近年來，隨著網際網路的不斷成長，網路開始逐漸取代各種傳統媒體，成為人們獲取新聞的主要管道之一。新聞網站越來越高的瀏覽率使得新聞網站成為一個良好的廣告平台。然而大部分的新聞網站主要的廣告策略都著重在顯眼性與數量上，卻忽略了使用者對於新聞主題的興趣。本論文認為應該根據不同的新聞主題去推薦相關的商品，才能有效吸引使用者。
為了找出與新聞主題相關的商品，本論文提出一個三階層的模型Event-Task-Product model，首先從新聞中抽取事件，然後找到相關的任務，最後找到與任務相關的商品。運用這個模型我們建立一個Event-Task-Product資料庫，基於這個資料庫我們提出一個商品推薦聊天機器人，可以根據使用者所觀看的新聞推薦相關商品。

In recent years, with the continuous growth of the Internet, the Internet has gradually replaced various traditional media and become one of the major channels for people to get news. The increasing browsing rate of news websites makes news websites become a good advertising platform. However, most of the main advertising strategies of news websites focus on the conspicuousness and quantity, but ignore the user's interest in news topics. This paper believes that relevant products should be recommended according to different news topics in order to effectively attract users.
In order to find out suitable products related to the news topic, this paper proposes a three-level model “Event-Task-Product model”, which first extracts events from the news, then finds related tasks, and finally finds the products related to the tasks. Using this model, we build an Event-Task-Product database. Based on this database, we propose a product recommendation chat bot that can recommend related products based on the news viewed by users.

[1] J.-J. Kuo and H.-H. Chen, Multidocument summary generation: using informative and event words. ACM Transactions on Asian Language Information Processing (TALIP), 7 (1). 2008.
[2] Chinese Knowledge Information Processing Group Academia Sinica Institute of Information Science. Technical Report no. 93-05中文詞類分析(三版)
[3] Chinese Knowledge Information Processing Group Academia Sinica Institute of Information Science. Technical Report no. 13-01句結構樹中的語意角色
[4] G. Kumaran and J. Allan. Text classification and named entities for new event detection. Proceeding SIGIR '04 Proceedings of the 27th annual international ACM SIGIR conference on Research and development in information retrieval, pp. 297-304. 2004.
[5] D. Rusu, J. Hodson, and A. Kimball. Unsupervised techniques for extracting and clustering complex events in news. In Proceedings of the Second Workshop on EVENTS: Definition, Detection, Coreference and Representation, pp. 26-34. 2014.
[6] C.-Y. Teng and H.-H. Chen. Analyzing Temporal Collocations in Weblogs. In ICWSM. 2007.
[7] Y. Wang, D. Fink, and E. Agichtein. Seeft: Planned social event discovery and attribute extraction by fusing twitter and web content. In Ninth International AAAI Conference on Web and Social Media. 2015.
[8] Y.-H. Lin and C.-H. Chang. Facebook Activity Event Extraction System. ROCLING 2016, pp. 229-243. 2016.
[9] C. Lucchese, S. Orlando, R. Perego, F. Silvestri, and G. Tolomei. Identifying task-based sessions in search engine query logs. In Proceedings of WSDM, pp. 277-286. 2011.
[10] Q. Guo and E. Agichtein. Ready to buy or just browsing?: detecting web searcher goals from interaction data. In Proceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval, pp. 130-137. 2010.
[11] A. Hassan, R. W. White, S. T. Dumais, and Y.-M. Wang. Struggling or exploring?: disambiguating long search sessions. In Proceedings of the 7th ACM international conference on Web search and data mining, pp. 53-62. 2014.
[12] T.-X. Wang, K.-Y. Tsai, and W.-H. Lu. Identifying Real-Life Complex Task Names with Task-Intrinsic Entities from Microblogs. Association for Computational Linguistics, pp. 470-475. 2014.
[13] D. M. Blei, A. Y. Ng, and M. I. Jordan. Latent dirichlet allocation. Journal of machine Learning research, 3(Jan), pp. 993-1022. 2003.
[14] F.O. Isinkaye, Y.O. Folajimi, and B.a. Ojokoh. Recommendation systems: Principles, methods and evaluation. Egyptian Informatics Journal, 16 (3), pp. 261-273. 2015.
[15] R.J. Mooney, L. Roy. Content-based book recommending using learning for text categorization. Proceedings of the fifth ACM conference on digital libraries, pp. 195-204. 2000.
[16] D. Billsus, M.J. Pazzani. User modeling for adaptive news access. User Model User-adapted Interact, 10 (2–3), pp. 147-180. 2000.
[17] U. Shardanand and P. Maes. Social information filtering: algorithms for automating “word of mouth”. Proceedings of the SIGCHI conference on human factors in computing systems, ACM Press/Addison-Wesley Publishing Co., pp. 210-217. 1995.
[18] J.A. Konstan, B.N. Miller, D. Maltz, J.L. Herlocker, L.R. Gordon, J. Riedl. Applying collaborative filtering to usenet news. Communications of the ACM, 40 (3), pp. 77-87. 1997.
[19] R. Burke. Hybrid recommender systems: survey and experiments. User Model User-adapted Interact, 12 (4), pp. 331-370. 2002.
[20] J. Bobadilla, F. Ortega, A. Hernando, A. Gutiérrez. Recommender systems survey. Knowl-Based Syst, 46, pp. 109-132. 2013.
[21] J.L. Herlocker, J.A. Konstan, L.G. Terveen, and J.T. Riedl. Evaluating collaborative filtering recommender systems. ACM Trans Inform Syst, 22 (1), pp. 5-53. 2004.
[22] Y.-M. Hsieh, M.-H. Bai, J. S. Chang, and K.-J. Chen. Improving pcfg chinese parsing with context-dependent probability reestimation. Proceedings of CLP’12, pp. 216-221. 2012.
[23] W.-T. Chen, S.-C. Lin, S.-L. Huang, Y.-S. Chung, and K.-J. Chen. E-HowNet and Automatic Construction of a Lexical Ontology. the 23rd International Conference on Computational Linguistics, Beijng, China. 2010.
[24] W. Wei, G. Cong, X. Li, S. K. Ng, and G. Li. Integrating Community Question and Answer Archives. In Proceedings of the Twenty-Fifth AAAI Conference on Artificial Intelligence. 2011.
[25] O. Kolesnikova. Survey of word co-occurrence measures for collocation detection Computacion y Sistemas, 20 (3), pp. 327-344. 2016.