Type
Text
Type
Dissertation
Advisor
Rachev, Svetlozar | Mullhaupt, Andrew P | Xing, Haipeng | Xiao, Keli.
Date
2015-05-01
Keywords
Applied mathematics | 130-30 Fund, Fisher Information, Log Band Fraction, Long Only Portfolio, Low Volatility Strategy
Department
Department of Applied Mathematics and Statistics.
Language
en_US
Source
This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree.
Identifier
http://hdl.handle.net/11401/77488
Publisher
The Graduate School, Stony Brook University: Stony Brook, NY.
Format
application/pdf
Abstract
Structured matrix plays an important role in statistics, especially in covariance estimation. Band fraction representation is one of the efficient structures for matrices. In this dissertation, we study the metric tensor for the band fraction representation for the covariance matrix. We propose a new structure, the log band fraction representation, which gives smaller information distance and Hellinger distance than factor model and band fraction representation. We apply the log band fraction estimation in the portfolio optimization problem. We propose our long only strategy and 130-30 strategy, which significantly outperform the benchmarks, i.e. | SPY, SPLV, and CSM. Transaction cost is considered in the portfolio construction process. The strategies proposed in this dissertation are fully investable. | 93 pages
Recommended Citation
Yu, Riyu, "Log Band Fraction Approximation For Covariance Estimation and Low Volatility Strategy" (2015). Stony Brook Theses and Dissertations Collection, 2006-2020 (closed to submissions). 3300.
https://commons.library.stonybrook.edu/stony-brook-theses-and-dissertations-collection/3300