臺灣博碩士論文加值系統

本篇論文提出一個建立在Instagram上的照片推薦系統，方便讓使用者透過輸入標籤或是點擊圖片來查詢相關圖片。使用者可以透過推薦系統的功能來獲得Instagram上面的資訊。系統除了使用Instagram API收集可用的標籤，還使用the Free Dictionary來建立知識庫中標籤之間的關係。而系統提供兩種相關性的查詢；user-defined correlation和system-defined correlation。最後，實驗證明user-defined correlation和system-defined correlation方法都具有良好的使用者滿意程度。

In this paper, we propose a picture recommendation system built on Instagram, which facilitate users querying correlated pictures by keying in hashtags or clicking images. Users can access the value-added information (or pictures) on Instagram through the recommendation platform. In addition to collect available hashtags using the Instagram API, the system also uses the Free Dictionary to build the relationships between all the hashtags in a knowledge base. Thus, two kinds of correlations can be provided for a query in the system; i.e., user-defined correlation and system-defined correlation. Finally, the experimental results show that both user-defined correlation and system-defined correlation methods have good satisfaction degrees with users.

摘要..........................................................................i
Abstract....................................................................ii
誌謝........................................................................iii
Content.....................................................................iv
List of Figures..............................................................v
1 Introduction...............................................................1
2 System Overviews...........................................................2
3 Offline Processing.........................................................3
3.1 Preprocessor.....................................................3
3.2 Correlation Matrix Generator.....................................3
3.2.1 User-defined and System-defined Hashtag Matrices.......3
3.2.2 Correlation Calculation................................4
3.3 Term Weight Calculation..........................................6
4 Online Processing..........................................................7
4.1 Login and Authorization..........................................7
4.2 Query Processing.................................................7
4.2.1 User-defined Correlation...............................7
4.2.2 System-defined Correlation.............................8
4.3 Term Ranking.....................................................8
5 Implementation and Evaluation.............................................10
5.1 User Interface..................................................10
5.2 Evaluation Results..............................................11
6 Conclusions...............................................................15
References..................................................................16

[1]R. Agrawal, A. Gupta, Y. Prabhu, and M. Varma, “Multi-label learning with millions of labels: recommending advertiser bid phrases for web pages,” Proc. the 22th International Conference on World Wide Web, pp.13-24, 2013.
[2]S. Brin and L. Page, “Reprint of: the anatomy of a large-scale hypertextual web search engine,” Computer Networks, Vol. 56, No. 18, pp. 3825-3833, 2012.
[3]S. Calegari and G. Pasi, “Personal ontologies: generation of user profiles based on the YAGO ontology,” Information Processing & Management, Vol. 49, No. 3, pp.640-658, 2013.
[4]C. Carpineto and G. Romano, “A survey of automatic query expansion in information retrieval,” Computing Surveys, Vol. 44, No. 1, pp. 1-50, 2012.
[5]S. Haiduc, G. Bavota, R. Oliveto, A. Marcus, and A. De Lucia, “Evaluating the specificity of text retrieval queries to support software engineering tasks,” Proc. the 34th International Conference on Software Engineering, pp. 1273-1276, 2012.
[6]S. Haiduc, G. Bavota, A. Marcus, R. Oliveto, A. De Lucia, T. Menzies, “Automatic query reformulations for text retrieval in software engineering,” Proc. the 35th International Conference on Software Engineering, pp. 842-851, 2013.
[7]L. Meng, R. Huang, and J. Gu, “A review of semantic similarity measures in WordNet,” International Journal of Hybrid Information Technology, Vo1. 6, No. 1, pp. 1-12, 2013.
[8]A. Montejo-Raez, E. Martínez-Cámara, M. Teresa Martín-Valdivia, and L. Alfonso Ureña-López, “Ranked WordNet graph for sentiment polarity classification in Twitter,” Computer Speech & Language, Vol. 28, No. 1, pp. 93-107, 2014.
[9]S. Poria, A. Gelbukh, E. Cambria, P. Yang, A. Hussain, and T. Durrani, “Merging SenticNet and WordNet-affect emotion lists for sentiment analysis,” Proc. the 11th International Conference on Signal Processing, pp. 1251-1255, 2012.
[10]The Free Dictionary, http://www.thefreedictionary.com/.
[11]Instagram API, https://www.instagram.com/developer/.
[12]Picture Recommendation System,
http://mdb.csie.yuntech.edu.tw/instagram/