<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>Adaptive linear neural networks lay the foundation for the development of the uniqueness of algorithmic composition and emotion recognition. In this paper, we first analyze the process of emotion recognition and the development of algorithmic compositions to establish the emotion recognition dataset. Secondly, the algorithm of the adaptive linear neural network is selected, including the analysis of the adaptive linear neuron model and gradient and most rapid descent method and LMS algorithm. The analysis focuses on the LMS algorithm flow, convergence conditions and performance parameters of the LMS algorithm. Finally, the sentiment recognition results of four models, SVM, CNN, LSTM and Adaline neural network, based on different dimensional self-encoder features, are analyzed. To verify whether the classification method of self-encoder + Adaline neural network can find the information connection between various emotions and improve the efficiency of emotion recognition. The classification method of self-encoder + Adaline neural network can improve the recognition rate by up to 85% for noise-reducing self-encoder features in 500 dimensions.</p>
</abstract>

Adaptive linear neural networks lay the foundation for the development of the uniqueness of algorithmic composition and emotion recognition. In this paper, we first analyze the process of emotion recognition and the development of algorithmic compositions to establish the emotion recognition dataset. Secondly, the algorithm of the adaptive linear neural network is selected, including the analysis of the adaptive linear neuron model and gradient and most rapid descent method and LMS algorithm. The analysis focuses on the LMS algorithm flow, convergence conditions and performance parameters of the LMS algorithm. Finally, the sentiment recognition results of four models, SVM, CNN, LSTM and Adaline neural network, based on different dimensional self-encoder features, are analyzed. To verify whether the classification method of self-encoder + Adaline neural network can find the information connection between various emotions and improve the efficiency of emotion recognition. The classification method of self-encoder + Adaline neural network can improve the recognition rate by up to 85% for noise-reducing self-encoder features in 500 dimensions.

Research on algorithm composition and emotion recognition based on adaptive networks

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

Adaptive linear neural networks lay the foundation for the development of the uniqueness of algorithmic composition and emotion recognition. In this paper, we...

Research on algorithm composition and emotion recognition based on adaptive networks

Pubblicato online: 15 ott 2023

Pagine: -

Ricevuto: 14 gen 2023

Accettato: 25 apr 2023

DOI: https://doi.org/10.2478/amns.2023.2.00649

Parole chiave
Adaptive linear neural network, LMS algorithm, Algorithm composition, Most rapid descent method, Self-encoder

© 2023 Shuxin Hou et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Research on algorithm composition and emotion recognition based on adaptive networks

Pubblicato online: 15 ott 2023

Pagine: -

Ricevuto: 14 gen 2023

Accettato: 25 apr 2023

DOI: https://doi.org/10.2478/amns.2023.2.00649

Parole chiaveAdaptive linear neural network, LMS algorithm, Algorithm composition, Most rapid descent method, Self-encoder

© 2023 Shuxin Hou et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Parole chiave
Adaptive linear neural network, LMS algorithm, Algorithm composition, Most rapid descent method, Self-encoder