept to measure User's QoE.
In this theme, we focus on reducing the channel zapping time based on user's behavior prediction. In some proposed scheme, they try to pre-join some channels in advance and the channel hit rate at the Set Top Box (STB) will increase when the user's behavior in line with their expectation. But the temporary change of the user’s behavior is not in their consideration. For this purpose, we try to record user's behavior and then modify the probability distribution of user's behavior when user has a channel zapping. According to the user's behavior probability distribution, STB can shorten channel zapping time by making a user's behavior prediction for user's next behavior when user's operation on IPTV has a change. Simulation results show that our solution increases the channel hit rate and decreases the channel zapping time than some proposed schemes

[1] Gerard O’Driscoll, “Next Generation IPTV Services and Technologies”, Join wiley & Sons, New Jersey, 2008.
[2] C. Cho, I. Han, Y. Jun, and H. Lee, ” Improvement of channel zapping time in IPTV service using the adjacent groups join-leave method.” , The 6th International Conference on Advanced Communication Technology (2004), Vol. 2, pp. 971-975,2004 C.
[3] Chuan-Ching Sue, et al., “A new IPTV channel zapping scheme for EPON”, International Conference on Ubiquitous and Future Networks (ICUFN 2009), pp. 7-9, Jun. 2009.
[4] J. Caja, “Optimization of IPTV Multicast Traffic Transport over Next Generation Metro Networks.”, 12th International Telecommunications Network Strategy and Planning Symposium (Networks 2006), pp. 1-6, Nov. 2006.
[5] H. Joo, D. Lee, and H. Song, “Effective IPTV channel management method over heterogeneous environments”, Proc. Of SPIE on Multimedia Systems and Applications X, vol. 6777, 677705, Sep. 2007.
[6] Yonghee Lee, et al., “Reducing IPTV Channel Switching Time Using H.264 Scalable Video Coding”, IEEE Transactions on Consumer Electronics, vol.54, No. 2, pp.912-919, May 2008.
[7] Smith, D.E., “IP TV Bandwidth Demand: Multicast and Channel Surfing,” INFOCOM 2007. 26th IEEE International Conference on Computer Communications, pp.2546-2550, 6-12 May 2007.
[8] M. Sarni, B. Hilt, and P. Lorenz, “A Novel Channel Switching Scenario in Multicast IPTV Networks,” Networking and Services, 2009. ICNS ‘09. Fifth International Conference on, pp.396-401, 20-25 April 2009.
[9] Roy D. Yates, David J. Goodman,“Probability and Stochastic Process”, John Wiley & Sons, New Jersey, 2005.
[10] Donald Gross, et al., “Queuing Theory”, John Wiley & Sons, New Jersey, 2008.
[11] J. Kishigami, “The Role of QoE on IPTV Services style,” Multimedia, 2007. ISM 2007. Ninth IEEE International Symposium on, pp.11-13, 10-12 Dec. 2007.
[12] M.Z. Ahmad, et al., “Prediction-based channel zapping latency reduction techniques for IPTV systems — A survey,” Emerging Technologies, 2009. ICET 2009. International Conference on, pp.466-470, 19-20 Oct. 2009.
[13] http://www.microsoft.com
[14] Ying-Mei Tu, Hsin-Nan Chen, “The effect of downtime frequency at fixed machine availability in queuing systems,” Service Operations and Logistics, and Informatics, 2008. IEEE/SOLI 2008. IEEE International Conference on, vol.2, pp.2696-2701, 12-15 Oct. 2008.
[15] R. Jain, “Quality of Experience,” IEEE Multimedia, 11:96-97.
[16] Yang Xiao, et al., “Internet Protocol Television (IPTV): The Killer Application for the Next-Generation Internet,” Communications Magazine, IEEE, vol.45, no.11, pp.126-134, November 2007.
[17] L.B. Sofman, B. Krogfoss, “Analytical Model for Hierarchical Cache Optimization in IPTV Network,” Broadcasting, IEEE Transactions on, vol.55, no.1, pp.62-70, March 2009.
[18] Z. Begic, M. Bolic, and H. Bajric, “Effect of multicast on IPTV channel change performance,” ELMAR, 2008. 50th International Symposium, vol.1, p.151-155, 10-12 Sept. 2008.
[19] Q. Chen, A.F. Norcio, “A neural network approach for user modeling Systems”, Man, and Cybernetics, 1991. Decision Aiding for Complex Systems, Conference Proceedings., 1991 IEEE International Conference on, pp.1429-1434 vol.2, 13-16 Oct 1991.
[20] I. Aleksander, and H. Morton, “An introduction to neural computing”, Intl Thomson Computer Pr (T); 2 Sub edition, October 1995.