RESUMEN
Chimp optimization algorithm (ChOA) is a robust nature-inspired technique, which was recently proposed for addressing real-world challenging engineering problems. Due to the novelty of the ChOA, there is room for its improvement. Recognition and classification of marine mammals using artificial neural networks (ANNs) are high-dimensional challenging problems. In order to address this problem, this paper proposed the using of ChOA as ANN's trainer. However, evolving ANNs using metaheuristic algorithms suffers from high complexity and processing time. In order to address this shortcoming, this paper proposes the fuzzy logic to adjust the ChOA's control parameters (Fuzzy-ChOA) for tuning the relationship between exploration and exploitation phases. In this regard, we collect underwater marine mammals sounds and then produce an experimental dataset. After pre-processing and feature extraction, the ANN is used as a classifier. Besides, for having a fair comparison, we used a benchmark audio database of marine mammals. The comparison algorithms include ChOA, coronavirus optimization algorithm, harris hawks optimization, black widow optimization algorithm, Kalman filter benchmark algorithms, and also comparative benchmarks include convergence speed, local optimal avoidance ability, classification rate, and receiver operating characteristics (ROC). The simulation results show that the proposed fuzzy model can tune the boundary between the exploration and extraction phases. The convergence curve and ROC confirm that the convergence rate and performance of the designed recognizer are better than benchmark algorithms.
RESUMEN
There is a variety of fuzzy classifiers, one of which is Adaptive Neuro-Fuzzy Inference system (ANFIS) classifier. One of the main challenges in designing such data classifiers is selection of effective and appropriate type and location of membership functions and its training method to reduce the classification error. In this paper, a new technique (based on intelligent methods) is presented and implemented to select and locate the membership functions and simultaneous training using a new method based on Inclined Planes System Optimization (IPO) to minimize errors of an ANFIS classifier for the first time. The presented method is evaluated for classification of data sets with different reference classes and different length feature vectors, which have acceptable complexity. According to the results of the research, the presented method has a higher level of accuracy and efficiency in selecting the type and location of membership functions (based on intelligent methods) and simultaneous training with IPO, compared to other methods, such as particle swarm optimization, genetic algorithm, differential evolution, and ACOR algorithms.
Asunto(s)
Lógica Difusa , Heurística , Redes Neurales de la Computación , Reconocimiento de Normas Patrones AutomatizadasRESUMEN
Recently, professional team sport organizations have invested their resources to analyze their own and opponents' performance. So, developing methods and algorithms for analyzing team sports has become one of the most popular topics among data scientists. Analyzing football is hard because of its complexity, number of events in each match, and constant flow of circulation of the ball. Finding roles of players with the purpose of analyzing the performance of a team or making a meaningful comparison between players is crucial. In this article, an automatic big data clustering method, based on a swarm intelligence algorithm, is proposed to automatically cluster the data set of players' performance centers in different matches and extract different kinds of roles in football. The proposed method created using particle swarm optimization algorithm has two phases. In the first phase, the algorithm searches the solution space to find the number of clusters and, in the second phase, it finds the positions of the centroids. To show the effectiveness of the algorithm, it is tested on six synthetic data sets and its performance is compared with two other conventional clustering methods. After that, the algorithm is used to find clusters of a data set containing 93,000 objects, which are the centers of players' performance in about 4900 matches in different European leagues.