Whale optimization algorithm (WOA) tends to fall into the local optimum and fails to converge quickly in solving complex problems. To address the shortcomings, an improved WOA (QGBWOA) is proposed in this work. First, quasi-opposition-based learning is introduced to enhance the ability of WOA to search for optimal solutions. Second, a Gaussian barebone mechanism is embedded to promote diversity and expand the scope of the solution space in WOA. To verify the advantages of QGBWOA, comparison experiments between QGBWOA and its comparison peers were carried out on CEC 2014 with dimensions 10, 30, 50, and 100 and on CEC 2020 test with dimension 30. Furthermore, the performance results were tested using Wilcoxon signed-rank (WS), Friedman test, and post hoc statistical tests for statistical analysis. Convergence accuracy and speed are remarkably improved, as shown by experimental results. Finally, feature selection and multi-threshold image segmentation applications are demonstrated to validate the ability of QGBWOA to solve complex real-world problems. QGBWOA proves its superiority over compared algorithms in feature selection and multi-threshold image segmentation by performing several evaluation metrics.

The resting-state functional magnetic resonance imaging (rs-fMRI) is a pivotal tool that can reveal brain dysfunction in the computer-aided diagnosis of the autism spectrum disorder (ASD). However, the instability of data collection devices, complexity of pathogenesis, and ambiguity in the causes of the disease always introduce considerable uncertainty in identifying ASD using rs-fMRI. Due to the strong ability of Takagi–Sugeno–Kang fuzzy inference systems (TSK FISs) in handling the uncertainty of knowledge and expression, we build an ASD classification model based on TSK FISs and further propose a novel multicenter ASD classification method FCG-MTGS-TSK. Specifically, the correlation information of multiple imaging centers is considered by introducing multitask group sparse learning, and the features across multiple imaging centers are thus jointly selected. An augmented lagrange multiplier (ALM) method is further developed to find the optimal solution of the model. Compared with the other existing methods, the proposed method has the advantages of strong interpretability and high classification accuracy. The experimental results also identify the most discriminative functional connectivity in multicenter ASD classification.

In recent years, a range of novel and pseudonovel optimization algorithms has been proposed for solving engineering problems. Swarm intelligence optimization algorithms (SIAs) have become popular methods, and the whale optimization algorithm (WOA) is one of the highly discussed SIAs. However, regardless of novelty concerns about this method, the basic WOA is a weak method compared to top differential evolutions and particle swarm variants, and it suffers from the problem of poor initial population quality and slow convergence speed. Accordingly, in this paper, to increase the diversity of WOA versions and enhance the performance of WOA, a new WOA variant, named LXMWOA, is proposed, and based on the Lévy initialization strategy, the directional crossover mechanism, and the directional mutation mechanism. Specifically, the introduction of the Lévy initialization strategy allows initial populations to be dynamically distributed in the search space and enhances the global search capability of the WOA. Meanwhile, the directional crossover mechanism and the directional mutation mechanism can improve the local exploitation capability of the WOA. To evaluate its performance, using a series of functions and three models of engineering optimization problems, the LXMWOA was compared with a broad array of competitive optimizers. The experimental results demonstrate that the LXMWOA is significantly superior to its exploration and exploitation capability peers. Therefore, the proposed LXMWOA has great potential to be used for solving engineering problems.

Immunotherapy is an effective way to treat non-small cell lung cancer (NSCLC). The efficacy of immunotherapy differs from person to person and may cause side effects, making it important to predict the efficacy of immunotherapy before surgery. Radiomics based on machine learning has been successfully used to predict the efficacy of NSCLC immunotherapy. However, most studies only consider the radiomic features of the individual patient, ignoring the inter-patient correlations. Besides, they usually concatenate different features as the input of a single-view model, failing to consider the complex correlation among features of multiple types. To this end, we propose a multi-view adaptive weighted graph convolutional network (MVAW-GCN) for the prediction of NSCLC immunotherapy efficacy. Specifically, we divide the radiomic features into several views according to the type of the filtered images they extracted from. We construct a graph in each view based on the radiomic features and phenotypic information. An attention mechanism is introduced to automatically assign weights to each view. Considering the view-shared and view-specific knowledge of radiomic features, we propose separable graph convolution that decomposes the output of the last convolution layer into two components, i.e., the view-shared and view-specific outputs. We maximize the consistency and enhance the diversity among different views in the learning procedure. The proposed MVAW-GCN is evaluated on 107 NSCLC patients, including 52 patients with valid efficacy and 55 patients with invalid efficacy. Our method achieved an accuracy of 77.27% and an area under the curve (AUC) of 0.7780, indicating its effectiveness in NSCLC immunotherapy efficacy prediction.

Thyroid nodule is a clinical disorder with a high incidence rate, with large number of cases being detected every year globally. Early analysis of a benign or malignant thyroid nodule using ultrasound imaging is of great importance in the diagnosis of thyroid cancer. Although the b‐mode ultrasound can be used to find the presence of a nodule in the thyroid, there is no existing method for an accurate and automatic diagnosis of the ultrasound image. In this pursuit, the present study envisaged the development of an ultrasound diagnosis method for the accurate and efficient identification of thyroid nodules, based on transfer learning and deep convolutional neural network. Initially, the Total Variation‐ (TV‐) based self‐adaptive image restoration method was adopted to preprocess the thyroid ultrasound image and remove the boarder and marks. With data augmentation as a training set, transfer learning with the trained GoogLeNet convolutional neural network was performed to extract image features. Finally, joint training and secondary transfer learning were performed to improve the classification accuracy, based on the thyroid images from open source data sets and the thyroid images collected from local hospitals. The GoogLeNet model was established for the experiments on thyroid ultrasound image data sets. Compared with the network established with LeNet5, VGG16, GoogLeNet, and GoogLeNet (Improved), the results showed that using GoogLeNet (Improved) model enhanced the accuracy for the nodule classification. The joint training of different data sets and the secondary transfer learning further improved its accuracy. The results of experiments on the medical image data sets of various types of diseased and normal thyroids showed that the accuracy rate of classification and diagnosis of this method was 96.04%, with a significant clinical application value.