在社群大數據的影響下，消費者大量的發表評論於社群網路上。而為了得知消費者的意見傾向，情感分析被大量的應用於文本資料的分析上。然而在過去的文獻中較少應用深度學習於中文評論上，因此本研究檢驗深度學習應用於情感分析的效果。
	本研究開發網路爬蟲程式蒐集Google Play上總計196,651條評論，以深度學習、貝氏演算法、支援向量機，三種方法進行情感分析，並比較效果與正確率。
	實驗結果顯示，貝氏演算法的正確率為74.12%、支援向量機得到76.46%、深度學習預測模型的正確率為94%。從而證明深度學習在情感分析上的預測效果最為出色。
	本論文的研究貢獻為 (1)本研究提出一套使用深度學習於中文手機應用程式消費者評論情感分析方法，實驗結果證實本研究所提出的深度學習方法能有效提升消費者評論情感分析正確率。(2)本研究透過文本資料分析，建構出適用於手機應用程式領域的正負向意見詞典

Influenced by Big Data, there are a large number of customers shared their product reviews on social media. Therefore, many researchers implement sentiment analysis technique on consumer reviews to understand the opinion tendency. However, there are few research about implement deep learning method on Chinese customer reviews. It is therefore the intent of the present study to examine the effect of sentiment analysis with deep learning method.
	The study used web mining technique collected 196,651 reviews on Google Play. In addition, we use deep learning, Naïve Bayes, Support vector machine methods for sentiment analysis and compared the result.
	The present study display the accuracy of the Naïve Bayes is 74.12%, the accuracy of Support vector machine is 76.46%, and the accuracy of deep learning is 94%. Our finding confirm that sentiment analysis with deep learning is outstanding.
	There are three contributions in present finding. First, the present study confirm sentiment analysis with deep learning on Chinese cell phone application customers reviews may improve the accuracy of prediction. Second, the present study create a sentiment dictionary of cell phone application. Third, the study compared the result of average sampling data and non-average sampling data. We found that deep learning method with non-average sampling data reached the better performance.

目次


 第一章 緒論	1
1.1	研究背景	1
1.2	研究動機	1
1.3	研究目的	2
1.4	研究架構	3
 第二章 文獻探討	4
2.1	情感分析	4
2.2	意見辭典	5
2.3	貝氏演算法(Naïve Bayes)	6
2.4	類神經網路	7
2.5	支援向量機(Support Vector Machine)	7
2.6	深度學習	9
2.6.1	活化函數(Activation Function)	10
2.6.2	卷積神經網路(Convolutional Neural Networks)	12
2.6.3	遞迴式類神經網路(Recurrent Neural Networks)	14
2.6.4	長短期記憶(Long Short Term Memory)	15
2.6.5	閘循環單元(Gated Recurrent Unit)	16
2.6.6	雙向遞迴式類神經網路(Bidirectional Recurrent  Neural Network)	17
2.7	本章小節	19
 第三章 研究方法與系統架構	21
3.1	研究方法	21
3.2	系統架構	24
3.3	手機應用程式評論蒐集	25
3.3.1	評論蒐集	25
3.3.2	資料前處理	27
3.4	新詞萃取	29
3.5	特徵值產生	29
3.6	產生詞向量(Word Embedding)	31
3.7	建立深度學習模型與	32
3.8	深度模型參數設定	34
3.8.1	損失函數(Loss Function)	34
3.8.2	優化器(Optimizer)	35
 第四章 資料分析與結果	38
4.1	資料分配與實驗說明	38
4.1.1	資料分配	38
4.1.2	實驗說明	43
4.2	實驗分析結果	44
4.2.1	基於貝氏演算法(Naïve Bayes)的意見分析	44
4.2.2	基於支援向量機(Support Vector Machine)的情感分析	46
4.2.3	基於長短期記憶(Long Short Term Memory)的情感分析	46
4.2.4	基於GRU之情感分析	52
4.2.5	基於雙向長短期記憶之情感分析	56
4.2.6	各活化函數與優化器搭配結果比對	60
4.2.7	各機器學習效果比對	64
 第五章 結論與意涵	65
5.1	結論	65
5.2	研究貢獻	66
5.3	管理意涵	66
5.4	未來研究方向	67
 參考文獻	68


圖目次
圖 1支援向量機分類圖	8
圖 2非線性分類資料	9
圖 3 雙彎曲函數	10
圖 4 雙彎曲正切函數	11
圖 5 線性修正函數	12
圖 6 基於自然語言處理的CNN架構	13
圖 7 遞迴式類神經網路架構	15
圖 8 長短期記憶	16
圖 9 GRU架構圖	17
圖 10 BRNN架構圖	18
圖 11 資訊系統發展方法研究生命週期循環圖	22
圖 12系統發展研究法研究流程圖	23
圖 13深度學習於社群網路消費者評論之情感分析研究	24
圖 14 Google Play 分類查詢介面	26
圖 15 Google Play 手機應用程式使用者評論畫面	27
圖 16 原始資料正負面評論數量分布圖	28
圖 17 結構化後的使用者評論資料	28
圖 18支援向量機訓練資料格式	31
圖 19 詞向量結果截圖	32
圖 20 深度學習模型建構流程圖	34
圖 21 深度模型建立流程圖	37
圖 22 無評論留言之示意圖	39
圖 23 負向評論留言字數折線圖	40
圖 24 正向評論留言字數折線圖	41
圖 25 訓練資料正負數量	42
圖 26 原始資料與實驗用語料比較分布圖	43
圖 27非平均採樣實驗A7 的LSTM訓練圖	48
圖 28非平均採樣實驗A9 的LSTM訓練圖	49
圖 29 平均採樣實驗B7 的LSTM訓練圖	50
圖 30平均採樣實驗B9 的LSTM訓練圖	51
圖 31非平均採樣實驗C7 的GRU訓練圖	53
圖 32非平均採樣實驗C8的GRU訓練圖	53
圖 33 平均採樣實驗D7的GRU訓練圖	55
圖 34 平均採樣實驗D8的GRU訓練圖	55
圖 35非平均採樣實驗E7的BLSTM訓練圖	57
圖 36非平均採樣實驗E8的BLSTM訓練圖	57
圖 37 平均採樣實驗F7的BLSTM訓練圖	59
圖 38 平均採樣實驗F8的BLSTM訓練圖	59
圖 39各機器學習方法準確率之比對	64

表目次
表 1 知網情緒詞範例表	6
表 2 NTUSD意見詞範例表	6
表 3中文情感分析研究	19
表 4英文情感分析研究	19
表 5 iSGoPaSD情緒詞範例詞	29
表 6 特徵值列表	30
表 7特徵值訓練資料範例	31
表 8 詞向量範例	32
表 9詞向量高相關性範例詞彙	32
表 10 深度學習模型參數表	36
表 11 Padding轉換範例	37
表 12 負向評論字數分布	39
表 13 正向評論字數分布	40
表 14 貝氏演算法預測模型正確率表	45
表 15 使用者評分與預測極性衝突範例	45
表 16 交叉驗證結果分析	46
表 17 實驗硬體配備	47
表 18 非平均採樣資料集的LSTM結果	48
表 19 平均採樣資料集的LSTM結果	50
表 20 非平均採樣資料集的GRU結果	52
表 21 平均採樣資料集的GRU結果	54
表 22非平均採樣資料集的BLSTM結果	56
表 23 平均採樣資料集的BLSTM結果	58
表 24 Sigmoid在各種組合下的結果	61
表 25 Relu在各種組合下的結果	62
表 26 TanH在各種組合下的結果	63
表 27 各機器學習方法準確率比對	64

參考文獻

中文文獻

侯彥旭. (2012). 以評價理論為基礎的電子本文分類. 中原大學.  Available from Airiti AiritiLibrary database. (2012年)
葉庭瑋. (2013). 基於意見詞修飾關係之微網誌情感分析技術. (碩士), 國立臺北科技大學, 台北市.   
蕭瑞祥, 姜青山, 曹金豐, & 陳柏翰. (2015). 基於中文語法規則的情感評價單元抽取方法之研究. 資訊管理學報, 22(3), 243-272. 
簡之文. (2012). 部落格文章情感分析之研究. (碩士), 淡江大學, 新北市.

英文文獻

Agarwal, A., Xie, B., Vovsha, I., Rambow, O., & Passonneau, R. (2011). Sentiment analysis of twitter data. Paper presented at the Proceedings of the workshop on languages in social media.
Andreevskaia, A., & Bergler, S. (2006). Mining WordNet for a Fuzzy Sentiment: Sentiment Tag Extraction from WordNet Glosses. Paper presented at the EACL.
Bottou, L. (2010). Large-scale machine learning with stochastic gradient descent Proceedings of COMPSTAT'2010 (pp. 177-186): Springer.
Bourlard, H. A., & Morgan, N. (2012). Connectionist speech recognition: a hybrid approach (Vol. 247): Springer Science & Business Media.
Campbell, C., & Ying, Y. (2011). Learning with support vector machines. Synthesis lectures on artificial intelligence and machine learning, 5(1), 1-95. 
Chauvin, Y., & Rumelhart, D. E. (1995). Backpropagation: theory, architectures, and applications: Psychology Press.
Chien, L.-F. (1997). PAT-tree-based keyword extraction for Chinese information retrieval. Paper presented at the ACM SIGIR Forum.
Cho, K., Van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014). Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078. 
Chung, J., Gulcehre, C., Cho, K., & Bengio, Y. (2014). Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555. 
Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3), 273-297. 
Dave, K., Lawrence, S., & Pennock, D. M. (2003). Mining the peanut gallery: Opinion extraction and semantic classification of product reviews. Paper presented at the Proceedings of the 12th international conference on World Wide Web.
De Boer, P.-T., Kroese, D. P., Mannor, S., & Rubinstein, R. Y. (2005). A tutorial on the cross-entropy method. Annals of operations research, 134(1), 19-67. 
Domingos, P., & Pazzani, M. (1997). On the optimality of the simple Bayesian classifier under zero-one loss. Machine learning, 29(2-3), 103-130. 
Dong, Z., Dong, Q., & Hao, C. (2006). HowNet and the Computation of Meaning. 
Dong, Z., Dong, Q., & Hao, C. (2010). Hownet and its computation of meaning. Paper presented at the Proceedings of the 23rd International Conference on Computational Linguistics: Demonstrations.
Gamon, M. (2004). Sentiment classification on customer feedback data: noisy data, large feature vectors, and the role of linguistic analysis. Paper presented at the Proceedings of the 20th international conference on Computational Linguistics.
Glorot, X., Bordes, A., & Bengio, Y. (2011). Domain adaptation for large-scale sentiment classification: A deep learning approach. Paper presented at the Proceedings of the 28th International Conference on Machine Learning (ICML-11).
Graves, A., & Schmidhuber, J. (2005). Framewise phoneme classification with bidirectional LSTM and other neural network architectures. Neural networks, 18(5), 602-610. 
Guzman, E., & Maalej, W. (2014). How do users like this feature? a fine grained sentiment analysis of app reviews. Paper presented at the Requirements Engineering Conference (RE), 2014 IEEE 22nd International.
Han, J., Pei, J., & Kamber, M. (2011). Data mining: concepts and techniques: Elsevier.
Hinton, G. E., & Salakhutdinov, R. R. (2006). Reducing the dimensionality of data with neural networks. science, 313(5786), 504-507. 
Hochreiter, S. (1998). The vanishing gradient problem during learning recurrent neural nets and problem solutions. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 6(02), 107-116. 
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735-1780. 
Kangale, A., Kumar, S. K., Naeem, M. A., Williams, M., & Tiwari, M. K. (2015). Mining consumer reviews to generate ratings of different product attributes while producing feature-based review-summary. International Journal of Systems Science, 47(13), 3272-3286. doi:10.1080/00207721.2015.1116640
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59-68. doi:10.1016/j.bushor.2009.09.003
Karpathy, A. (2016). Cs231n: Convolutional neural networks for visual recognition. Online Course. 
Kaur, A., & Duhan, N. (2015). A survey on sentiment analysis and opinion mining. International Journal of Innovations & Advancement in Computer Science, 4, 107-116. 
Kim, Y. (2014). Convolutional neural networks for sentence classification. arXiv preprint arXiv:1408.5882. 
Kohavi, R. (1996). Scaling Up the Accuracy of Naive-Bayes Classifiers: A Decision-Tree Hybrid. Paper presented at the KDD.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Paper presented at the Advances in neural information processing systems.
Lawrence, S., Giles, C. L., Tsoi, A. C., & Back, A. D. (1997). Face recognition: A convolutional neural-network approach. IEEE transactions on neural networks, 8(1), 98-113. 
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444. doi:10.1038/nature14539
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324. 
Lewis, D. D. (1998). Naive (Bayes) at forty: The independence assumption in information retrieval. Paper presented at the European conference on machine learning.
Liu, B. (2012). Sentiment analysis and opinion mining. Synthesis lectures on human language technologies, 5(1), 1-167. 
Liwicki, M., Graves, A., Bunke, H., & Schmidhuber, J. (2007). A novel approach to on-line handwriting recognition based on bidirectional long short-term memory networks. Paper presented at the Proc. 9th Int. Conf. on Document Analysis and Recognition.
McCallum, A., & Nigam, K. (1998). A comparison of event models for naive bayes text classification. Paper presented at the AAAI-98 workshop on learning for text categorization.
McCulloch, W. S., & Pitts, W. (1943). A logical calculus of the ideas immanent in nervous activity. The bulletin of mathematical biophysics, 5(4), 115-133. 
Mihalcea, R., & Liu, H. (2006). A Corpus-based Approach to Finding Happiness. Paper presented at the AAAI Spring Symposium: Computational Approaches to Analyzing Weblogs.
Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781. 
Ng, A. (2000). CS229 Lecture notes. CS229 Lecture notes, 1(1), 1-3. 
Ng, A. Y., & Jordan, M. I. (2002). On discriminative vs. generative classifiers: A comparison of logistic regression and naive bayes. Advances in neural information processing systems, 2, 841-848. 
Nunamaker Jr, J. F., Chen, M., & Purdin, T. D. (1990). Systems development in information systems research. Journal of management information systems, 7(3), 89-106. 
Pang, B., & Lee, L. (2004). A sentimental education: Sentiment analysis using subjectivity summarization based on minimum cuts. Paper presented at the Proceedings of the 42nd annual meeting on Association for Computational Linguistics.
Pang, B., Lee, L., & Vaithyanathan, S. (2002). Thumbs up?: sentiment classification using machine learning techniques. Paper presented at the Proceedings of the ACL-02 conference on Empirical methods in natural language processing-Volume 10.
Poria, S., Cambria, E., & Gelbukh, A. (2016). Aspect extraction for opinion mining with a deep convolutional neural network. Knowledge-Based Systems, 108, 42-49. doi:10.1016/j.knosys.2016.06.009
Rubinstein, R. Y., & Kroese, D. P. (2013). The cross-entropy method: a unified approach to combinatorial optimization, Monte-Carlo simulation and machine learning: Springer Science & Business Media.
Rumelhart, D., Hinton, G., & Williams, R. (1986). Learning internal representation by back propagation. Parallel distributed processing: exploration in the microstructure of cognition, 1. 
Schuster, M., & Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11), 2673-2681. 
Shelke, N. M., Deshpande, S., & Thakre, V. (2012). Survey of techniques for opinion mining. International Journal of Computer Applications, 57(13). 
Sondak, N. E., & Sondak, V. K. (1989). Neural networks and artificial intelligence. Paper presented at the ACM SIGCSE Bulletin.
Specht, D. F. (1991). A general regression neural network. IEEE transactions on neural networks, 2(6), 568-576. 
Tai, K. S., Socher, R., & Manning, C. D. (2015). Improved semantic representations from tree-structured long short-term memory networks. arXiv preprint arXiv:1503.00075. 
Tomar, D., Singhal, S., & Agarwal, S. (2014). Weighted least square twin support vector machine for imbalanced dataset. International Journal of Database Theory and Application, 7(2), 25-36. 
Utgoff, P. E., & Stracuzzi, D. J. (2002). Many-layered learning. Neural Computation, 14(10), 2497-2529. 
Vapnik, V., Golowich, S. E., & Smola, A. (1997). Support vector method for function approximation, regression estimation, and signal processing. Advances in neural information processing systems, 281-287. 
Yih, W.-t., He, X., & Meek, C. (2014). Semantic Parsing for Single-Relation Question Answering. Paper presented at the ACL (2).
Zhao, J., Dong, L., Wu, J., & Xu, K. (2012). Moodlens: an emoticon-based sentiment analysis system for chinese tweets. Paper presented at the Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining.
Zheng, X., Chen, H., & Xu, T. (2013). Deep Learning for Chinese Word Segmentation and POS Tagging. Paper presented at the EMNLP.