Data mining for the demand-based problem and holt's forecasting with events

Name: Thanrada Chaikajonwat

ThaSH: Data mining

ThaSH: Data mining -- Statistical metheds

Abstract: This thesis proposes modified data mining techniques for addressing the location problem and sales forecasting with special events. First, the typical K-means clustering algorithm is applied and also modified to find optimal locations for distribution centers for a Thailand-based convenience store franchise. Three clustering approaches using different distance metrics, namely, Euclidean, Manhattan, Chebyshev, along with their three modified ones that incorporate the demands, are compared for effectiveness and efficiency. The Weighted Chebyshev approach offers the best results in terms of expected distribution cost and the Davies-Bouldin index, while Euclidean is the most efficient. เท addition, this research proposes three modified Holt's-based methods to better forecast time-series data affected by events, such as the COVID-19 pandemic, on Thailand's automotive industry. The proposed methods show improved accuracy values in terms of mean absolute percentage error (MAPE) and symmetric mean absolute percentage error (SMAPE), compared to the traditional Holt's method. The Holt’s with seasonality and events yields the best MAPE of 8.64% and SMAPE of 8.90%

King Mongkut's Institute of Technology Ladkrabang. Central Library

Address: BANGKOK

Email: Lifelong@kmitl.ac.th

Name: Chartchai Leenawong

Role: Thesis Advisor

Created: 2023

Modified: 2025-03-24

Issued: 2024-04-29

วิทยานิพนธ์/Thesis

application/pdf

eng

DegreeName: Doctor of philosophy

Level: Doctoral Degree

Descipline: Applied Mathematics

Grantor: King Mongkut's Institute of Technology Ladkrabang

©copyrights King Mongkut's Institute of Technology Ladkrabang

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	Thanrada Chaikajonwat.pdf	9.03 MB	13	2025-07-29 14:04:15

ใช้เวลา

0.021905 วินาที