Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Building and Real Estate | en_US |
| dc.contributor.advisor | Chi, Hung-lin (BRE) | en_US |
| dc.creator | Lin, Haolei | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/12652 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Capabilities and limitations of ChatGPT in classifying crane-related accident reports | en_US |
| dcterms.abstract | Cranes hold an indispensable and vital role in the realm of construction engineering. Efficiently summarizing and managing the causal factors behind crane-related accidents can significantly mitigate unnecessary hazards. This study focuses on the application of ChatGPT in the classification of crane-related accident reports, summarizing the fundamental framework of utilizing the Prompt for ChatGPT for conducting classification tasks. The aim is to diminish manual labeling time for researchers while constructing their own deep learning models, and concurrently enhance the accuracy of ChatGPT's classification abilities. | en_US |
| dcterms.abstract | This research encompasses two distinct phases. The initial phase entails a preliminary data analysis of crane-related incident reports from 2002 to 2021, sourced from the Occupational Safety and Health Administration (OSHA). Subsequently, ChatGPT is applied to conduct classification experiments, with only the basic requirements of the Prompt being edited. The evaluation metric employed is the F1-Score, and the Prompt improvement focuses on addressing two specific limitations of ChatGPT: the difficulty in classifying due to vague classification standards, and high levels of expertise in labeling. | en_US |
| dcterms.abstract | In the second phase, iterative experiments are carried out by using ChatGPT repeatedly refining and adjusting the Prompts. Ultimately, the study concludes that transforming a single classification task into a composite of multiple classification tasks, endowing ChatGPT with an assumed identity, and allocating adequate computational time are effective methods within the framework for crafting Prompts that enhance the accuracy of ChatGPT's classification performance. | en_US |
| dcterms.extent | iv, 35 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2023 | en_US |
| dcterms.educationalLevel | M.Sc. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.LCSH | Cranes, derricks, etc -- Accidents | en_US |
| dcterms.LCSH | Natural language generation (Computer science) -- Computer programs | en_US |
| dcterms.LCSH | Artificial intelligence | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 7117.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 1.15 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/12652