大型總體經濟計量模型為國內主要預測機構的主要工具，但模型設定困難且耗時，若無法及時更新可能造成系統性偏誤。再者進行分析時多以相同頻率變數為主，若可以將不同頻率變數納入候選變數供模型挑選，將使模型不再受限於資料頻率，且模型能納入更及時的資訊。本文採用資料驅動的方式建置總體經濟計量模型，採用兩階段變數選取使模型降維，降低模型設定上的困難。第一階段使用 LASSO-MIDAS 方法同時將不同頻率變數混合及挑選變數；接著再使用逐步迴歸法進行第二階段的挑選，找出調整後判斷係數最高的變數組合避免模型過度配適。此外為使模型能包含國外經濟環境，以主成分分析法從國外變數中萃取出國外因子，加入模型中作為外生變數使用。最後再將模型擴展為能於樣本外預測多期，進行重要經濟變數的預測。
實證結果發現，變數組合中選入高頻變數的內生變數，相較於沒有高頻變數的模型，預測表現上有獲得改善；本模型對重要經濟變數的預測與其他機構比較，預測值屬於合理範圍內，預測表現並不遜於其他機構；比較本模型使用不同求解期間，模型中變數組合雖不同，預測表現依然穩定。以上結果顯示將高頻變數加入模型中具有改善模型預測能力的可能性，且本模型的穩定性高。未來對於建置高維度總體經濟計量模型，可考慮本文所建議的方法。

Macroeconometric model is the main tool used by major forecasting organizations in domestic, but it is difficult and time-consuming to deal with model specification and identification. On the other hand, it may cause systematic bias if the model is not updated in time. In addition, the same frequency variables are used in most of the analyses. If different frequency variables can be included as candidate variables for model selection, the model will no longer be limited by the data frequency, and the model can incorporate more timely information. In this paper, we adopt a data-driven approach to build the macroeconometric model by using two stages of variable selection that leave model dimension reduction and reduce the difficulty of model specification and identification. In the first stage, the LASSO-MIDAS method is used to mix different frequency variables and select variables at the same time. In the second stage, we use stepwise regression to find the combination of variables with the highest coefficient after adjustment to avoid overfitting the model. Furthermore, the foreign factors are extracted from the foreign variables by principal component analysis and added to the model as exogenous variables to include the foreign economic environment. Finally, the model is extended to be able to forecast multiple periods out of the sample.

中文部分
吳中書、單易、鄭淑如、梅家瑗、蘇文瑩、高志祥、羅雅惠、黃純宜與王淑娟
(2001)，「臺灣總體經濟計量動態季模型」，《臺灣經濟預測與政策》，31(1)，
111-159。
林建甫 (2006)，「台灣總體經濟金融模型之建立」，《中央銀行季刊》，28(1)，5-41。
林建甫 (2010)，「總體經濟計量模型的建立與應用」，《經濟論文叢刊》，38(1)，1-64。
徐士勛、管中閔與羅雅惠 (2005)，「以擴散指標為基礎之總體經濟預測」，《臺灣經濟預
測與政策》，36(1)，1-28。
吳俊毅、朱浩榜(2020)，「即時預報台灣經濟成長率：MIDAS模型之應用」，《中央銀行
季刊》，42-1，59-84。
盧亦盛(2021)，「高維度變數下總體經濟預測模型的建立」，國立成功大學研究所碩士論
文。
黃裕烈、管中閔(2019)，「美國聯準會會議紀要的文字探勘與台灣經濟變數預測」，《經
濟論文叢刊》，47(3)，363-391。
陳宜廷、徐士勛、劉瑞文與莊額嘉，「經濟成長率預測之評估與更新」，《經濟論文叢
刊》，39(1)，1-44。
英文部分
Andreoua, E., E. Ghysels and A. Kourtellos (2010), “Regression Models with Mixed Sampling Frequencies,” Journal of Econometrics, 158, 246–261.
Andreou, E., E. Ghysels and A. Kourtellos (2013), “Should Macroeconomic Forecasters Use Daily Financial Data and How?” Journal of Business & Economic Statistics, 31:2, 240-251.
Boriss, S. (2017), “Short-term Forecasting with Mixed-frequency Data: A MIDASSO Approach,” Applied Economics, 49:13, 1326-1343.
Ferrara, L. and C. Marsilli (2019), “Nowcasting Global Economic Growth: A Factor‐Augmented Mixed‐Frequency Approach,” The World Economy, 42, 846–875.
Ghysels, E., P. Santa-Clara and V. Rossen (2004), “The MIDAS Touch: Mixed Data Sampling Regression Models,” Discussion Paper UNC/UCLA.
Ghysels, E., A. Sinko and R. Valkanov (2007), “MIDAS Regressions: Further Results and New Directions,” Econometric Reviews, 26:1, 53-90.
Gu, S., B. Kelly and D. Xiu (2020), “Empirical Asset Pricing via Machine Learning,” Review of Financial Studies, 33:5, 2223–2273.
Hocking, R. R. and R. N. Leslie (1967), “Selection of the Best Subset in Regression Analysis,” Technometrics, 9:4, 531-540.
Hwang, J. and T. Hu (2015), “A Stepwise Regression Algorithm for High-Dimensional Variable Selection,” Journal of Statistical Computation and Simulation, 85:9, 1793-1806.
Hyndman, R. J. and G. Athanasopoulos (2013). Forecasting: Principles and Practice, 3rd edition, OTexts: Melbourne, Australia. otexts.com/fpp3/. Accessed on 2022/6/26.
Li, J. and W. Chen (2014), “Forecasting Macroeconomic Time Series: LASSO-Based Approaches and Their Forecast Combinations with Dynamic Factor Models,” International Journal of Forecasting, 30:4, 996-1015.
Marsilli, C. (2014), “Variable Selection in Predictive MIDAS Models,” Document de Travail, No. 520.
McCracken, M. W. and S. Ng (2016), “FRED-MD: A Monthly Database for Macroeconomic Research,” Journal of Business & Economic Statistics, 34:4, 574-589.
McCracken, M. W. and S. Ng (2020), “FRED-QD: A Quarterly Database for Macroeconomic Research,” NBER working paper, No. 26872.
Medeiros, M. C., G. F. R. Vasconcelos, Á. Veiga and E. Zilberman (2021), “Forecasting Inflation in a Data-Rich Environment: The Benefits of Machine Learning Methods,” Journal of Business & Economic Statistics, 39:1, 98-119.
Nakajima, Y. and N. Sueishi (2020), “Forecasting the Japanese Macroeconomy Using High-Dimensional Data,” The Japanese Economic Review, 73, 299–324.
Pope, P. T. and J. T. Webster (1972), “The Use of an F-Statistic in Stepwise Regression Procedures,” Technometrics, 14:2, 327-340.
Siliverstovs, B. (2016), “Short-term forecasting with mixed-frequency data: a MIDASSO approach,” Applied Economics, 49:13, 1326-1343.
Stock, J. H. and M. W. Watson (1998), “Diffusion Indexes,” NBER working paper, No. 6702.
Stock, J. H. and M. W. Watson (2002), “Forecasting Using Principal Components From a Large Number of Predictors,” Journal of the American Statistical Association, 97: 460, 1167-1179.
Tibshirani, R. (1996), “Regression Shrinkage and Selection via the Lasso, “Journal of the Royal Statistical Society, Series B (Methodological), 58:1, 267-288.
資料來源
中華民國統計資訊網
https://statdb.dgbas.gov.tw/pxweb/Dialog/statfile9L.asp
AREMOS 台灣經濟統計資料庫
http://net.aremos.org.tw/
ECONOMIC RESEARCH
https://research.stlouisfed.org/econ/mccracken/fred-databases
國發會
https://index.ndc.gov.tw/n/zh_tw
yahoo 財經
https://hk.finance.yahoo.com/