摘要
At present, the lawsuit documents provided by the parties during the court filing process still require to be reviewed through manual work. Faced with a large amount of lawsuit documents and limited reviewers, the review is less efficient. This paper proposes an automatic case review system based on deep learning, which can automatically review the lawsuit documents from the parties and identify the type of the case independently, instead of being manually reviewed. And it can recommend similar cases for reference by calculating the text similarity, which can help judges make fair decisions. In this paper, convolutional neural network (CNN), a deep learning model, is used to extract the key features for text classification in lawsuit documents and the features are input into the Softmax layer to classify the case. Then the system matches the case knowledge database to help the parties check for vacancies and provide suggestions, which realizes the automatic case review and improve the review efficiency.
延伸閱讀
- 吳旻聖（2021）。A Multimodal and Time Series based Deep Learning Hotel Recommendation System〔碩士論文，國立清華大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0016-2412202116182381
- Guo, Y., & He, G. (2021). A Review of Image Recognition based on Deep Learning. International Core Journal of Engineering, 7(4), 260-264. https://doi.org/10.6919/ICJE.202104_7(4).0036
- 蔣惟丞（2009）。The Development of a Rapid Machine Learning System〔碩士論文，國立虎尾科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0028-2407200919281700
- 雷承維（2020）。Unsupervised Deep Learning System for AOI Detection〔碩士論文，國立臺灣師範大學〕。華藝線上圖書館。https://doi.org/10.6345/NTNU202001296
- Cao, R. (2020). A Summary of Research on Deep Learning in Time Series Learning Methods. International Core Journal of Engineering, 6(9), 201-208. https://doi.org/10.6919/ICJE.202009_6(9).0024