Advances in Optimisation Algorithms and Techniques for Deep Learning

Nwankpa, Chigozie

doi:10.25046/aj050570

Cited by 13 publications

(3 citation statements)

References 57 publications

(77 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The activation function [53] for the hidden dense layer is the rectified linear unit (ReLU). For binary classification, the output dense layer is the sigmoid function, and the models are compiled using binary cross-entropy as the loss function and the adaptive movement estimation optimizer Adam [54]. In the case of multi-class classification with 3 or 5 classes, the models are compiled using the sparse categorical cross-entropy function, and for the output dense layer, we use the so f tmax function.…”

Section: Models 441 Supervised Classificationmentioning

confidence: 99%

Machine-Learning-Based Approaches for Multi-Level Sentiment Analysis of Romanian Reviews

Briciu,

Călin,

Miholca

et al. 2024

Mathematics

View full text Add to dashboard Cite

Sentiment analysis has increasingly gained significance in commercial settings, driven by the rising impact of reviews on purchase decision-making in recent years. This research conducts a thorough examination of the suitability of machine learning and deep learning approaches for sentiment analysis, using Romanian reviews as a case study, with the aim of gaining insights into their practical utility. A comprehensive, multi-level analysis is performed, covering the document, sentence, and aspect levels. The main contributions of the paper refer to the in-depth exploration of multiple sentiment analysis models at three different textual levels and the subsequent improvements brought with respect to these standard models. Additionally, a balanced dataset of Romanian reviews from twelve product categories is introduced. The results indicate that, at the document level, supervised deep learning techniques yield the best outcomes (specifically, a convolutional neural network model that obtains an AUC value of 0.93 for binary classification and a weighted average F1-score of 0.77 in a multi-class setting with 5 target classes), albeit with increased resource consumption. Favorable results are achieved at the sentence level, as well, despite the heightened complexity of sentiment identification. In this case, the best-performing model is logistic regression, for which a weighted average F1-score of 0.77 is obtained in a multi-class polarity classification task with three classes. Finally, at the aspect level, promising outcomes are observed in both aspect term extraction and aspect category detection tasks, in the form of coherent and easily interpretable word clusters, encouraging further exploration in the context of aspect-based sentiment analysis for the Romanian language.

show abstract

Section: Models 441 Supervised Classificationmentioning

confidence: 99%

Machine-Learning-Based Approaches for Multi-Level Sentiment Analysis of Romanian Reviews

Briciu,

Călin,

Miholca

et al. 2024

Mathematics

View full text Add to dashboard Cite

show abstract

“…This error function is iteratively minimized by optimization algorithms, which aims to find optimal weights that improve the learning process of the model and minimize the error between predicted and observable values. The MLP model uses several types of optimization algorithms or also called learning algorithms such as Gradient Descending (SGD), RMSprop and Adam, among others [26].…”

Section: The Multilayer Perceptron (Mlp) Modelmentioning

confidence: 99%

Global solar radiation time series forecasting using different architectures of the multilayer perceptron model

Penalva

Lozano

Murillo

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

In this work, the multilayer perceptron model was used to forecast the time series of global solar radiation for a near future about a week. Different architectures of this model were built through varying its different hyperparameters such as optimizers, activation functions, number of neurons and neuron dropout in which their performance was evaluated using error metrics. It was found that the architectures (60, SGD, Sigmoid), (10, Adam, Relu) and (60, SGD, Sigmoid) presented an R2 around 0.877, 0.873 and 0.872, respectively. The architecture with neuron dropout (150, SGD, Sigmoid, 0.2) presented a higher performance among all the architectures evaluated and its R2 value was 0.884. Architectures with higher performance are used to predict future values of solar radiation.

show abstract

“…e psychological community has never stopped discussing the definition of emotions, and the criteria for classifying emotions have become a point of debate among scholars. Some of the more recognized theories of emotion classification are the basic emotion theory and the dimensional spatial emotion theory [18]. e basic emotion theory holds that emotions have archetypal forms in their occurrence; that is, there are several pan-human basic emotion types.…”

Section: Analysis Of the Chaotic Fish Swarm Algorithm Mental Health S...mentioning

confidence: 99%

A Chaotic Fish Swarm Algorithm-Based Model for Assessing the Mental Health Status of Older Adults

Zhang

Yao

et al. 2022

Scientific Programming

View full text Add to dashboard Cite

In this paper, the chaotic fish swarm algorithm is used to conduct in-depth research and analysis on the assessment of the mental health of the elderly. Firstly, the principle, search method, and characteristics of the harmonic search algorithm are analysed, and it is proposed to use the excellent local fine-tuning ability of the harmonic search algorithm to improve the local search accuracy of the artificial fish swarm algorithm. Then, the concept of chaos factor is introduced to improve the global search of the artificial fish swarm algorithm efficiency, using its global search capability without repeated traversal to form a new hybrid fish swarm algorithm. The comparison of experimental results shows that the improved algorithm can effectively guide the robot to avoid obstacles and quickly find the best path or a better path. The improved hybrid algorithm is more efficient and reliable than other algorithms in path planning and can handle more a complex environment model. When considering sample selection bias, ordinary least squares (OLS) regression may underestimate the extent to which social participation affects the mental health of older adults. Further research found that there is heterogeneity in the influence of social participation on the mental health of the elderly. In addition, different types of social participation have different effects on the mental health of the elderly. Simply making friends, physical exercise, and recreational participation in social activities can significantly improve the mental health of the elderly. The improvement is the strongest.

show abstract

Advances in Optimisation Algorithms and Techniques for Deep Learning

Cited by 13 publications

References 57 publications

Machine-Learning-Based Approaches for Multi-Level Sentiment Analysis of Romanian Reviews

Machine-Learning-Based Approaches for Multi-Level Sentiment Analysis of Romanian Reviews

Global solar radiation time series forecasting using different architectures of the multilayer perceptron model

A Chaotic Fish Swarm Algorithm-Based Model for Assessing the Mental Health Status of Older Adults

Contact Info

Product

Resources

About