The immense popularity of convolutional neural network (CNN) models has sparked a growing interest in optimizing their hyperparameters. Discovering the ideal values for hyperparameters to achieve optimal CNN training is a complex and time-consuming task, often requiring repetitive numerical experiments. As a result, significant attention is currently being devoted to developing methods aimed at tailoring hyperparameters for specific CNN models and classification tasks. While existing optimization methods often yield favorable image classification results, they do not provide guidance on which hyperparameters are worth optimizing, the appropriate value ranges for those hyperparameters, or whether it is reasonable to use a subset of training data for the optimization process. This work is focused on the optimization of hyperparameters during transfer learning, with the goal of investigating how different optimization methods and hyperparameter selections impact the performance of fine-tuned models. In our experiments, we assessed the importance of various hyperparameters and identified the ranges within which optimal CNN training can be achieved. Additionally, we compared four hyperparameter optimization methods—grid search, random search, Bayesian optimization, and the Asynchronous Successive Halving Algorithm (ASHA). We also explored the feasibility of fine-tuning hyperparameters using a subset of the training data. By optimizing the hyperparameters, we observed an improvement in CNN classification accuracy of up to 6%. Furthermore, we found that achieving a balance in class distribution within the subset of data used for parameter optimization is crucial in establishing the optimal set of hyperparameters for CNN training. The results we obtained demonstrate that hyperparameter optimization is highly dependent on the specific task and dataset at hand.

This paper presents the application of a genetic algorithm (GA) and a support vector machine (SVM) to the recognition of blood cells based on the image of the bone marrow aspirate. The main task of the GA is the selection of the features used by the SVM in the final recognition and classification of cells. The automatic recognition system has been developed, and the results of its numerical verification are presented and discussed. They show that the application of the GA is a powerful tool for the selection of the diagnostic features, leading to a significant improvement of the accuracy of the whole system.

Ensemble Neural Network Approach for Accurate Load Forecasting in a Power System: The paper presents an improved method for 1-24 hours load forecasting in the power system, integrating and combining different neural forecasting results by an ensemble system. We will integrate the results of partial predictions made by three solutions, out of which one relies on a multilayer perceptron and two others on self-organizing networks of the competitive type. As the expert system we will apply different integration methods: simple averaging, SVD based weighted averaging, principal component analysis and blind source separation. The results of numerical experiments, concerning forecasting the hourly load for the next 24 hours of the Polish power system, will be presented and discussed. We will compare the performance of different ensemble methods on the basis of the mean absolute percentage error, mean squared error and maximum percentage error. They show a significant improvement of the proposed ensemble method in comparison to the individual results of prediction. The comparison of our work with the results of other papers for the same data proves the superiority of our approach.

The paper discusses methods of data mining for prediction of air pollution. Two tasks in such a problem are important: generation and selection of the prognostic features, and the final prognostic system of the pollution for the next day. An advanced set of features, created on the basis of the atmospheric parameters, is proposed. This set is subject to analysis and selection of the most important features from the prediction point of view. Two methods of feature selection are compared. One applies a genetic algorithm (a global approach), and the other-a linear method of stepwise fit (a locally optimized approach). On the basis of such analysis, two sets of the most predictive features are selected. These sets take part in prediction of the atmospheric pollutants PM10, SO2, NO2 and O3. Two approaches to prediction are compared. In the first one, the features selected are directly applied to the random forest (RF), which forms an ensemble of decision trees. In the second case, intermediate predictors built on the basis of neural networks (the multilayer perceptron, the radial basis function and the support vector machine) are used. They create an ensemble integrated into the final prognosis. The paper shows that preselection of the most important features, cooperating with an ensemble of predictors, allows increasing the forecasting accuracy of atmospheric pollution in a significant way.

Fungi and oomycetes release volatiles into their environment which could be used for olfactory detection and identification of these organisms by electronic-nose (e-nose). The aim of this study was to survey volatile compound emission using an e-nose device and to identify released molecules through solid phase microextraction–gas chromatography/mass spectrometry (SPME–GC/MS) analysis to ultimately develop a detection system for fungi and fungi-like organisms. To this end, cultures of eight fungi (Armillaria gallica, Armillaria ostoyae, Fusarium avenaceum, Fusarium culmorum, Fusarium oxysporum, Fusarium poae, Rhizoctonia solani, Trichoderma asperellum) and four oomycetes (Phytophthora cactorum, P. cinnamomi, P. plurivora, P. ramorum) were tested with the e-nose system and investigated by means of SPME-GC/MS. Strains of F. poae, R. solani and T. asperellum appeared to be the most odoriferous. All investigated fungal species (except R. solani) produced sesquiterpenes in variable amounts, in contrast to the tested oomycetes strains. Other molecules such as aliphatic hydrocarbons, alcohols, aldehydes, esters and benzene derivatives were found in all samples. The results suggested that the major differences between respective VOC emission ranges of the tested species lie in sesquiterpene production, with fungi emitting some while oomycetes released none or smaller amounts of such molecules. Our e-nose system could discriminate between the odors emitted by P. ramorum, F. poae, T. asperellum and R. solani, which accounted for over 88% of the PCA variance. These preliminary results of fungal and oomycete detection make the e-nose device suitable for further sensor design as a potential tool for forest managers, other plant managers, as well as regulatory agencies such as quarantine services.