| บทคัดย่อ |
การวิจัยมีวัตถุประสงค์ เพื่อนำเสนอ และเปรียบเทียบช่วงความเชื่อมั่นสำหรับประมาณค่าเฉลี่ยประชากร ภายใต้การสุ่มตัวอย่างแบบใช้ความน่าจะในการเลือกไม่เท่ากัน และไม่ใส่คืน รวมถึงสัดส่วนตัวอย่างมีขนาดเล็ก ช่วงความเชื่อมั่นที่ผู้วิจัยนำเสนอประกอบด้วย ช่วงความเชื่อมั่นแบบเชิงเส้น และช่วงความเชื่อมั่นแบบไม่เป็นเชิงเส้น ผู้วิจัยเปรียบเทียบประสิทธิภาพของช่วงความเชื่อมั่นที่นำเสนอ โดยใช้เทคนิคมอนติคาร์โลซิมูเลชั่น และใช้ขนาดตัวอย่างทั้งหมด 5 ระดับ ประกอบด้วย 30, 35, 40, 45, และ 50 เกณฑ์ที่ใช้ในการพิจารณาประสิทธิภาพช่วงความเชื่อมั่นที่นำเสนอได้แก่ ความกว้างเฉลี่ยของช่วงความเชื่อมั่นที่ระดับความเชื่อมั่น เท่ากับ 90% 95% และ 99% ผลการเปรียบเทียบประสิทธิภาพพบว่าช่วงความเชื่อมั่นแบบไม่เป็นเชิงเส้น มีประสิทธิภาพดีกว่าช่วงความเชื่อมั่นที่ไม่เป็นเชิงเส้น
In this research paper, we aim to propose and compare confidence interval estimators for estimating population mean under unequal probability sampling without replacement and overall negligible sampling fraction. The propose estimators comprise linear and ratio confidence interval. We compare the efficiency of the propose estimators by using mote carlo simulation with 5 levels sample size there are 30, 35, 40, 45 and 50. The considered criterions are based on expected length of confidence interval with 90%, 95% and 99% confidence levels. The simulation study show that for all sample size and confidence levels the nonlinear confidence interval is more efficiency than linear confidence interval.
|
| คำสำคัญ |
ตัวประมาณค่า,สัดส่วนตัวอย่าง,ความแปรปรวน,Estimator,Sampling fraction,Variance
|