The traditional particle swarm optimization (PSO) algorithm often exhibits defects such as of slow convergence and easily falling into a local optimum. To overcome these problems, this paper proposes an enhanced variant featuring adaptive selection. Initially, a composite chaotic mapping model integrating Logistic and Sine mappings is employed to initialize the population for diversity and exploration capability. Subsequently, the global and local search capabilities of the algorithm are balanced through the introduction of adaptive inertia weights. The population is then divided into three subpopulations—elite, ordinary, and inferior particles—based on their fitness values, with each group employing a distinct position update strategy. Finally, a particle mutation strategy is incorporated to avoid convergence to local optima. Experimental results demonstrate that our algorithm outperforms existing algorithms on the standard benchmark functions. In practical engineering applications, our algorithm also has demonstrated better performance than other meta heuristic algorithms.

With the exponential growth in online news, hotspot classification has become increasingly important. Although traditional machine-learning-based text categorisation methods, such as plain Bayes, support vector machines (SVMs), and classification trees provide some interpretability, they are often incompetent in handling complex semantic relations, which often leads to poor classification accuracy. To address this issue, we introduce a novel deep learning text classification method based on the convolutional neural network(CNN) and long short-term memory (LSTM) and Attention mechanism, which accurately predicts news popularity by combining the feature extraction ability of the CNN, the sequence modelling ability of LSTM, and the weighted summation ability of the attention mechanism. Experimental results show that compared with other deep learning models, the proposed method has higher accuracy and can better address the context of text data and the problem of poor classification accuracy. Along with model selection, feature engineering is also the key to improving classification accuracy. In this study, we propose a plain Bayesian text classification model based on feature extraction, using word embeddings to convert text into richer vector representations, and then combine it with the different distributions of plain Bayes, proving that the polynomial plain Bayes is the most suitable model for text classification. Therefore, we added the feature word classification expressiveness index to improve the TF-IDF feature extraction, resulting in a classification accuracy of 96%. This demonstrates that the effectiveness of the classification reaches 96.3%, which demonstrates that the improved model is better at understanding and classifying text.