研究生: |
姚道志 Tao-Chih Yao |
---|---|
論文名稱: |
建立以小樣本時間序列網路聲量資料為基之汽車銷量趨勢預測模型 Establish Forecast Models of Automobile Sales Trends based on Small Samples Time Series Network Volume Data |
指導教授: |
歐陽超
Chao Ou-Yang |
口試委員: |
歐陽超
Chao Ou-Yang 王孔政 Kung-Jeng Wang 郭人介 Ren-Jieh Kuo |
學位類別: |
碩士 Master |
系所名稱: |
管理學院 - 工業管理系 Department of Industrial Management |
論文出版年: | 2020 |
畢業學年度: | 108 |
語文別: | 中文 |
論文頁數: | 67 |
中文關鍵詞: | 網路聲量資料 、小樣本 、CNN 、LSTM 、子集合選擇 |
外文關鍵詞: | voice volume data, small samples, CNN, LSTM, Subset Selection |
相關次數: | 點閱:113 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
隨著近年網路使用者的大量增加,許多人在做出相對高價之消費決策時經常會選擇網路討論與瀏覽推文的方式做為消費的參考依據。對廠商而言,運用網路做行銷已經是主要的行銷通路之一,網路的宣傳與網民意見可以視為左右名聲的重要指標。尤其是高價之消費品如汽車,因其品牌競爭與車型種類的繁多,在網路上討論以協助購買決策已是常見的模式。特別是國內網路文章中關於汽車的討論一直是相當活躍的群體。
本研究基於上述理由,利用汽車品牌於國內網路聲量資料,進行預測並判斷銷售量趨勢,並希望運用這項預測資訊,可以提供生產與進口的廠商作為行銷與生產管理方面之參考,進而促進汽車產業之發展。
另外,本研究所採用之網路聲量資料來源,係國內知名之網路聲量分析公司,其提供之資訊在國內受到廣泛使用。本研究除用該公司的汽車品牌網路總聲量資料,將採用該公司標註之正聲量、負聲量做為情緒資料變數,以提高模型判斷準確率。而預測目標將使用交通部公路總局網站所公開之每月汽車掛牌數,並轉換成趨勢上升、保持或下跌標籤。
本研究所採用的網路聲量情緒資料具有時間序列資料之特性,所以在預測模型方面將使用一維卷積神經網絡(1 Dimension Convolution Neural Networks, 1D CNN),以及長短期記憶網絡(Long Short Term Memory Network, LSTM)作為預測模型的主軸,並加入CNN與LSTM的複合模型,對架構進行測試與研究,活用這類深度學習模型之特性並提供實務上的案例,建立針對個案資料最適配之模型架構。同時,本研究屬於現實的小樣本時間序列資料,在方法與研究上會針對其特殊的資料特性選配,期待會對本研究之模型預測有所幫助。
Applying network opinions as a reference to shopping becomes popular in recent years. Many social networks medias such as discussion forums and blogs growth rapidly. More and more businesses have utilized social networks as their marketing strategies. The market for automobile sales is one of the areas. People will go to various social media to search for related information before making the final decision. Therefore, this research tends to apply network opinions from various social media to predict the trends of automobile sales.
This research will use the positive/negative emotion and total internet volume for six brands of automobiles and their automobiles sales data as the dataset. The time series of internet volume data are collected and analyzed every day and car sales data are recorded every month.
The proposed approach including several steps. First, convert car sales data into a trend up, hold, or down labels. Second, use genetic algorithms to reduce the number of variables to facilitate model training. Then a hybrid “One Dimension Convolution Neural Networks (1D CNN)” with “Long Short Term Memory Network (LSTM)” will be used to predict the sales trends. Finally, compare the prediction results with LSTM and 1D CNN models. The predicting results can be used as a reference for the marking and production departments of local automobile companies.
Chong, A. Y. L., Li, B. Y., Ngai, E. W. T., Ch'ng, E., & Lee, F. (2016). Predicting online product sales via online reviews, sentiments, and promotion strategies A big data architecture and neural network approach. International Journal of Operations & Production Management, 36(4), 358-+. Retrieved from <Go to ISI>://WOS:000379677500001. doi:10.1108/Ijopm-03-2015-0151
Fan, Z.-P., Che, Y.-J., & Chen, Z.-Y. (2017). Product sales forecasting using online reviews and historical sales data: A method combining the Bass model and sentiment analysis. Journal of Business Research, 74, 90-100.
Feldman, R. (2013). Techniques and applications for sentiment analysis. Communications of the Acm, 56(4). doi:10.1145/2436256.2436274
Hülsmann, M., Borscheid, D., & Friedrich, C. M. (2012). General sales forecast models for automobile markets and their analysis. Transactions on Machine Learning and Data Mining, 5 (2012)(2). Retrieved from https://opus.bsz-bw.de/fhdo/frontdoor/index/index/docId/414.
Huete-Alcocer, N. (2017). A Literature Review of Word of Mouth and Electronic Word of Mouth: Implications for Consumer Behavior. Front Psychol, 8, 1256. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/28790950. doi:10.3389/fpsyg.2017.01256
Karegowda, A. G., Jayaram, M., & Manjunath, A. (2010). Feature subset selection problem using wrapper approach in supervised learning. International journal of Computer applications, 1(7), 13-17.
Kirange, D. K., & Deshmukh, R. R. (2016). Sentiment Analysis of News Headlines for Stock Price Prediction.
Kiranyaz, S., Avci, O., Abdeljaber, O., Ince, T., Gabbouj, M., & Inman, D. J. (2019). 1D Convolutional Neural Networks and Applications: A Survey. arXiv preprint arXiv:1905.03554.
Kohavi, R., & John, G. H. (1997). Wrappers for feature subset selection. Artificial intelligence, 97(1-2), 273-324.
LeCun, Y., & Bengio, Y. (1998). Convolutional networks for images, speech, and time series. In A. A. Michael (Ed.), The handbook of brain theory and neural networks (pp. 255-258): MIT Press.
Liu, C. L., Hsaio, W. H., & Tu, Y. C. (2019). Time Series Classification With Multivariate Convolutional Neural Network. IEEE Transactions on Industrial Electronics, 66(6), 4788-4797. Retrieved from <Go to ISI>://WOS:000457598700061. doi:10.1109/Tie.2018.2864702
Mudinas, A., Zhang, D., & Levene, M. (2019). Market Trend Prediction using Sentiment Analysis: Lessons Learned and Paths Forward. doi:10.1145/nnnnnnn.nnnnnnn
Nasukawa, T., & Yi, J. (2003). Sentiment Analysis: Capturing Favorability Using Natural Language Processing. Paper presented at the K-CAP '03 Proceedings of the 2nd International Conference on Knowledge Capture.
OpView 社群口碑資料庫. Retrieved from https://www.opview.com.tw/
Parekh, R. G., Yang, J., & Honavar, V. (1997). Constructive neural network learning algorithms for multi-category real-valued pattern classification.
Sosa, P. M. (2017). Twitter Sentiment Analysis using Combined LSTM-CNN Models. In.
Stevenson, W. J. (2015). Operations Management, 12e.
Swapna, G., Soman, K. P., & Vinayakumar, R. (2018). Automated detection of diabetes using CNN and CNN-LSTM network and heart rate signals. Procedia Computer Science, 132, 1253-1262. Retrieved from http://www.sciencedirect.com/science/article/pii/S1877050918307737. doi:https://doi.org/10.1016/j.procs.2018.05.041
Thiesing, F. M., & Vornberger, O. (1997, 12-12 June 1997). Sales forecasting using neural networks. Paper presented at the Proceedings of International Conference on Neural Networks (ICNN'97).
Wang, F.-K., Chang, K.-K., & Tzeng, C.-W. (2011). Using adaptive network-based fuzzy inference system to forecast automobile sales. Expert Systems with Applications, 38(8), 10587-10593. doi:10.1016/j.eswa.2011.02.100
Winters, P. R. (1960). Forecasting Sales by Exponentially Weighted Moving Averages. Management Science, 6(3), 324-342. Retrieved from <Go to ISI>://WOS:A1960CKL0800008. doi:DOI 10.1287/mnsc.6.3.324
Yang, J., Parekh, R., & Honavar, V. (1999). DistAl: An inter-pattern distance-based constructive learning algorithm. Intelligent Data Analysis, 3(1), 55-73.
Yenter, A., & Verma, A. (2017). Deep CNN-LSTM with combined kernels from multiple branches for IMDB Review Sentiment Analysis. Paper presented at the 2017 IEEE 8th Annual Ubiquitous Computing, Electronics and Mobile Communication Conference (UEMCON).
Zhang, G. P. (2003). Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, 50, 159-175. Retrieved from <Go to ISI>://WOS:000180567700009. doi:10.1016/S0925-2312(01)00702-0
Zhao, B. D., Lu, H. Z., Chen, S. F., Liu, J. L., & Wu, D. Y. (2017). Convolutional neural networks for time series classification. Journal of Systems Engineering and Electronics, 28(1), 162-169. Retrieved from <Go to ISI>://WOS:000396969200018. doi:10.21629/Jsee.2017.01.18
交通部統計查詢網. Retrieved from https://stat.motc.gov.tw/mocdb/stmain.jsp?sys=100
林育琦. (2016). 數位足跡巨量資料應用之網路聲量對台灣上市公司股價漲跌趨勢之影響性研究.
許馨予. (2016). 股票價格波動性與新聞情緖分析之關聯性研究. 許馨予,