Greedy Online Classification of Persistent Market States Using Realized Intraday Volatility Features
(2020) In The Journal of Financial Data Science 2(3). p.25-39- Abstract
- In many financial applications, it is important to classify time-series data without any latency while maintaining persistence in the identified states. The authors propose a greedy online classifier that contemporaneously determines which hidden state a new observation belongs to without the need to parse historical observations and without compromising persistence. Their classifier is based on the idea of clustering temporal features while explicitly penalizing jumps between states by a fixed-cost regularization term that can be calibrated to achieve a desired level of persistence. Through a series of return simulations, the authors show that in most settings their new classifier remarkably obtains a higher accuracy than the correctly... (More)
- In many financial applications, it is important to classify time-series data without any latency while maintaining persistence in the identified states. The authors propose a greedy online classifier that contemporaneously determines which hidden state a new observation belongs to without the need to parse historical observations and without compromising persistence. Their classifier is based on the idea of clustering temporal features while explicitly penalizing jumps between states by a fixed-cost regularization term that can be calibrated to achieve a desired level of persistence. Through a series of return simulations, the authors show that in most settings their new classifier remarkably obtains a higher accuracy than the correctly specified maximum likelihood estimator. They illustrate that the new classifier is more robust to misspecification and yields state sequences that are significantly more persistent both in and out of sample. They demonstrate how classification accuracy can be further improved by including features that are based on intraday data. Finally, the authors apply the new classifier to estimate persistent states of the S&P 500 Index. (Less)
- Abstract (Swedish)
- In many financial applications, it is important to classify time-series data without any latency while maintaining persistence in the identified states. The authors propose a greedy online classifier that contemporaneously determines which hidden state a new observation belongs to without the need to parse historical observations and without compromising persistence. Their classifier is based on the idea of clustering temporal features while explicitly penalizing jumps between states by a fixed-cost regularization term that can be calibrated to achieve a desired level of persistence. Through a series of return simulations, the authors show that in most settings their new classifier remarkably obtains a higher accuracy than the correctly... (More)
- In many financial applications, it is important to classify time-series data without any latency while maintaining persistence in the identified states. The authors propose a greedy online classifier that contemporaneously determines which hidden state a new observation belongs to without the need to parse historical observations and without compromising persistence. Their classifier is based on the idea of clustering temporal features while explicitly penalizing jumps between states by a fixed-cost regularization term that can be calibrated to achieve a desired level of persistence. Through a series of return simulations, the authors show that in most settings their new classifier remarkably obtains a higher accuracy than the correctly specified maximum likelihood estimator. They illustrate that the new classifier is more robust to misspecification and yields state sequences that are significantly more persistent both in and out of sample. They demonstrate how classification accuracy can be further improved by including features that are based on intraday data. Finally, the authors apply the new classifier to estimate persistent states of the S&P 500 Index. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/8724d7aa-68dd-407b-bdfa-fb7bf5e810e0
- author
- Nystrup, Peter LU ; N. Kolm, Petter and Lindström, Erik LU
- organization
- publishing date
- 2020
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Statistical methods, simulations, big data/machine learning
- in
- The Journal of Financial Data Science
- volume
- 2
- issue
- 3
- pages
- 15 pages
- publisher
- Portfolio Management Research
- external identifiers
-
- scopus:85099169612
- ISSN
- 2640-3943
- DOI
- 10.3905/jfds.2020.2.3.025
- language
- English
- LU publication?
- yes
- id
- 8724d7aa-68dd-407b-bdfa-fb7bf5e810e0
- alternative location
- https://jfds.pm-research.com/content/2/3/25
- date added to LUP
- 2020-08-19 11:28:34
- date last changed
- 2023-09-14 15:07:10
@article{8724d7aa-68dd-407b-bdfa-fb7bf5e810e0, abstract = {{In many financial applications, it is important to classify time-series data without any latency while maintaining persistence in the identified states. The authors propose a greedy online classifier that contemporaneously determines which hidden state a new observation belongs to without the need to parse historical observations and without compromising persistence. Their classifier is based on the idea of clustering temporal features while explicitly penalizing jumps between states by a fixed-cost regularization term that can be calibrated to achieve a desired level of persistence. Through a series of return simulations, the authors show that in most settings their new classifier remarkably obtains a higher accuracy than the correctly specified maximum likelihood estimator. They illustrate that the new classifier is more robust to misspecification and yields state sequences that are significantly more persistent both in and out of sample. They demonstrate how classification accuracy can be further improved by including features that are based on intraday data. Finally, the authors apply the new classifier to estimate persistent states of the S&P 500 Index.}}, author = {{Nystrup, Peter and N. Kolm, Petter and Lindström, Erik}}, issn = {{2640-3943}}, keywords = {{Statistical methods, simulations, big data/machine learning}}, language = {{eng}}, number = {{3}}, pages = {{25--39}}, publisher = {{Portfolio Management Research}}, series = {{The Journal of Financial Data Science}}, title = {{Greedy Online Classification of Persistent Market States Using Realized Intraday Volatility Features}}, url = {{http://dx.doi.org/10.3905/jfds.2020.2.3.025}}, doi = {{10.3905/jfds.2020.2.3.025}}, volume = {{2}}, year = {{2020}}, }