時間序列智能化演算法應用於螢幕掛燈短期銷售預測

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：188

、訪客IP：18.191.168.29

姓名

林秋萍(Chiu-Ping Lin) 查詢紙本館藏

畢業系所

企業管理學系在職專班

論文名稱

時間序列智能化演算法應用於螢幕掛燈短期銷售預測

相關論文

★ 從生態共生觀點發展組織合作模式	★ 影響產業垂直分工因素之探討
★ 發展台籍專業管家的人力派遣模式	★ 影響軍事機關審計品質之關鍵性因素及其相對效率衡量
★ 台灣光碟片設備供應商競爭策略分析	★ 國防科技研發機構組織向心力之研究
★ 衡量半導體零組件通路商之相對經營績效	★ 衡量半導體晶圓代工廠商生產單位之營運績效
★ 商業銀行營運型態與組織調整對分行營運效率的影響	★ 發展行動條碼為基礎的隨傳隨用視訊平台之應用
★ 飛機引擎定子零件維修之訂價調整方式	★ 國際化、創新活動與公司績效：台灣上市公司董事會結構論析
★ 探討網際網路事件達成關鍵多數之影響因素	★ 生態循環農業產銷策略個案研究
★ 探討以政策規範高耗能產業參與再生能源開發之綜效－以太陽光電發電系統為例	★ 適地性服務(LBS)之顧客體驗整合性設計—以中華電信行動導遊為例

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 ( 永不開放)

摘要(中)

隨著全球市場迅速變遷及科技進步，銷售預測在企業財務風險管理、庫存控制與生產製造規劃中扮演關鍵角色。精準的銷售預測不僅能提升企業成本效益與客戶滿意度，還可減少生產庫存風險，進而增強企業競爭力與財務穩定。因此，有效的銷售預測對於應對市場需求變化及優化庫存與生產管理至關重要。
近年來，由於電子產品的廣泛運用及居家辦公模式的興起，消費者對於護眼閱讀及智慧照明產品的需求明顯增加。在選購如電腦螢幕等電子產品時，消費者往往也會同步考慮購買螢幕護眼照明裝置。面對此趨勢，個案公司建立安全庫存水位以應對市場需求的波動，確保產品供應穩定性。因此，時間序列模型應用於螢幕掛燈的短期銷售預測顯得尤為重要。然而，除了季節性變化和市場趨勢的影響，同時需要考量如疫情等特殊事件的衝擊，使得短期銷售預測充滿挑戰。
本研究對市場A、B中的三種螢幕掛燈ScreenLight A、B、C，應用Prophet、SARIMA及Holt-Winters三種時間序列模型，訓練區間分為1年、1.5年及2年來進行未來三個月的短期銷售預測，以評估這些模型在處理季節性變化及突發事件時的效能。研究結果顯示，由於Prophet模型在對季節性變化和特殊事件的快速響應方面具有優異的適應力，使其在三種模型中展示了較高的預測精確性。相較之下，SARIMA和Holt-Winters模型雖在特定條件下表現良好，但在更廣泛的應用場景中適應性相對較弱。
本研究展示了短期銷售預測在螢幕掛燈市場中的重要性，同時為相關企業提供了具體且可行的策略，以增強其面對市場不確定性及外部衝擊時的快速反應能力。

摘要(英)

With the rapid changes in the global market and advances in technology, sales forecasting plays a crucial role in corporate financial risk management, inventory control, and production planning. Accurate sales forecasting not only enhances corporate cost-effectiveness and customer satisfaction but also reduces the risks of overproduction and insufficient inventory, thereby boosting competitive advantage and financial stability. Thus, effective sales forecasting is essential for responding to market demand fluctuations and optimizing inventory and production management.
In recent years, the widespread use of electronic products and the rise of the work-from-home model have significantly increased consumer demand for eye-protective reading and smart lighting products. When purchasing electronic products such as computer monitors, consumers often consider acquiring screen protective lighting devices simultaneously. In response to this trend, the case company has established safe inventory levels to stabilize product supply in the face of market demand fluctuations. Therefore, the application of time series algorithms for short-term sales forecasting of screen lighting lamps becomes particularly important. However, the forecasting process is challenging due to the impact of seasonal variations, market trends, and special events like pandemics.
This study applies three time series algorithms—Prophet, SARIMA, and Holt-Winters—to forecast the short-term sales for the next three months of three types of screen lighting lamps, ScreenLight A, B, and C, in markets A and B. The training intervals are set at 1 year, 1.5 years, and 2 years, to assess these models′ efficacy in handling seasonal changes and sudden events. The results show that the Prophet model, due to its excellent adaptability in responding quickly to seasonal changes and special events, demonstrated higher forecasting accuracy among the three algorithms. In contrast, while the SARIMA and Holt-Winters models performed well under specific conditions, their adaptability in broader application scenarios was relatively weaker.
This research underscores the importance of short-term sales forecasting in the screen lighting lamp market and provides specific and feasible strategies for related businesses to enhance their rapid response capabilities in the face of market uncertainties and external shocks.

關鍵字(中)

★ 短期銷售預測
★ 時間序列智能化演算法
★ Prophet
★ SARIMA
★ Holt-Winters
★ 平均絕對百分比誤差

關鍵字(英)

★ Short-term Sales Forecasting
★ Time Series Forecasting Algorithms
★ Prophet
★ SARIMA
★ Holt-Winters
★ MAPE

論文目次

中文摘要 i
Abstract ii
目錄 iv
圖目錄 vi
表目錄 vii
一、緒論 1
1-1 研究背景 1
1-2 研究動機 2
1-3 研究目的 2
二、文獻探討 4
2-1 時間序列分析 4
2-2 銷售預測的重要性 5
2-3 銷售預測模型的演進 6
2-4 總結 7
三、研究方法 8
3-1 銷售預測目標 8
3-2 資料來源與處理 10
3-3 時間序列模型 12
3-3-1 研究對象 12
3-3-2 模型建立 13
3-4 衡量指標 14
3-4-1 平均絕對百分比誤差(MAPE): 14
3-4-2 可視化擬合圖分析: 15
四、研究結果 16
4-1 銷售資料分析 16
4-2 模型驗證 20
4-2-1 Prophet時間序列預測模型 20
4-2-2 SARIMA時間序列預測模型 29
4-2-3 Holt-Winters時間序列預測模型 30
4-3 測試資料結果 36
4-4 預測模型參數 37
五、研究結論與未來方向 40
5-1 研究結論 40
5-2 研究限制 40
5-3 未來研究方向 41
參考文獻 42

參考文獻

〔1〕Enders, W. (2014). Applied Econometric Time Series. Wiley.
〔2〕Taylor, S. J., & Letham, B. (2018). Forecasting at scale. The American Statistician,
72(1), 37-45.
〔3〕Box, G. E., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time Series
Analysis: Forecasting and Control. John Wiley & Sons.
〔4〕Holt, C. C. (1957). Forecasting seasonals and trends by exponentially weighted moving
averages. ONR Research Memorandum.
〔5〕Winters, P. R. (1960). Forecasting sales by exponentially weighted moving averages.
Management Science, 6(3), 324-342.
〔6〕Brockwell, P. J., & Davis, R. A. (2002). Introduction to Time Series and Forecasting.
Springer.
〔7〕Chopra, S., & Meindl, P. (2013). Supply Chain Management: Strategy, Planning, and
Operation. Pearson.
〔8〕Lee, H. L., Padmanabhan, V., & Whang, S. (1997). The bullwhip effect in supply
chains. Sloan Management Review, 38(3), 93-102.
〔9〕Christopher, M., & Towill, D. R. (2001). An integrated model for the design of agile
supply chains. International Journal of Physical Distribution & Logistics Management, 31(4),
235-246.
〔10〕Mentzer, J. T., Moon, M. A., & Matsuno, K. (2008). Sales Forecasting Management:
Understanding the Techniques, Systems, and Management of the Sales Forecasting Process.
Sage Publications.

指導教授

張東生(Dong-Shang Chang)

審核日期

2024-6-17

推文