hzion-liaaa-aaaad-qbiyq-cai.icp0.io
Open in
urlscan Pro
2a0b:21c0:b002:2:5000:edff:fe0d:98de
Public Scan
URL:
https://hzion-liaaa-aaaad-qbiyq-cai.icp0.io/personal/Generation-of-Realistic-Synthetic-Financial-Time-Series/index.html
Submission: On December 22 via api from US — Scanned from FR
Submission: On December 22 via api from US — Scanned from FR
Form analysis 1 forms found in the DOM
<form class="search-content__form" onkeydown="return event.key != 'Enter';">
<label class="sr-only" for="search"> Enter your search term... </label>
<input type="search" id="search" class="search-input" tabindex="-1" placeholder="Enter your search term...">
</form>Text Content
SKIP LINKS
* Skip to primary navigation
* Skip to content
* Skip to footer
Liviu-Daniel ȘTEFAN
* CV
* Publications
* Projects
* Blog
Toggle search Toggle menu
LIVIU-DANIEL ȘTEFAN
Research Scientist @AIMultimedia Lab, CAMPUS Research Institute.
Follow
* Bucharest, Romania
* Download CV
* Email
* Affiliation
* Publons
* GitHub
* Stack Overflow
* Facebook
GENERATION OF REALISTIC SYNTHETIC FINANCIAL TIME-SERIES
1 minute read
Financial markets have always been a point of interest for automated systems.
Due to their complex nature, financial algorithms and fintech frameworks require
vast amounts of data to accurately respond to market fluctuations. This data
availability is tied to the daily market evolution so it is impossible to
accelerate its acquisition. In this paper, we discuss several solutions for
augmenting financial datasets via synthesizing realistic time-series with the
help of generative models. This problem is complex since financial time series
present very specific properties, e.g., fat-tail distribution, cross-correlation
between different stocks, specific autocorrelation, cluster volatility etc. In
particular, we propose solutions for capturing cross-correlations between
different stocks and for transitioning from fixed to variable length time-series
without resorting to sequence modeling networks, and adapt various network
architectures, e.g., fully connected and convolutional GANs, variational
autoencoders, and generative moment matching networks. Finally, we tackle the
problem of evaluating the quality of synthetic financial time-series. We
introduce qualitative and quantitative metrics, along with a portfolio trend
prediction framework which validates our generative models’ performance. We
carry out experiments on real-world financial data extracted from the US stock
market proving the benefits of these techniques.
Please find all the details about the data set in the draft of the main paper
[here].
@inproceedings{dogariu2022generation,
author = {Dogariu, Mihai and {\c{S}}tefan, Liviu-Daniel and Boteanu, Bogdan Andrei and Lamba, Claudiu and Kim, Bomi and Ionescu, Bogdan},
title = {Generation of Realistic Synthetic Financial Time-Series},
booktitle = {ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM)},
year = {2022}
}
Tags: blog
Categories: Personal
Updated: December 5, 2021
SHARE ON
Twitter Facebook LinkedIn
Previous Next
YOU MAY ALSO ENJOY
[FAVCI2D] FACE VERIFICATION WITH CHALLENGING IMPOSTERS AND DIVERSIFIED
DEMOGRAPHICS
2 minute read
With the recent General Data Protection Regulation (GDPR) directive in European
Union (EU) regarding data protection, establishing a new person identificat...
PH.D. THESIS - DEEP LEARNING FOR MULTIMEDIA CONTENT CLASSIFICATION
1 minute read
The era of big data presents an important challenge for data analysis,
processing, and learning, as information become available in such a volume. In
this...
WELCOME!
less than 1 minute read
Welcome to my webpage. Here, I will share my thought about the research I’m
conducting, tutorials for various technologies and tools, and code samples for
in...
Enter your search term...
* Follow:
* Facebook
* GitHub
* Stack Overflow
* Feed
© 2023 Liviu-Daniel ȘTEFAN