Theory and application of locally stationary time series factor models for large-scale financial data
| Project/Area Number |
16K00042
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Statistical science
|
|---|
| Research Institution | Niigata University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2016-04-01 – 2020-03-31
|
| Project Status |
Discontinued (Fiscal Year 2019)
|
| Budget Amount *help |
¥4,550,000 (Direct Cost: ¥3,500,000、Indirect Cost: ¥1,050,000)
Fiscal Year 2019: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2017: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2016: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
|
|---|
| Keywords | 時系列解析 / 因子モデル / 局所定常時系列 / 金融工学 / 高次元データ / 局所定常 / 次元縮小 / 単位根過程 |
|---|
| Outline of Final Research Achievements |
The theory of the locally stationary time series factor models for dimension reduction for large-scale financial data was prepared, and then the statistical asymptotic theory of the finite-dimensional locally stationary time series factor model was established. A simple eigenvalue analysis of a nonnegative definite multiplied covariance matrix was used to give estimators for both the number of factors and the factor loadings. Since many existing methods only examine the consistency of estimators, the results are uniform in the stationary and non-stationary cases. Therefore, we could not judge how non-stationarity affects the estimator. By investigating the asymptotic variance of the estimator, we clarifyiedthe difference in the properties of the proposed estimator under the assumption of the locally stationary time series factor model and under the assumption of the stationary time series factor model.
|
| Academic Significance and Societal Importance of the Research Achievements |
今後は，得られた基礎理論を高次元時系列データの次元縮小に応用する。また，金融時系列データの特徴である時間と共に相関構造が滑らかに変化していく様な現象を記述するのに適している局所定常イノベーションを持つ緩やかに爆発する過程の漸近理論を導いた。緩やかに爆発する過程により，バブル期の金融時系列データを記述し，バブル期の始まりと終焉の時期を識別するのに応用した。今後は，大規模金融データに対する局所定常時系列因子モデルをバブル期の前，中，後に，それぞれあてはめることにより，何故バブル期が生まれ，はじけたかの要因を明らかにする。得られた結果を将来のバブル期の予測に応用する。
|
Report
(4 results)
Research Products
(13 results)
