Fig 1 - available via license: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International

Content may be subject to copyright.

Source publication

Melanoma is a malignant and aggressive type of skin cancer. This paper describes an effective method for detection of melanoma. A hybrid classification algorithm was developed by using the SVM algorithm and a heuristic optimization algorithm. In this algorithm, the SVM algorithm which uses a Gaussian Radial Basis Function (RBF) was enhanced by the...

## Similar publications

Fault diagnosis of industrial process has long been a challenging issue owing to the industrial system that exhibits nonlinearity, coupled parameters and time-varying in the production process. This paper presents a novel dynamic fault diagnosis model (AUKF-RBF) based on radial basis function (RBF) neural network for Tennessee Eastman (TE) industri...

## Citations

... This MLP approach is more efficient and reliable than traditional methods, but it comes at a high cost and requires an importable detector system. The SVM with bacterial colony optimization (BCO) algorithm was introduced byİlkin et al. [11]. The PH2 and ISIC datasets with10 cross-fold validation are used to examine the evaluation of the SVM-BCO algorithm for malignant melanoma classification.The processing time was 11.9 s and the AUC was 98%, but the feature dimensionality was higher. ...

The most aggressive and malignant type of skin cancer is melanoma. When input data is in the form of images, image processing plays a vital role to detect and classifying cancer in the human body. Existing research discovered many weaknesses in complex data models, such as higher feature dimensionality, which required more data for training, resulting in lower detection accuracy, higher computational difficulties, portability, and processing time. Hence, we introduced fractal neural network-based galactic swarm optimization (FNN–GSO) algorithm for the detection of malignant melanoma such as superficial spreading, nodular, and lentigo malignant melanoma. The main aim of this work is to apply a deep learning technique to classify skin lesions for effective treatment and prognostication instead of the gold standard excision biopsy, which is currently used to diagnose this condition. Expert analysis, time consumption, and expensive processing associated with malignant melanoma classification and prediction are minimized in this manner. The proposed work involves four major components such as pre-processing, segmentation, feature extraction, and classification. The raw input images are pre-processed thereby the noise removal and contrast level enhancement are carried out. An adaptive watershed segmentation algorithm performs malignant melanoma segmentation. Following that, image features such as are extracted correctly using the DWT–GLCM feature extraction model. Finally, the malignant melanoma classifications are performed using the FNN–GSO algorithm. The proposed method's performance is evaluated using MATLAB software and different evaluation parameters. The proposed FNN–GSO algorithm demonstrates better classification results than other existing methods.

... In this experiment, precious, recall, false positive rate (FPR), and F-measure are selected as evaluation metrics of the proposed method, which are calculated as follows [22,23]: ...

A sports-assisted education method based on a support vector machine (SVM) is proposed to address the problem of complex and variable sports actions leading to easy ghosting of target detection and high dimensionality of feature extraction, which reduces the low accuracy of sports action recognition. The ViBe target detection algorithm is improved by using Wronskian function and the “4-linked algorithm” seed filling algorithm, which effectively solves the ghosting problem and obtains clearer human sports targets. By using the genetic algorithm to fuse the eight-star model with sports action features extracted by the Zernike moment, redundant features are reduced and differentiability between different classes is ensured. Sports action classification was achieved by using a one-to-one construction of an SVM classifier. The results show that the proposed method can effectively recognize sports movements with an average recognition accuracy of more than 96%, which can assist physical education and has a certain practical application value.

... (2) Assign flight direction and distance to all fruit flies and olfactory search is utilized to update [23] X i � X axis + rand(FR), ...

To improve the ability of market to avoid and prevent credit risk and strengthen the awareness of market risk early warning, SMOTE is used to process the unbalanced sample, and fruit fly optimization algorithm (FOA) is utilized to optimize the parameters of support vector machine (SVM), and thus an improved SVM market risk early warning model is proposed. The simulation results show that the proposed model has excellent stability and generalization ability, and it can predict market credit risk accurately. Compared with the prediction model based on FOA-SMOTE-BP and FOA-SMOTE-Logit, the proposed model performs better on the indicators of G value, F value, and AUC value, which provides a reference for market credit risk prediction.

... As convex quadratic programming can only be solved by a global optimal solution, which makes the process of solution simple, the global optimal solution can be derived by calculating the extrema. In solving convex quadratic programming, a combination of structural and empirical risk needs to be considered, and we can get the following equation after the Lagrangian function is introduced into equations (3)-(4) based on the Lagrangian duality [22]: ...

This paper presents a settlement prediction method based on PSO optimized SVM for improving the accuracy of foundation pit settlement prediction. Firstly, the method uses the SA algorithm to improve the traditional PSO algorithm, and thus, the overall optimization-seeking ability of the PSO algorithm is improved. Secondly, the improved PSO algorithm is used to train the SVM algorithm. Finally, the optimal SVM model is obtained, and the trained model is used in foundation pit settlement prediction. The results suggest that the settling results obtained from the optimized model are closer to the actual values and also more advantageous in indicators such as RMSE. The fitting value R2 = 0.9641, which is greater, indicates a better fitting effect. Thus, it is indicated that the improvement method is feasible.

Skin cancer affects the lives of millions of people every year, as it is considered the most popular form of cancer. In the USA alone, approximately three and a half million people are diagnosed with skin cancer annually. The survival rate diminishes steeply as the skin cancer progresses. Despite this, it is an expensive and difficult procedure to discover this cancer type in the early stages. In this study, a threshold-based automatic approach for skin cancer detection, classification, and segmentation utilizing a meta-heuristic optimizer named sparrow search algorithm (SpaSA) is proposed. Five U-Net models (i.e., U-Net, U-Net++, Attention U-Net, V-net, and Swin U-Net) with different configurations are utilized to perform the segmentation process. Besides this, the meta-heuristic SpaSA optimizer is used to perform the optimization of the hyperparameters using eight pre-trained CNN models (i.e., VGG16, VGG19, MobileNet, MobileNetV2, MobileNetV3Large, MobileNetV3Small, NASNetMobile, and NASNetLarge). The dataset is gathered from five public sources in which two types of datasets are generated (i.e., 2-classes and 10-classes). For the segmentation, concerning the “skin cancer segmentation and classification” dataset, the best reported scores by U-Net++ with DenseNet201 as a backbone architecture are 0.104, 94.16%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$94.16\%$$\end{document}, 91.39%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$91.39\%$$\end{document}, 99.03%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.03\%$$\end{document}, 96.08%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$96.08\%$$\end{document}, 96.41%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$96.41\%$$\end{document}, 77.19%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$77.19\%$$\end{document}, 75.47%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$75.47\%$$\end{document} in terms of loss, accuracy, F1-score, AUC, IoU, dice, hinge, and squared hinge, respectively, while for the “PH2” dataset, the best reported scores by the Attention U-Net with DenseNet201 as backbone architecture are 0.137, 94.75%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$94.75\%$$\end{document}, 92.65%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$92.65\%$$\end{document}, 92.56%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$92.56\%$$\end{document}, 92.74%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$92.74\%$$\end{document}, 96.20%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$96.20\%$$\end{document}, 86.30%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$86.30\%$$\end{document}, 92.65%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$92.65\%$$\end{document}, 69.28%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$69.28\%$$\end{document}, and 68.04%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$68.04\%$$\end{document} in terms of loss, accuracy, F1-score, precision, sensitivity, specificity, IoU, dice, hinge, and squared hinge, respectively. For the “ISIC 2019 and 2020 Melanoma” dataset, the best reported overall accuracy from the applied CNN experiments is 98.27%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$98.27\%$$\end{document} by the MobileNet pre-trained model. Similarly, for the “Melanoma Classification (HAM10K)” dataset, the best reported overall accuracy from the applied CNN experiments is 98.83%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$98.83\%$$\end{document} by the MobileNet pre-trained model. For the “skin diseases image” dataset, the best reported overall accuracy from the applied CNN experiments is 85.87%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$85.87\%$$\end{document} by the MobileNetV2 pre-trained model. After computing the results, the suggested approach is compared with 13 related studies.

The dealers are responsible for connecting customers to manufacturers in the automotive industry. An effective relationship between the dealers and the manufacturers can lead to customer satisfaction and loyalty. This research provides an approach to evaluate factors affecting the dealers-manufacturer relationship and develop a solution for predicting how manufacturers can continue or terminate the cooperation with their dealers in the after-sales service network, which is less discussed in the literature. We used predictive analytics tools and classification methods to predict dealers’ cooperation as a dependent variable in two classes based on a 2-year real-life collected data. Our results revealed that dealer performance violations, changing dealer principles, and the services revenue have the most effect on the cooperation of dealers in the automotive after-sales service network. Also, this research proposed a dealer management solution for manufacturers to show them the correct decisions for the dealers’ cooperation at the right time. This solution can be used in real-life cases in after-sales service networks to prevent using redundant resources, improve the weak performance of dealers, avoid unnecessary termination of cooperation, avoid the costs of attracting new dealers, and decrease customer dissatisfaction.