Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts

Hsuan Lei Shao; Yu Ying Huang; Sieh Chuen Huang

doi:10.1007/978-3-031-36190-6_1

Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts

Hsuan Lei Shao
, Yu Ying Huang
, Sieh Chuen Huang

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Drunk driving cases often arouse public concern in Taiwanese society. According to 2013 amended Paragraph 1 of Article 185–3, the Taiwan Criminal Code, a drunk person may face up to a maximum sentence of two years in prison if his/her exhalation contains alcohol of 0.25 mg per liter or more, or blood alcohol concentration is 0.05% or more. The huge volume of “drunk driving” cases becomes a considerable workload for the court and therefore it may be worthwhile developing an automatic sentencing supportive system. This research attempts to train a deep-learning model to predict sentences by inputting the section of “recidivist/facts of the judgement.” The TextCNN (Convolutional Neural Networks) model reached a 73% accuracy rate in four-category sentencing prediction. This research suggests that adopting two kinds of pre-processing methods and a well-trained model directly to unstructured judgement texts without word segmentation can result in a good performance. It opens the possibility of applying different machine learning techniques to legal texts.

Original language	English
Title of host publication	New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers
Editors	Katsutoshi Yada, Yasufumi Takama, Koji Mineshima, Ken Satoh
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	3-15
Number of pages	13
ISBN (Print)	9783031361890
DOIs	https://doi.org/10.1007/978-3-031-36190-6_1
Publication status	Published - 2023
Externally published	Yes
Event	The 13th International Symposium on Artificial Intelligence supported by the Japanese Society for Artificial Intelligence, JSAI-isAI 2021 - Yokohama, Japan Duration: Nov 13 2021 → Nov 15 2021

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13856 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	The 13th International Symposium on Artificial Intelligence supported by the Japanese Society for Artificial Intelligence, JSAI-isAI 2021
Country/Territory	Japan
City	Yokohama
Period	11/13/21 → 11/15/21

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 3 Good Health and Well-being

Keywords

convolutional neural network
deep learning
drunk driving
sentencing system
Text-classification
TextCNN

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-36190-6_1

Cite this

Shao, H. L., Huang, Y. Y., & Huang, S. C. (2023). Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts. In K. Yada, Y. Takama, K. Mineshima, & K. Satoh (Eds.), New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers (pp. 3-15). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13856 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-36190-6_1

Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts. / Shao, Hsuan Lei; Huang, Yu Ying; Huang, Sieh Chuen.
New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers. ed. / Katsutoshi Yada; Yasufumi Takama; Koji Mineshima; Ken Satoh. Springer Science and Business Media Deutschland GmbH, 2023. p. 3-15 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13856 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Shao, HL, Huang, YY & Huang, SC 2023, Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts. in K Yada, Y Takama, K Mineshima & K Satoh (eds), New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13856 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 3-15, The 13th International Symposium on Artificial Intelligence supported by the Japanese Society for Artificial Intelligence, JSAI-isAI 2021, Yokohama, Japan, 11/13/21. https://doi.org/10.1007/978-3-031-36190-6_1

Shao HL, Huang YY, Huang SC. Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts. In Yada K, Takama Y, Mineshima K, Satoh K, editors, New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers. Springer Science and Business Media Deutschland GmbH. 2023. p. 3-15. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-36190-6_1

Shao, Hsuan Lei ; Huang, Yu Ying ; Huang, Sieh Chuen. / Prediction Model for Drunk Driving Sentencing : Applying TextCNN to Chinese Judgement Texts. New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers. editor / Katsutoshi Yada ; Yasufumi Takama ; Koji Mineshima ; Ken Satoh. Springer Science and Business Media Deutschland GmbH, 2023. pp. 3-15 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d3682da1569443aea39953d9bbf9feac,

title = "Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts",

abstract = "Drunk driving cases often arouse public concern in Taiwanese society. According to 2013 amended Paragraph 1 of Article 185–3, the Taiwan Criminal Code, a drunk person may face up to a maximum sentence of two years in prison if his/her exhalation contains alcohol of 0.25 mg per liter or more, or blood alcohol concentration is 0.05\% or more. The huge volume of “drunk driving” cases becomes a considerable workload for the court and therefore it may be worthwhile developing an automatic sentencing supportive system. This research attempts to train a deep-learning model to predict sentences by inputting the section of “recidivist/facts of the judgement.” The TextCNN (Convolutional Neural Networks) model reached a 73\% accuracy rate in four-category sentencing prediction. This research suggests that adopting two kinds of pre-processing methods and a well-trained model directly to unstructured judgement texts without word segmentation can result in a good performance. It opens the possibility of applying different machine learning techniques to legal texts.",

keywords = "convolutional neural network, deep learning, drunk driving, sentencing system, Text-classification, TextCNN",

author = "Shao, \{Hsuan Lei\} and Huang, \{Yu Ying\} and Huang, \{Sieh Chuen\}",

note = "Publisher Copyright: {\textcopyright} 2023, Springer Nature Switzerland AG.; The 13th International Symposium on Artificial Intelligence supported by the Japanese Society for Artificial Intelligence, JSAI-isAI 2021 ; Conference date: 13-11-2021 Through 15-11-2021",

year = "2023",

doi = "10.1007/978-3-031-36190-6\_1",

language = "English",

isbn = "9783031361890",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "3--15",

editor = "Katsutoshi Yada and Yasufumi Takama and Koji Mineshima and Ken Satoh",

booktitle = "New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers",

address = "Germany",

}

TY - GEN

T1 - Prediction Model for Drunk Driving Sentencing

T2 - The 13th International Symposium on Artificial Intelligence supported by the Japanese Society for Artificial Intelligence, JSAI-isAI 2021

AU - Shao, Hsuan Lei

AU - Huang, Yu Ying

AU - Huang, Sieh Chuen

PY - 2023

Y1 - 2023

N2 - Drunk driving cases often arouse public concern in Taiwanese society. According to 2013 amended Paragraph 1 of Article 185–3, the Taiwan Criminal Code, a drunk person may face up to a maximum sentence of two years in prison if his/her exhalation contains alcohol of 0.25 mg per liter or more, or blood alcohol concentration is 0.05% or more. The huge volume of “drunk driving” cases becomes a considerable workload for the court and therefore it may be worthwhile developing an automatic sentencing supportive system. This research attempts to train a deep-learning model to predict sentences by inputting the section of “recidivist/facts of the judgement.” The TextCNN (Convolutional Neural Networks) model reached a 73% accuracy rate in four-category sentencing prediction. This research suggests that adopting two kinds of pre-processing methods and a well-trained model directly to unstructured judgement texts without word segmentation can result in a good performance. It opens the possibility of applying different machine learning techniques to legal texts.

AB - Drunk driving cases often arouse public concern in Taiwanese society. According to 2013 amended Paragraph 1 of Article 185–3, the Taiwan Criminal Code, a drunk person may face up to a maximum sentence of two years in prison if his/her exhalation contains alcohol of 0.25 mg per liter or more, or blood alcohol concentration is 0.05% or more. The huge volume of “drunk driving” cases becomes a considerable workload for the court and therefore it may be worthwhile developing an automatic sentencing supportive system. This research attempts to train a deep-learning model to predict sentences by inputting the section of “recidivist/facts of the judgement.” The TextCNN (Convolutional Neural Networks) model reached a 73% accuracy rate in four-category sentencing prediction. This research suggests that adopting two kinds of pre-processing methods and a well-trained model directly to unstructured judgement texts without word segmentation can result in a good performance. It opens the possibility of applying different machine learning techniques to legal texts.

KW - convolutional neural network

KW - deep learning

KW - drunk driving

KW - sentencing system

KW - Text-classification

KW - TextCNN

UR - https://www.scopus.com/pages/publications/85172413111

UR - https://www.scopus.com/pages/publications/85172413111#tab=citedBy

U2 - 10.1007/978-3-031-36190-6_1

DO - 10.1007/978-3-031-36190-6_1

M3 - Conference contribution

AN - SCOPUS:85172413111

SN - 9783031361890

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 3

EP - 15

BT - New Frontiers in Artificial Intelligence - JSAI-isAI 2021 Workshops, JURISIN, LENLS18, SCIDOCA, Kansei-AI, AI-BIZ, Revised Selected Papers

A2 - Yada, Katsutoshi

A2 - Takama, Yasufumi

A2 - Mineshima, Koji

A2 - Satoh, Ken

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 13 November 2021 through 15 November 2021

ER -

Prediction Model for Drunk Driving Sentencing: Applying TextCNN to Chinese Judgement Texts

Abstract

Publication series

Conference

UN SDGs

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this