| dc.contributor | 國貿系 | - |
| dc.creator (作者) | 顏佑銘 | - |
| dc.date (日期) | 2020-01 | - |
| dc.date.accessioned | 12-May-2026 15:55:08 (UTC+8) | - |
| dc.date.available | 12-May-2026 15:55:08 (UTC+8) | - |
| dc.date.issued (上傳時間) | 12-May-2026 15:55:08 (UTC+8) | - |
| dc.identifier.uri (URI) | https://ah.lib.nccu.edu.tw/item?item_id=182495 | - |
| dc.description.abstract (摘要) | 在這個計畫中，我會發展新的統計檢定方法，在一致性損失(評價) 函數下，來比較期望分位數及分位數預測之精確性。此統計檢定的虛無假設如下: 在所有的極端一致性損失函數（Ehmet al., 2016）之下，基準預測至少會與競爭預測有相同的表現。如果這樣的虛無假設成立，則在所有的一致性損失函數下，基準預測也至少會與競爭預測有相同的表現。因此，在此虛無假設之下，當使用不同的一致性損失函數時，競爭預測表現得不如基準預測的結果，將不會因此而改變。在實際執行此統計檢定上，我將會建構Kolmogorov-Smirnov 類型的檢定統計量。之後我會建立此檢定統計量的漸近性質，並進行模擬，以探討此檢定統計量在不同情況下之表現。我還會運用所提出之統計檢測，重新審視一些常用的預測變數對S&P500 指數風險溢價之預測能力。本計畫未來研究將包括：(1) 比較預測者在不同領域的預測表現，及探討相關經濟意義和政策意涵；(2) 更為深入地探討所提出之檢定統計量的理論性質，例如它們的統計檢定力；(3) 探討是否可以改進所提出之檢定統計量的效率性；(4) 嘗試將無條件檢定擴展到條件相等精確性檢定。 | - |
| dc.description.abstract (摘要) | In the project, I plan to develop new statistical tests for comparing perfor-mances of forecasting expectiles and quantiles of a random variable under consis-tent loss (scoring) functions. The null hypothesis of the tests is that a benchmark forecast at least performs equally well as a competitive one under all extremal consistent loss functions (Ehm, et.al., 2016). It can be shown that if such a null holds, the benchmark will also perform at least equally well as the competitor under all consistent loss functions. Thus under the null, when different consistent loss functions are used, the result that the competitor does not outperform the benchmark will not be altered. To implement the tests, I will construct Kolmogorov-Smirnov type statistics. Then I will build asymptotic theories for the test statistics and conduct simulations to show how the test statistics perform under different situations. I will also apply the proposed test on a real empirical analysis: re-examining abilities of some predictors on forecasting risk premium of the S&P500 index. Future research will include: (1) empirical comparisons of forecasters' perfor-mances in different research fields and the related economic insights and policy implications; (2) more discussions on theoretical properties of the proposed test statistics, for example, their statistical power; (3) investigating the possibility of more efficient test statistics; (4) extending the unconditional tests to conditional equivalent forecast accuracy tests. | - |
| dc.format.extent | 116 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.relation (關聯) | 科技部, MOST106-2410-H004-014-MY2, 106.08-108.07 | - |
| dc.subject (關鍵詞) | 一致性損失函數; 期望分位數; 極端一致性損失函數; 預測精確性; 分位數 | - |
| dc.subject (關鍵詞) | Consistent loss function; Expectile; Extremal consistent loss function; Forecast accuracy; Quantile | - |
| dc.title (題名) | 運用極端一致性損失函數來檢定期望分位數及分位數預測之精確性 | - |
| dc.title (題名) | Testing Forecast Accuracy of Expectiles and Quantiles with the Extremal Consistent Loss Functions | - |
| dc.type (資料類型) | report | - |
