:::

詳目顯示

回上一頁
題名:應用卷積神經網路深度學習方法作電子書分類研究
作者:林世陽
作者(外文):Lin, Shih-Yang
校院名稱:國立高雄科技大學
系所名稱:商業智慧學院博士班
指導教授:柯博昌
柯國民
學位類別:博士
出版日期:2022
主題關鍵詞:深度學習卷積神經網路電子書分類Deep learningConvolutional Neural NetworkE-Book Classification
原始連結:連回原系統網址new window
相關次數:
  • 被引用次數被引用次數:期刊(0) 博士論文(0) 專書(0) 專書論文(0)
  • 排除自我引用排除自我引用:0
  • 共同引用共同引用:0
  • 點閱點閱:3
摘要
隨著時代的演進,人類社會經歷了無數的改變,當中不變的是人類一直以來嘗試利用不同的方法保留知識並流傳至後代,而書本在其中佔有極重要的地位。為了能讓讀者更快選出所需書本,準確的分類必不可少。但用人工的方式分類沒有效率且有機會因為認知不一樣產生誤判。
在目前電子書自動分類的研究中,已分類的書本中得到書名、關鍵詞與摘要全部詞語來產生書本的特徵矩陣,再對未分類書本用相同預處理手法產生特徵矩陣作為輸入並比較,此方法依賴已分類書本的特徵矩陣內容,若未分類書本中有重要的特徵詞語而找不到對應已分類書本產生的特徵詞語,分類器就很有可能誤判。另外亦有使用書名及目錄產生文件向量 (document vector)待徵的技術,並以詞頻的方式從未分類電子書中的書名及簡介抽取關鍵字,再透過 word2vec 轉換出可以融合至書本封面圖片的特徵矩陣。圖文特徵融合後,再利用深度學習的卷積神經網路(CNN)來訓練,在圖片辨識上用卷積神經網路非常有優勢,有效減少了訓練的時間同時又提高了準確率。
實體書與電子書的分類一樣得依靠特徵,將影像轉換為數據後,利用演算法,搭配多層次的卷積運算,包括一個或數個卷積層與池化層之組合提取特徵,再以完全連接方式將特徵資料輸入至神經網路進行訓練,進而達成更高的影像辨識。
關鍵字:深度學習、卷積神經網路、電子書分類
Abstract
As time progress, humans have achieved numerous success, one of them being innovating methods to pass on knowledge, such as books. In this age of information, there will only be more and more books being published or store on the Internet. In order to allow readers to choose their reading materials with precision, this research aims to utilize deep learning to classify existing Chinese literatures.
Modern day ebooks are usually classified through their text features, but ebooks also contain graphs and pictures that could be used as features. This research utilize the cover of ebooks as a feature and classify them based on term frequency, title, and summary. This research also utilize Word2vec (published by Google) to calculate and formulate a picture for deep learning to process and classify. After fusing images and texts, Convolutional Neural Network(CNN) will train the dataset to classify ebooks. CNN provides a superior method to train datasets, it reduces the time require to train dataset and increases the precision of such training.
By utilizing the strong feature extraction function in the convolutional neural network (CNN), we are able to extract automatically contours, edge lines, and local features layer by layer through filters. We then fully connect the data and input them to the neural network and train the CNN. Due to the effectiveness of image recognition, it is easy to find recurring features and classify the differences.
Keywords: Deep learning; Convolutional Neural Network; E-Book Classification
References

1.Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C.,...& Zheng, X. (2016). Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467.
2.Audebert, N., Herold, C., Slimani, K., & Vidal, C. (2019). Multimodal deep networks for text and image-based document classification. arXiv preprint arXiv:1907.06370.
3.Audebert, N., Herold, C., Slimani, K., & Vidal, C. (2019). Multimodal deep networks for text and image-based document classification. arXiv preprint arXiv:1907.06370.
4.Bengio, Y., Goodfellow, I., & Courville, A. (2017). Deep learning (Vol. 1). Massachusetts, USA:: MIT press.
5.Bouvrie, J. (2006). Notes on convolutional neural networks.
6.Bouvrie, J. (2006). Notes on convolutional neural networks.
7.Chatterjee, S., Jose, P. G., & Datta, D. (2019). Text Classification Using SVM Enhanced by Multithreading and CUDA. International Journal of Modern Education & Computer Science, 11(1).
8.Chin, S. H., Chen, C., Ko, P. C., & Lin, S. Y. (2019). Design of museum advertisement picture management system based on web. Journal of Visual Communication and Image Representation, 63, 102595.
9.Fukushima, K., & Miyake, S. (1982). Neocognitron: A self-organizing neural network model for a mechanism of visual pattern recognition. In Competition and cooperation in neural nets, 267-285.
10.Gallo, I., Calefati, A., Nawaz, S., & Janjua, M. K. (2018). Image and encoded text fusion for multi-modal classification. In 2018 Digital Image Computing: Techniques and Applications (DICTA) (pp. 1-7). IEEE.
11.Ko, K. M., Ko, P. C., Lin, S. Y., & Hong, Z. (2019). Quality-guided image classification toward information management applications. Journal of Visual Communication and Image Representation, 63, 102594.
12.Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 1097-1105.
13.Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25, 1097-1105.
14.LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989). Backpropagation applied to handwritten zip code recognition. Neural computation, 1(4), 541-551.
15.Lin, S. Y., Du, Y., Ko, P. C., Wu, T. J., Ho, P. T., & Sivakumar, V. (2020). Fog Computing Based Hybrid Deep Learning Framework in effective inspection system for smart manufacturing. Computer Communications, 160, 636-642.
16.Rosenblatt, F. (1958). The perceptron: a probabilistic model for information storage and organization in the brain. Psychological review, 65(6), 386.
17.Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
18.Uijlings, J. R., Van De Sande, K. E., Gevers, T., & Smeulders, A. W. (2013). Selective search for object recognition. International journal of computer vision, 104(2), 154-171.
19.Weng, J., Ahuja, N., & Huang, T. S. (1992, June). Cresceptron: a self-organizing neural network which grows adaptively. In [Proceedings 1992] IJCNN International Joint Conference on Neural Networks (Vol. 1, pp. 576-581). IEEE.
20.Guo, L. M., Liu, W., Wu, P. C., Zhang, L., “Machine Learning and Its application in Library: Take TensorFlow as an Example”, Journal of Academic Libraries, (6), 2017.
21.Guo, J. J., “The Study of Automatic Book Classification Using Majority Vote Strategy”, Journal of Library and Information Science Research 13: 1 (December 2018): 87-124, 2018.

 
 
 
 
第一頁 上一頁 下一頁 最後一頁 top