-
Wavelet-based Autoencoder and EfficientNet for Schizophrenia Detection from EEG Signals
Authors:
Umesh Kumar Naik M,
Shaik Rafi Ahamed
Abstract:
Schizophrenia (SZ) is a complex mental disorder that necessitates accurate and timely diagnosis for effective treatment. Traditional methods for SZ classification often struggle to capture transient EEG features and face high computational complexity. This study proposes a convolutional autoencoder (CAE) to address these challenges by reducing dimensionality and computational complexity. Additiona…
▽ More
Schizophrenia (SZ) is a complex mental disorder that necessitates accurate and timely diagnosis for effective treatment. Traditional methods for SZ classification often struggle to capture transient EEG features and face high computational complexity. This study proposes a convolutional autoencoder (CAE) to address these challenges by reducing dimensionality and computational complexity. Additionally, we introduce a novel approach utilizing spectral scalograms (SS) combined with EfficientNet (ENB) architectures. The SS, obtained through continuous wavelet transform, reveals temporal and spectral information of EEG signals, aiding in the identification of transient features. ENB models, through transfer learning (TL), extract discriminative features and improve SZ classification accuracy. Experimental evaluation on a comprehensive dataset demonstrates the efficacy of our approach, achieving a five-fold mean cross-validation accuracy of 98.5\% using CAE with a soft voting classifier and 99\% employing SS with the ENB7 model. These results suggest the potential of our methods to enhance SZ diagnosis, surpassing traditional deep learning (DL) and TL techniques. By leveraging CAE and ENBs, this research offers a robust framework for objective SZ classification, promoting early intervention and improved patient outcomes.
△ Less
Submitted 24 July, 2024;
originally announced July 2024.
-
Finite-Horizon Constrained MDPs With Both Additive And Multiplicative Utilities
Authors:
Uday Kumar M,
Sanjay P Bhat,
Veeraruna Kavitha,
Nandyala Hemachandra
Abstract:
This paper considers the problem of finding a solution to the finite horizon constrained Markov decision processes (CMDP) where the objective as well as constraints are sum of additive and multiplicative utilities. Towards solving this, we construct another CMDP, with only additive utilities under a restricted set of policies, whose optimal value is equal to that of the original CMDP. Furthermore,…
▽ More
This paper considers the problem of finding a solution to the finite horizon constrained Markov decision processes (CMDP) where the objective as well as constraints are sum of additive and multiplicative utilities. Towards solving this, we construct another CMDP, with only additive utilities under a restricted set of policies, whose optimal value is equal to that of the original CMDP. Furthermore, we provide a finite dimensional bilinear program (BLP) whose value equals the CMDP value and whose solution provides the optimal policy. We also suggest an algorithm to solve this BLP.
△ Less
Submitted 15 March, 2023; v1 submitted 14 March, 2023;
originally announced March 2023.
-
Approximate Solutions To Constrained Risk-Sensitive Markov Decision Processes
Authors:
Uday Kumar M,
Sanjay P Bhat,
Veeraruna Kavitha,
Nandyala Hemachandra
Abstract:
This paper considers the problem of finding near-optimal Markovian randomized (MR) policies for finite-state-action, infinite-horizon, constrained risk-sensitive Markov decision processes (CRSMDPs). Constraints are in the form of standard expected discounted cost functions as well as expected risk-sensitive discounted cost functions over finite and infinite horizons. The main contribution is to sh…
▽ More
This paper considers the problem of finding near-optimal Markovian randomized (MR) policies for finite-state-action, infinite-horizon, constrained risk-sensitive Markov decision processes (CRSMDPs). Constraints are in the form of standard expected discounted cost functions as well as expected risk-sensitive discounted cost functions over finite and infinite horizons. The main contribution is to show that the problem possesses a solution if it is feasible, and to provide two methods for finding an approximate solution in the form of an ultimately stationary (US) MR policy. The latter is achieved through two approximating finite-horizon CRSMDPs which are constructed from the original CRSMDP by time-truncating the original objective and constraint cost functions, and suitably perturbing the constraint upper bounds. The first approximation gives a US policy which is $ε$-optimal and feasible for the original problem, while the second approximation gives a near-optimal US policy whose violation of the original constraints is bounded above by a specified $ε$. A key step in the proofs is an appropriate choice of a metric that makes the set of infinite-horizon MR policies and the feasible regions of the three CRSMDPs compact, and the objective and constraint functions continuous. A linear-programming-based formulation for solving the approximating finite-horizon CRSMDPs is also given.
△ Less
Submitted 29 September, 2022;
originally announced September 2022.
-
Performance Analysis of MIMO Radar Waveform using Accelerated Particle Swarm Optimization Algorithm
Authors:
B. Roja Reddy,
Uttara Kumari . M
Abstract:
The Accelerated Particle Swarm Optimization Algorithm is promoted to numerically design orthogonal Discrete Frequency Waveforms and Modified Discrete Frequency Waveforms (DFCWs) with good correlation properties for MIMO radar. We employ Accelerated Particle Swarm Optimization algorithm (ACC_PSO), Particles of a swarm communicate good positions, velocity and accelerations to each other as well as d…
▽ More
The Accelerated Particle Swarm Optimization Algorithm is promoted to numerically design orthogonal Discrete Frequency Waveforms and Modified Discrete Frequency Waveforms (DFCWs) with good correlation properties for MIMO radar. We employ Accelerated Particle Swarm Optimization algorithm (ACC_PSO), Particles of a swarm communicate good positions, velocity and accelerations to each other as well as dynamically adjust their own position, velocity and acceleration derived from the best of all particles. The simulation results show that the proposed algorithm is effective for the design of DFCWs signal used in MIMO radar.
△ Less
Submitted 15 September, 2012;
originally announced September 2012.
