研究歷程中，文獻探討需同時具有「相關」且「重要」之著作，因此，研究人員在引用時需同時顧及文獻之完整性與新穎性。現今的學術文獻搜尋引擎與相關文獻推薦之研究著重於推薦具關聯性(Relevance)之文獻，然無法從具關聯性的文獻中評估新穎性程度，進而使得研究人員仍需對推薦內容進行評估與篩選。
為滿足學術文獻推薦系統同時具備重要性與新穎性，本論文提出以權威擴散法(Citation Authority Diffusion)對於引用網路(Citation Network)之特定研究主題推薦出同時具重要性且能滿足新穎性的文獻清單。本研究為每一項文獻的引用行為定義出權威矩陣(Authority Matrix)，再由引用網路中統計每一位作者被提及的次數並做數量的標準化，對照文獻中各自包含多少作者與標準化分數，即為雙方引用文獻的權威與影響力。權威擴散法基於權威矩陣的定義，以可信度傳遞推論演算法(Belief Propagation)將引用網路內每個文獻的權威利用其引用對象進行權威度的擴散。引用網路收斂後的結果，重要(具權威性)且新穎的文章將會被推薦。
由實驗結果可看出本方法對於某項研究所推薦的文獻清單，對應於該研究自身的參考文獻之間的共同被引用機率(Co-cited Probability)，隨著推薦文章筆數增加，其演變程度僅為線性衰退。因此，本方法能持續推薦出重要且新穎的文獻。有別於其它相關注重在推薦累積大量引用率之以引用網路為基礎之方法，本方法以推薦出具新穎性之相關文獻為核心價值。

Survey of academic literatures or papers should be considered with both relevance and importance of references. Researchers may cite the related references with both integrity and novelty for their research. However, the current public search engines of scholarly papers and some corresponded researches only recommend the relevant papers toward the target research. That is, researchers should evaluate the novelty of the recommended relevant papers by themselves.
In this thesis, we propose a citation-network-based methodology, called Citation Authority Diffusion (CAD), to rapidly discover the limited key papers for the target research, and measure the novelty on literature survey. A defined Authority Matrix (AM) is used to standardize duplication rate of authors and to describe the authority relation between the citing and the cited papers. Based on AM, our CAD methodology leverages the Belief Propagation to diffuse the authority among the citation network. Therefore, CAD transforms the converged citation network to a novelty paper list to researchers.
The experimental results show that CAD can find out more novelty paper than other paper recommendation methods under the top-k result. The co-cited probability of CAD and structure-based approaches represent in linear decay as enlarging the number of recommended papers. In contract, the content-based approach comes with exponential decay as the growth of the number of recommended papers. According to the analysis and experiment results, CAD could indeed keep recommending related papers with novelty suited for the literature survey.

[1] The acl anthology network. http://clair.si.umich.edu/clair/anthology/.
[2] arxiv.org e-print archive. http://arxiv.org/.
[3] Citeseerx. http://citeseerx.ist.psu.edu/.
[4] Current index to statistics. http://www.statindex.org/.
[5] E-prints in library and information science. http://eprints.rclis.org/.
[6] Fao home. http://www.fao.org/.
[7] Freecite. http://freecite.library.brown.edu/.
[8] Google scholar. http://scholar.google.com/.
[9] International union of crystallography. http://www.iucr.org/.
[10] Project euclid. http://projecteuclid.org/.
[11] Universitt trier: Profil. http://www.informatik.uni-trier.de/.
[12] Adomavicius, G., and Tuzhilin, A. Context-aware recommender systems.
Recommender Systems Handbook (2011), 217–253.
[13] Berry, M., and Castellanos, M. Survey of text mining II. Springer, 2008.
[14] Brody, T., Jiao, Z., Krichel, T., and Warner, S. Syntax and vocabu-
lary of the academic metadata format, 2001.
[15] Elkiss, A., Shen, S., Fader, A., Erkan, G., et al. Blind men and
elephants: What do citation summaries tell us about a research article? Journal
of the American Society for Information Science and Technology (JASIST) 59,
1 (2008), 51–62.
[16] Frantzi, K., Ananiadou, S., and Tsujii, J. The c-value/nc-value method
of automatic recognition for multi-word terms. Research and Advanced Tech-
nology for Digital Libraries (ECDL) (2009), 520–520.
[17] Glaser, H., Jaffri, A., and Millard, I. Managing co-reference on the
semantic web. In Linked Data on the Web (LDOW) (2009), CEUR-WS, pp. 1–6.
[18] He, Q., Pei, J., Kifer, D., Mitra, P., and Giles, L. Context-aware
citation recommendation. In International Conference on World Wide Web
(WWW) (2010), ACM, pp. 421–430.
[19] Huang, Z., and Qiu, Y. A multiple-perspective approach to constructing
and aggregating citation semantic link network. Future Generation Computer
Systems (FGCS) 26, 3 (2010), 400–407.
[20] Jones, K., and Willett, P. Readings in information retrieval. Morgan
Kaufmann Pub, 1997.
[21] Kaplan, D., Iida, R., and Tokunaga, T. Automatic extraction of cita-
tion contexts for research paper summarization: a coreference-chain based ap-
proach. In Workshop on text and citation analysis for scholarly digital libraries
(NLPIR4DL) (2009), Association for Computational Linguistics, pp. 88–95.
[22] Keshav, S. How to read a paper. ACM SIGCOMM Computer Communication
Review (CCR) 37, 3 (2007), 83–84.
[23] Konstan, J., Miller, B., Maltz, D., Herlocker, J., Gordon, L.,
and Riedl, J. Grouplens: applying collaborative filtering to usenet news.
Communications of the ACM 40, 3 (1997), 77–87.
[24] Koren, Y. Factorization meets the neighborhood: a multifaceted collaborative
filtering model. In International Conference on Knowledge Discovery and Data
Mining (SIGKDD) (2008), ACM, pp. 426–434.
[25] Li, L., Chu, W., Langford, J., and Schapire, R. A contextual-bandit
approach to personalized news article recommendation. In International Con-
ference on World Wide Web (WWW) (2010), ACM, pp. 661–670.
[26] Liang, H., Xu, Y., Li, Y., and Nayak, R. Collaborative filtering rec-
ommender systems using tag information. In IEEE/WIC/ACM International
Conference on Web Intelligence (IAT) (2009), vol. 3, IEEE, pp. 59–62.
[27] Mooij, J. M. libDAI: A free and open source C++ library for discrete ap-
proximate inference in graphical models. Journal of Machine Learning Research
(JMLR) 11 (Aug. 2010), 2169–2173.
[28] Naak, A., Hage, H., and Aimeur, E. A multi-criteria collaborative filter-
ing approach for research paper recommendation in papyres. E-Technologies:
Innovation in an Open World (2009), 25–39.
[29] Pan, C., and Li, W. Research paper recommendation with topic analysis. In
Computer Design and Applications (ICCDA) (2010), vol. 4, IEEE, pp. V4–264.
[30] Pearl, J. Probabilistic reasoning in intelligent systems: networks of plausible
inference. Morgan Kaufmann, 1988.
[31] Qazvinian, V., and Radev, D. Scientific paper summarization using cita-
tion summary networks. In International Conference on Computational Lin-
guistics (COLING) (2008), Association for Computational Linguistics, pp. 689–
696.
[32] Radev, D., Muthukrishnan, P., and Qazvinian, V. The acl anthology
network corpus. In Workshop on text and citation analysis for scholarly dig-
ital libraries (NLPIR4DL) (2009), Association for Computational Linguistics,
pp. 54–61.
[33] Rasanen, E., Kikta, R., Sorvari, A., Salmenkaita, J., Huhtala, Y.,
Mannila, H., Toivonen, H., Oinonen, K., and Murto, J. Location-
based novelty index value and recommendation system and method, Mar. 4
2008. US Patent App. 20,080/214,210.
[34] Sarwar, B., Karypis, G., Konstan, J., and Reidl, J. Item-based col-
laborative filtering recommendation algorithms. In International Conference
on World Wide Web (WWW) (2001), ACM, pp. 285–295.
[35] Shi, X., Leskovec, J., and McFarland, D. Citing for high impact. In
Joint Conference on Digital Libraries (JCDL) (2010), ACM, pp. 49–58.
[36] Su, X., and Khoshgoftaar, T. A survey of collaborative filtering tech-
niques. Advances in Artificial Intelligence 2009 (2009), 4.
[37] Toutanova, K., and Manning, C. Enriching the knowledge sources used
in a maximum entropy part-of-speech tagger. In Proceedings of the 2000 Joint
SIGDAT conference on Empirical methods in natural language processing and
very large corpora: held in conjunction with the 38th Annual Meeting of the
Association for Computational Linguistics-Volume 13 (2000), Association for
Computational Linguistics, pp. 63–70.
[38] Tso-Sutter, K., Marinho, L., and Schmidt-Thieme, L. Tag-aware
recommender systems by fusion of collaborative filtering algorithms. In ACM
Symposium on Applied Computing (SAC) (2008), ACM, pp. 1995–1999.
[39] Voorhees, E. The trec-8 question answering track report. National Institute
of Standards and Technology Special Publication (NISTSP) (2000), 77–82.
[40] Wan, S., Paris, C., Muthukrishna, M., and Dale, R. Designing a
citation-sensitive research tool: an initial study of browsing-specific information
needs. In Workshop on text and citation analysis for scholarly digital libraries
(NLPIR4DL) (2009), Association for Computational Linguistics, pp. 45–53.