This work evaluates the compression techniques on ConvNeXt models in image classification tasks using the CIFAR-10 dataset. Structured pruning, unstructured pruning, and dynamic quantization methods are evaluated to reduce model size and computational complexity while maintaining accuracy. The experiments, conducted on cloud-based platforms and edge device, assess the performance of these techniques. Results show significant reductions in model size, with up to 75% reduction achieved using structured pruning techniques. Additionally, dynamic quantization achieves a reduction of up to 95% in the number of parameters. Fine-tuned models exhibit improved compression performance, indicating the benefits of pre-training in conjunction with compression techniques. Unstructured pruning methods reveal trends in accuracy and compression, with limited reductions in computational complexity. The combination of OTOV3 pruning and dynamic quantization further enhances compression performance, resulting 89.7% reduction in size, 95% reduction with number of parameters and MACs, and 3.8% increase with accuracy. The deployment of the final compressed model on edge device demonstrates high accuracy 92.5% and low inference time 20ms, validating the effectiveness of compression techniques for real-world edge computing applications.

To understand the public’s perception of COVID-19 tracing applications, previous studies were primarily based on exploratory research, surveys or machine learning methods, which are semantically weak and time-consuming. To increase the reliability of this analytical methodology, hybrid-based Twitter sentiment analysis can be applied. In this paper, we propose a hybrid model for sentiment analysis by using Valence Aware Dictionary for Sentiment Reasoning (VADER) + Support Vector Machine (SVM). We demonstrate from the numerical analysis that a VADER and SVM-based hybrid model provides the best performance with 82.3% accuracy, 0.84 precision, 0.83 recall and 0.82 F1-score. The use of hybrid-based methods is shown to be effective in analysing the public’s perception towards COVID-19 contact tracing applications using tweets collected from the UK, USA and India. Positive responses clearly outweighed negatives responses towards contact tracing, but this was contradicted by the low uptake of apps in all three nations. Our analysis, however, showed that neutral responses were 52% of the collected tweets; these tweets did not express positive or negative opinions, and subsequent tweets from the same users could not be verified, thus limiting the number of analyzed tweets available.