ระบบจัดเก็บและค้นหางานวิจัยอัจฉริยะ

The Development a Job Position Recommendation System Using Collaborative Filtering Technique

ผศ.ดร.ณัฐรฐนนท์ กานต์รวีกุลธนา | Autoencoder Collaborative filtering Job position Matrix factorization

จำนวนผู้เข้าชม

The Development a Job Position Recommendation System Using Collaborative Filtering Technique

Collection	RMUTK Research Repository (RMUTK IR) วารสารวิชาการ — e-Journal Articles
ID RMUTK Digital	RMUTK000137
Title	The Development a Job Position Recommendation System Using Collaborative Filtering Technique
Alternative title	The Development a Job Position Recommendation System Using Collaborative Filtering Technique
Authors	ผศ.ดร.ณัฐรฐนนท์ กานต์รวีกุลธนา ผู้แต่งหลัก อ.ดร.จินตนา พลศรี อ.ภัททิรา แก้วเกิด ผศ.ดร.ภูริวัตร คัมภีรภาพพัฒน์ อ.รุ่งทิพย์ โคบาล Corresponding
Faculty	คณะบริหารธุรกิจ
Journal Title	Science & Technology Asia (STA)
ISSN	2586-9027
Volume / Issue / Pages	Vol.31 \| No.3 \| pp.147–159
Published	2026-09-30
Year	2569
Level	ระดับนานาชาติ (SCOPUS)
Quartile	Q3
DOI	https://ph02.tci-thaijo.org/index.php/SciTechAsia/article/view/255265
Abstract	Unemployed individuals and recent graduates often lack the experience or skills necessary to effectively search and filter job positions on job search websites, hindering their ability to access job listings and requirements. To address this issue, a research process was conducted involving extracting job information from the top job search website in Thailand. The data was then prepared, transformed, and subjected to Matrix Factorization. The data was randomly split into 80% for training and 20% for testing, out of a total of 30,000 records. This data was used to test the performance of a recommendation system utilizing the Collaborative Filtering technique. The Learning Curve, which measures the efficiency of model learning through Training Loss and Validation Loss, showed a continuous decrease in value until it reached its lowest validation loss of 0.0442, indicating a well-fitted model. This research suggests that the model can be effectively used for predicting and creating a prototype of a job recommendation system that connects job opportunities via API to efficiently meet the needs of job seekers and recent graduates.
Abstract (EN)	Unemployed individuals and recent graduates often lack the experience or skills necessary to effectively search and filter job positions on job search websites, hindering their ability to access job listings and requirements. To address this issue, a research process was conducted involving extracting job information from the top job search website in Thailand. The data was then prepared, transformed, and subjected to Matrix Factorization. The data was randomly split into 80% for training and 20% for testing, out of a total of 30,000 records. This data was used to test the performance of a recommendation system utilizing the Collaborative Filtering technique. The Learning Curve, which measures the efficiency of model learning through Training Loss and Validation Loss, showed a continuous decrease in value until it reached its lowest validation loss of 0.0442, indicating a well-fitted model. This research suggests that the model can be effectively used for predicting and creating a prototype of a job recommendation system that connects job opportunities via API to efficiently meet the needs of job seekers and recent graduates.
Keywords (TH)	Autoencoder Collaborative filtering Job position Matrix factorization Recommendation system
Keywords (EN)	Autoencoder Collaborative filtering Job position Matrix factorization Recommendation system
Access Level	Open Access (เปิดสาธารณะ)

ลิงก์แหล่งเผยแพร่

เอกสารฉบับเต็ม (ลิงก์ภายนอก)

https://ph02.tci-thaijo.org/index.php/SciTechAsia/article/view/255265

เปิดลิงก์