Relation extraction from medical texts using semi­supervised learning

Name: Siriwon Taewijit

keyword: Adverse drug reaction (ADR)

; Medical text mining

; Distant supervision

; Multiple­instance learning (MIL)

; Relation extraction

; Semi­supervised learning

; Transductive inference

Abstract: Information extraction and knowledge discovery from large­scale clinical texts are beneficial and needy processes. In dealing with text mining, text preprocessing is the additional task to transform unstructured textual data into a structured one. Fortunately, the traditional data mining process can further utilize this machine­readable format. Two significant difficulties of relation extraction are the lack of domain experts for labeling examples, exceptionally, in a large volume of unlabeled data, mainly clinical texts. This issue is well-known when it deals with the supervised learning method. Another challenge is the intractable processing of unstructured clinical texts. This dissertation, firstly, studies how to handle textual data in the large­scale texts. Generally speaking, the thesis aims to propose an automatic leveraging the existing knowledge to overcome shortcoming the lack of domain expert for data labeling. Secondly, the thesis proposes an efficient method for relation classification which aims to reveal semantic relation label of the given entity pairs from clinical texts. To this end, the dissertation presents a semi­supervised learning framework as a solution for semantic relation extraction from texts, herein, clinical textual data. Furthermore, the dissertation introduces efficient parameters estimation for a generative model that argues the traditional text assumption. This contribution can dramatically improve a model performance. Lastly, the thesis contributes to examine on multiple approaches of unlabeled data augmentation in order to be able to control large­scale of unlabeled data with the effectiveness

Thammasat University. Thammasat University Library

Address: BANGKOK

Email: preserv@tu.ac.th

Name: Thanaruk Theeramunkong

Role: advisor

Name: Ikeda, Mitsuru

Role: co-advisor

Created: 2020

Modified: 2022-11-04

Issued: 2022-11-04

วิทยานิพนธ์/Thesis

application/pdf

DOI: 10.14457/TU.the.2020.66

eng

DegreeName: Doctor of Philosophy

Level: Doctoral Degree

Descipline: Engineering and Technology

Grantor: Thammasat University

©copyrights Thammasat University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	9260siriwon.pdf	1.6 MB

ใช้เวลา

0.050508 วินาที