線上問答社群在這幾年非常地流行，而它有也多種類型。在這篇論文，我們專注在線上醫療問答社群。線上醫療問答社群提供了一個平台給想要尋求健康相關資訊的人們。人們喜歡使用這樣的平台因為它方便與免費的特性。有兩種使用這個平台的方法，第一是在平台上發問並等待專業醫生們的回答，另一種則是可以在平台上搜尋曾經被回答過的與自己問題相關的問題。對後者而言，人們喜好那些被過去的發問者所接受的答案。但是平台上大多數的問題卻沒有被接受的答案，這對那些想要搜尋問題的人來說變得很不方便。為了解決這個問題，我們想要從沒有被接受的答案中找出高品質的答案。在這篇論文中，我們提出深度學習的方法來達到我們的目的。為了訓練答案品質分數預測模型，我們先將有被接受的答案視為正面答案並用隱含狄利克雷分布模型標註負面答案。接著透過結合雙向長短期記憶模型與卷積神經網路我們可以抓到問題與答案的語意特徵。之後我們綜合問題與答案的語意特徵，用此來預測這個答案的品質分數。我們從最大的中文線上醫療問答社群中蒐集資料並且把它分為多個醫療科別做更詳細的分析。最後透過實驗來展示我們分類與標註負面答案的方法的有效性。我們的結果超越了其他研究而且我們還探討了不同類別之間的差異。

The online question answering (QA) community has been popular in recent years. There are many types of the QA communities. In this thesis, we focus on the online health question answering (HQA) community. The HQA community provides a platform for health consumers to inquire about health information. They prefer to use this platform because of its convenience and cost-effectiveness. There are two ways to use this platform. One is to post a question and wait for answers to be provided by authenticated doctors. The other is to search for relevant questions with answers. For the latter, the health consumer may prefer an accepted answer marked by the previous health consumer. However, there is a large proportion of questions without an accepted answer marked. It becomes inconvenient for those who want to search for relevant questions with an accepted answer. To address this issue, we aim to find high-quality answers from the answers without accepted answers been marked. In this thesis, we propose a deep learning approach to achieve this goal. For training the answer quality prediction model, we first view the accepted answer as the positive answer and propose a method to label the negative answer. Next, through combining Bidirectional Long Short-Term Memory (Bi-LSTM) and Convolutional Neural Network (CNN), we capture the semantic information on the question and the answer. After that, we combine the information of the question and the answer to predict the quality score of the answer. We collect data from one of the biggest Chinese HQA community and divide it into medical departments for detailed analysis. Finally, we conduct experiments to show the effectiveness of categorization and the method to label the negative answer. Moreover, our results outperform other studies and we further discuss the difference between categories.

[1] Amr Azzam, Neamat Tazi, Ahmad Hany Hossny, “A Question Routing Technique Using Deep Neural Network,” Database Systems for Advanced Applications (DASFAA), 1, pp. 35-49, 2017
[2] David M. Blei, Andrew Y. Ng, Michael I. Jordan, “Latent Dirichlet Allocation,” Journal of Machine Learning Research, 3, pp. 993-1022, 2003
[3] Mohan John Blooma, Alton Yeow-Kuan Chua, Dion Hoe-Lian GohA, “A Predictive Framework for Retrieving the Best Answer,” Symposium on Applied Computing (SAC), pp. 1107-1111, 2008
[4] Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, Yoshua Bengio, “Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling,” CoRR, abs/1412.3555, 2014
[5] Ruichu Cai, Binjun Zhu, Lei Ji, Tianyong Hao, Jun Yan, Wenyin Liu, “An CNN-LSTM Attention Approach to Understanding User Query Intent from Online Health Communities,” ICDM Workshops, pp. 430-437, 2017
[6] Minwei Feng, Bing Xiang, Michael R. Glass, Lidan Wang, and Bowen Zhou, “Applying deep learning to answer selection: A study and an open task,” IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), pp. 813-820, 2015
[7] R Flesch, “A New Readability Yardstick,” Journal of Applied Psychology, 32(3), pp. 221-233, 1948
[8] Junqing He, Mingming Fu, Manshu Tu, “Applying deep matching networks to Chinese medical question answering: a study and a dataset,” BMC Medical Informatics & Decision Making, 19-S (2), pp. 91-100, 2019
[9] Sepp Hochreiter, Jurgen Schmidhuber, “Long Short-Term Memory,” Neural Computation, 9, pp. 1735-1780, 1997
[10] Ze Hu, Zhan Zhang, Haiqin Yang, Qing Chen, Rong Zhu, Decheng Zuo, “Predicting the Quality of Online Health Expert Question-Answering Services with Temporal Features in a Deep Learning Framework,” Neurocomputing, 275, pp.2769-2782, 2018
[11] Zellig S Harris, “Distributional Structure,” Word, 10(2-3): 146 - 162, 1954
[12] Jiwoon Jeon, W. Bruce Croft, Joon Ho Lee, Soyeon Park, “A Framework to Predict the Quality of Answers with Non-Textual Features,” International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), pp. 228-235, 2006
[13] Diederik P. Kingma, Jimmy Ba, “Adam: A Method for Stochastic Optimization,” International Conference on Learning Representations (ICLR), 2015
[14] Nal Kalchbrenner, Edward Grefenstette, Phil Blunsom, “A Convolutional Neural Network for Modelling Sentences,” Meeting of the Association for Computational Linguistics (ACL), 1, pp. 655-665, 2014
[15] Yoon Kim, “Convolutional Neural Networks for Sentence Classification,” Empirical Methods in Natural Language Processing (EMNLP), pp. 1746-1751, 2014
[16] Pengfei Liu, Xipeng Qiu, Xuanjing Huang, “Recurrent Neural Network for Text Classification with Multi-Task Learning,” International Joint Conferences on Artificial Intelligence (IJCAI), pp. 2873-2879, 2016
[17] Ryan Lowe, Nissan Pow, Iulian V Serban, and Joelle Pineau, “The Ubuntu Dialogue Corpus: A Large Dataset for Research in Unstructured Multi-turn Dialogue Systems,” Special Interest Group on Discourse and Dialogue (SIGDIAL), pp. 285-294, 2015
[18] GH McLaughlin, “SMOG Grading - a New Readability Formula,” Journal of Reading, 12(8), pp. 639-646, 1969
[19] Jonas Mueller, Aditya Thyagarajan, “Siamese Recurrent Architectures for Learning Sentence Similarity,” Association for the Advancement of Artificial Intelligence (AAAI), pp. 2786-2792, 2016
[20] Tomas Mikolov, Kai Chen, Greg Corrado, Jeffrey Dean, “Efficient Estimation of Word Representations in Vector Space,” International Conference on Learning Representations (ICLR), 2013
[21] Fatemeh Riahi, Zainab Zolaktaf, M. Mahdi Shafiei, Evangelos E. Milios, “Finding Expert Users in Community Question Answering,” International World Wide Web Conferences (WWW), pp. 791-798, 2012
[22] Bin Shao, Jiafei Yan, “Recommending Answerers for Stack Overflow with LDA Model,” Chinese Conference on Computer Supported Cooperative Work and Social Computing (ChineseCSCW), pp. 80-86, 2017
[23] Chirag Shah, Jeffrey Pomerantz, “Evaluating and Predicting Answer Quality in Community QA,” International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), pp. 411-418, 2010
[24] Tirath Prasad Sahu, Naresh Kumar Nagwani, Shrish Verma, “Selecting Best Answer: An Empirical Analysis on Community Question Answering Sites,” IEEE Access, 4, pp. 4797-4808, 2016
[25] Duyu Tang, Bing Qin, Xiaocheng Feng, Ting Liu, “Effective LSTMs for Target-Dependent Sentiment Classification,” International Conference on Computational Linguistics (COLING), pp. 3298-3307, 2016
[26] Hapnes Toba, Zhaoyan Ming, Mirna Adriani, Tat-Seng Chua, “Discovering High Quality Answers in Community Question Answering Archives Using a Hierarchy of Classifiers,” Information Sciences, 261, pp. 101-115, 2014
[27] Ming Tan, Bing Xiang, Bowen Zhou, “LSTM-based Deep Learning Models for Non-factoid Answer Selection,” International Conference on Learning Representations (ICLR), 2016
[28] Yuanhe Tian, Weicheng Ma, Fei Xia, Yan Song, “ChiMed: A Chinese Medical Corpus for Question Answering,” BioNLP, pp. 250-260, 2019
[29] Mengqiu Wang, Noah A. Smith, Teruko Mitamura, “What is the Jeopardy Model? A Quasi-Synchronous Grammar for QA,” The Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL), pp. 22-32, 2007
[30] Peng Wang, Bo Xu, Jiaming Xu, Guanhua Tian, Cheng-Lin Liu, Hongwei Hao, “Semantic Expansion Using Word Embedding Clustering and Convolutional Neural Network for Improving Short Text Classification,” Neurocomputing, 174, pp. 806-814, 2016
[31] Xingyou Wang, Weijie Jiang, Zhiyong Luo, “Combination of Convolutional and Recurrent Neural Network for Sentiment Analysis of Short Texts,” International Conference on Computational Linguistics (COLING), pp. 2428-2437, 2016
[32] Zhiguo Wang, Wael Hamza, Radu Florian, “Bilateral Multi-Perspective Matching for Natural Language Sentences,” International Joint Conferences on Artificial Intelligence (IJCAI), pp. 4144-4150, 2017
[33] Dong Ye, Sheng Zhang, Hui Wang, Jiajun Cheng, Xin Zhang, Zhaoyun Ding, Pei Li, “Multi-Level Composite Neural Networks for Medical Question Answer Matching,” IEEE International Conference on Data Science in Cyberspace (DSC), pp. 139-145, 2018
[34] Seunghyun Yoon, Joongbo Shin, Kyomin Jung, “Learning to Rank Question-Answer Pairs Using Hierarchical Recurrent Encoder with Latent Topic Clustering,” Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT), pp.1575-1584, 2018
[35] Chenwei Zhang, Wei Fan, Nan Du, Philip S. Yu, “Mining User Intentions from Medical Queries,” International World Wide Web Conferences (WWW), pp. 1373-1384, 2016
[36] Chenwei Zhang, Nan Du, Wei Fan, Yaliang Li, Chun-Ta Lu, Philip S. Yu, “Bringing Semantic Structures to User Intent Detection in Online Medical Queries,” IEEE International Conference on Big Data (Big Data), pp. 1019-1026, 2017
[37] Rui Zhang, Honglak Lee, Dragomir R. Radev, “Dependency Sensitive Convolutional Neural Networks for Modeling Sentences and Documents,” North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pp. 1512-1521, 2016
[38] Sheng Zhang, Xin Zhang, Hui Wang, Jiajun Cheng, Pei Li, Zhaoyun Ding, “Chinese Medical Question Answer Matching Using End-to-End Character-Level Multi-Scale CNNs,” Applied Sciences, 7(8), 2017
[39] Tom Chao Zhou, Michael R. Lyu, Irwin King, “A Classification-based Approach to Question Routing in Community Question Answering,” International World Wide Web Conferences (WWW), pp. 783-790, 2012
[40] Thomas Zhang, Jason H. D. Cho, Chengxiang Zhai, “Understanding User Intents in Online Health Forums,” ACM International Conference on Bioinformatics and Computational Biology (BCB), pp. 220-229, 2014