近年來學術論文的發表數量劇增，研究學者越來越難以從大量文獻中篩選出需要的學術論文，因此目前有不少研究提出以推薦系統的方法來幫助研究學者過濾資訊，而大多數的研究選擇使用詞袋模型(Bag-of-Words)來建立學術論文的內容特徵，並使用TF-IDF對這些特徵進行加權，這種模型的優點是簡單直觀、可解釋性高，缺點則是特徵稀疏、字詞相互獨立。
　　因此本研究提出了一種基於Word2vec的學術論文推薦系統，該模型能夠將字詞映射至低維度向量空間之中並且保留字詞間的線性關係，我們利用此模型的特性來找出相似內容特徵的論文並進行推薦。
　　最後，本研究設計了一項相關性實驗，該實驗將學術論文中的參考文獻視為相關論文，利用非人為的方式客觀的評估模型的相關性表現，而實驗結果也證明了Word2vec的有效性。

In recent years, the number of published academic papers has increased dramatically. How to find the needed articles in a large number of research papers has always been a time-consuming task. Therefore, academic resources discovery has been an open and challenging problem. At present, most studies of academic papers recommendation used Bag-of-words model to represent the features of papers. The advantages of this model are simple and intuitive. However, the disadvantages are the sparseness of the feature space and not considering the dependence between words. Therefore, in this thesis, based on the concept of Word2vec, a research paper recommender system is proposed. The proposed method uses content information to train the neural network model (Word2vec Skip-gram), then the features will be embedded in the vector space. The advantage of this method is that there is a linear correlation between features, and the similarity between research papers can be calculated efficiently. To show the advantages of the proposed method, a set of experiments is performed. The experimental results indicate that the proposed method can recommend papers to users effectively and efficiently.

第一章　緒論　1
1.1　研究背景　1
1.2　研究動機與目的　1
1.3　論文架構　2
第二章　文獻探討　5
2.1　推薦系統　5
2.1.1　協同過濾推薦系統　5
2.1.2　基於內容推薦系統　14
2.1.3　混合推薦系統　16
2.2　Word2vec詞向量模型　17
2.3　學術論文推薦系統　19
第三章　學術論文推薦系統　21
3.1　資料集　21
3.2　特徵工程與模型建立　21
3.3　使用者特徵與相似度計算　24
3.4　相關性反饋　24
3.5　系統架構　25
3.5.1　使用者介面　25
3.5.2　資料處理端　25
3.5.3　操作流程　25
第四章　系統實驗與評估　29
4.1　實驗資料集介紹　29
4.2　基於內容推薦之比較模型　30
4.3　模型建立與相似度計算時間　30
4.4　學術論文推薦範例　34
4.5　推薦系統滿意度調查　39
4.5.1　精確度(precision)、召回率(recall)　39
4.5.2　評估方法　40
4.5.3　推薦系統滿意度實驗結果　40
4.6　推薦系統相關性實驗　42
第五章　結論與未來展望　45
第六章　參考文獻　47

[1]IEEE Xplore. (2020). Retrieved from https://ieeexplore.ieee.org/Xplore/home.jsp
[2]BEEL, Joeran, et al. paper recommender systems: a literature survey. International Journal on Digital Libraries, 2016, 17.4: 305-338.
[3]Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781.
[4]Yoon, Y. C., & Lee, J. W. (2018, January). Movie Recommendation Using Metadata Based Word2Vec Algorithm. In 2018 International Conference on Platform Technology and Service(PlatCon) (pp. 1-6). IEEE.
[5]Chen, H. W., Wu, Y. L., Hor, M. K., & Tang, C. Y. (2017, July). Fully content-based movie recommender system with feature extraction using neural network. In 2017 International conference on machine learning and cybernetics (ICMLC) (Vol. 2, pp. 504-509). IEEE.
[6]Iaquinta, L., De Gemmis, M., Lops, P., Semeraro, G., Filannino, M., & Molino, P. (2008, September). Introducing serendipity in a content-based recommender system. In 2008 Eighth International Conference on Hybrid Intelligent Systems (pp. 168-173). IEEE.
[7]Covington, P., Adams, J., & Sargin, E. (2016, September). Deep neural networks for youtube recommendations. In Proceedings of the 10th ACM conference on recommender systems (pp. 191-198).
[8]Cheng, H. T., Koc, L., Harmsen, J., Shaked, T., Chandra, T., Aradhye, H., ... & Anil, R. (2016, September). Wide & deep learning for recommender systems. In Proceedings of the 1st workshop on deep learning for recommender systems (pp. 7-10).
[9]Jannach, D., Zanker, M., Felfernig, A., & Friedrich, G. (2010). Recommender systems: an introduction. Cambridge University Press.
[10]Chhavi Saluja. Collaborative Filtering based Recommendation Systems exemplified. (2018). Retrieved from https://towardsdatascience.com/collaborative-filtering-based-recommendation-systems-exemplified-ecbffe1c20b1
[11]Linden, G., Smith, B., & York, J. (2003). Amazon. com recommendations: Item-to-item collaborative filtering. IEEE Internet computing, 7(1), 76-80.
[12]Matrix Factorization|Recommendation Systems|Google Developers. (2020). Retrieved from https://developers.google.com/machine-learning/recommendation/collaborative/matrix
[13]Sun Junyi.fxsjy/jieba:结巴中文分词.(2020). Retrieved from https://github.com/fxsjy/jieba
[14]Aprilianti, M., Mahendra, R., & Budi, I. (2016, October). Implementation of weighted parallel hybrid recommender systems for e-commerce in indonesia. In2016 International Conference on Advanced Computer Science and Information Systems (ICACSIS) (pp. 321-326). IEEE.
[15]McCormick, C. (2016). Word2vec tutorial-the skip-gram model.
[16]Lee, J., Lee, K., & Kim, J. G. (2013). Personalized academic research paper recommendation system. arXiv preprint arXiv:1304.5457.
[17]Yang, C., Wei, B., Wu, J., Zhang, Y., & Zhang, L. (2009, June). CARES: a ranking-oriented CADAL recommender system. In Proceedings of the 9th ACM/IEEE-CS joint conference on Digital libraries (pp. 203-212).
[18]McNee, S. M., Albert, I., Cosley, D., Gopalkrishnan, P., Lam, S. K., Rashid, A. M., ... & Riedl, J. (2002, November). On the recommending of citations for research papers. In Proceedings of the 2002 ACM conference on Computer supported cooperative work (pp. 116-125).
[19]Pohl, S., Radlinski, F., & Joachims, T. (2007, June). Recommending related papers based on digital library access records. In Proceedings of the 7th ACM/IEEE-CS joint conference on Digital libraries (pp. 417-418).
[20]arXiv.org. (2020). Retrieved from https://arxiv.org/
[21]林士傑，基於學術研究發展之論文推薦系統，國立東華大學碩士論文，民國107年。
[22]Rehurek, R., & Sojka, P. (2010). Software framework for topic modelling with large corpora. In Proceedings of the LREC 2010 Workshop on New Challenges for NLP Frameworks.