| บทคัดย่อ |
งานวิจัยนี้มีวัตถุประสงค์ 1) เพื่อสร้างองค์ความรู้การนำสมุนไพรไปใช้ประโยชน์ของชุมขนบ้านวังกวาง ตำบลวังกวาง อำเภอน้ำหนาว จังหวัดเพชรบูรณ์ และ 2) เพื่อพัฒนาระบบสารสนเทศเพื่อการเรียนรู้ความหลากหลายของสมุนไพรและรูปแบบการใช้ประโยชน์สมุนไพรส่งเสริมสุขภาพในรูปแบบสื่ออินเทอร์แอคทีฟมัลติมีเดีย การวิจัยนี้เป็นการวิจัยและพัฒนา โดยมีกระบวนการวิจัยแบ่งเป็น 4 ขั้นตอน ได้แก่ การศึกษารวบรวมข้อมูลสมุนไพร การออกแบบและพัฒนาระบบ การประเมินประสิทธิภาพระบบ และการประเมินความพึงพอใจผู้ใช้ เครื่องมือการวิจัย ได้แก่ เว็บไซต์นำเสนอองค์ความรู้สมุนไพรไทย จำนวน 100 แหล่ง ทำการวัดประสิทธิภาพในการสกัดความรู้เชิงความหมาย และทำนายเพื่อจำแนกชื่อสมุนไพร ด้วยอัลกอริทึมการเรียนรู้ของเครื่อง ได้แก่ นิวรอลเนตเวิร์ต เอสวีเอ็ม เคเนียเรสเนสเบอร์ และแผนผังต้นไม้เพื่อการตัดสินใจ การสร้างออนโทโลยรด้วยโปรแกรมโปรทีเจ ภาษาที่ใช้พัฒนาโปรแกรม ได้แก่ ภาษาไพธอน พีเอชพี และจาวาสคริปต์ ผลการวิจัยพบว่า ระบบสารสนเทศประกอบด้วย 4 ฟังก์ชัน ได้แก่ การจัดการข้อมูลสมุนไพร การจัดการข้อมูลปราชญ์ชาวบ้าน/ผู้เชี่ยวชาญ การสืบค้นข้อมูลสมุนไพร และรายงานสถิติการสืบค้นข้อมูลสมุนไพร ออนโทโลยีสมุนไพรไทย แบ่งเป็น 2 ชั้นความรู้ และมีโหนดความรู้จำนวนทั้งหมด 8 โหนด กระบวนการเรียนรู้ของเครื่องบนเหมืองความรู้เชิงความหมายสมุนไพรไทย ประกอบด้วย 4 ส่วน ได้แก่ การประมวลผลเชิงความหมาย การสกัดเหมืองความรู้ การสร้างชุดข้อมูลสมุนไพร และ การทำนายผลลัพธ์ด้วยการเรียนรู้ของเครื่อง ผลการประเมินประสิทธิภาพในการค้นคืนความรู้เชิงความหมายสมุนไพรไทยโดยใช้ออนโทโลยี และกฎSWRL ได้ค่า F-measure เท่ากับ 87.9% และประสิทธิภาพในการทำนาย เพื่อจำแนกคลาสคำตอบสมุนไพร ด้วยอัลกอริทึม NN ได้ค่าความถูกต้องสูงที่สุด เท่ากับ 90.0% และผู้ใช้มีความพึงพอใจต่อระบบในภาพรวมอยู่ในระดับมาก
This research aimed 1) to create the herb knowledge for utllalizating of
Wang Kwang Sub-district, Nam Nao District Phetchabun Province 2) to develop the information system for learning diverse herb knowledge in a form of interactive media.This research is based on research and development method. The research process was classified in 4 steps; studying and collecting the herb knowledge, designing and developing the herb information system (HIS), assessing the HIS’s performance, and assessing the users’ satisfaction to HIS. The research instruments were 100 Thai herb knowledge websites, measuring the performance in extracting the semantic knowledge and predicting for classifying the herb classes by the neural network (NN), the support vector machine (SVM), the K-nearest neighbor (KNN), and the decision tree (DT), and creating the herb ontology by Protégé. The system was developed by Python, PHP, and Javascript. This research found that HIS consisted of 4 functions are; managing the herb knowledge, herb knowledgeable persons and specialists, searching the herb knowledge, and displaying the statistical report in searching herb knowledge from users. The herb ontology was divided into 2 knowledge classes and it has 8 knowledge nodes. The processing of ML on the semantic Thai herb knowledge mining consisted of 4 parts are the semantic processing, extracting the knowledge mining, creating the Thai herb dataset, and predicting the results by the ML. The results of assessing the semantic Thai herb knowledge retrieval performance by the ontology and the SWRL rules has the F-measure value at 87.9%. The NN has the highest Thai herb classes prediction accuracy at 90.0%. The users have satisfaction to HIS in overview at high level.
|
| คำสำคัญ |
ระบบสารสนเทศ,เหมืองความรู้,ออนโทโลยี,การเรียนรู้ของเครื่อง,สมุนไพรไทย,Information System,Knowledge Mining,Ontology,Machine Learning,Thai Herb
|