This document discusses using deep learning methods to classify and predict chaotic time series for medical applications. The objectives are to develop algorithms to train models for chaotic time series classification, improve on previous research results, induce noise to test model performance, and compare different deep learning models. It also aims to detect heart diseases using these methods. The document reviews previous literature on classifying chaotic time series with deep learning. It identifies gaps in current research, such as only considering simple chaotic systems, not classifying other flows, needing more noise addition research, and requiring comparison of models for classification and prediction of chaotic time series. It also notes using these methods for heart rate variability and other disease time series is an under-researched application area.

1. CLASSIFICATION AND PREDICTION OF CHAOTIC TIME
SERIES USING DEEP LEARNING METHODS FOR MEDICAL
APPLICATIONS
PAWAN BHARADWAJ
ASST. PROF. NIEIT MYSURU

2. Chaos
1. Chaos theory is a branch of mathematics focusing on the study of chaos — dynamical systems
whose apparently random states of disorder and irregularities are actually governed by
underlying patterns and deterministic laws that are highly sensitive to initial conditions.
2. Sensitivity to initial conditions means that each point in a chaotic system is arbitrarily closely
approximated by other points that have significantly different future paths or trajectories. Thus,
an arbitrarily small change or perturbation of the current trajectory may lead to significantly
different future behavior.

4. AIM AND OBJECTIVE
AIM
1. To classify and predict chaotic time series using deep learning methods.
2. Applications of classification of chaotic time series for medical applications.
OBJECTIVES
1. Developing/designing algorithms to train the recognition models for chaotic time series.
2. To improve on the results obtained by previous researches.
3. To induce noise and check performance of models for chaotic time series.
4. To compare various deep learning models of deep learning for chaotic time series.
5. Detection of heart diseases by the above said methods.

5. Literature survey
SL.NO Title of paper Author Year
1 Classification of chaotic time
series with deep learning
Nicolas Boullé,
Vassilios Dallas, Yuji
Nakatsukasa, D.
Samaddar
2019, Elsevier
https://doi.org/10.1016/j.physd.2019.13
2261
2. Deep Learning of Chaos
Classification
Woo Seok Lee1 and
Sergej Flach
Mach. Learn.: Sci. Technol.1(2020)
045019https://doi.org/10.1088/2632-
2153/abb6d3
3. Deep Chaos Synchronization Majid Mobini;
Georges Kaddoum
IEEE, 2020. DOI:
10.1109/OJCOMS.2020.3028554
4. Deep Convolutional Neural
Networks for Chaos
Identification in Signal
Processing
Andrey V. Makarenko IEEE, 2018
10.23919/EUSIPCO.2018.8553098

6. RESEARCH GAP IDENTIFIED
1. Current research is restricted to simple chaotic systems like Lorentz and sine maps.
2. Autonomous and other periodic flows are not classified.
3. Addition of noise to the time series are to be properly researched.
4. Comparative study of classification and prediction of chaotic time series using various deep
learning models.
5. Applications to Heart rate variability and other diseases whose time series are chaotic in
nature is a niche area in deep learning classification.

7. THANK YOU