A5041843953- Metaheuristic Optimization Algorithms Research
- Advanced Clustering Algorithms Research
- Scheduling and Timetabling Solutions
- Biometric Identification and Security
- Evolutionary Algorithms and Applications
- Data Mining Algorithms and Applications
- User Authentication and Security Systems
- Dermatoglyphics and Human Traits
- Text and Document Classification Technologies
- Cell Image Analysis Techniques
- Constraint Satisfaction and Optimization
- Computational Drug Discovery Methods
- Vehicle Routing Optimization Methods
- Advanced Multi-Objective Optimization Algorithms
- Brain Tumor Detection and Classification
- Alzheimer's disease research and treatments
- Video Analysis and Summarization
- Web Data Mining and Analysis
- Forensic and Genetic Research
- Advanced Computing and Algorithms
- Parkinson's Disease Mechanisms and Treatments
- Remote-Sensing Image Classification
- Sentiment Analysis and Opinion Mining
- Gaze Tracking and Assistive Technology
- Advanced Authentication Protocols Security
Institute for the Future
2025
Tecnológico de Monterrey
2017-2025
Universidad de Ciego de Ávila
2009-2015
National Institute of Astrophysics, Optics and Electronics
2014-2015
Improving fingerprint matching algorithms is an active and important research area in recognition. Algorithms based on minutia triplets, matcher family, present some drawbacks that impact their accuracy, such as dependency to the order of minutiae feature, insensitivity reflection directions relative sides triangle. To alleviate these drawbacks, we introduce this paper a novel algorithm, named M3gl. This algorithm contains three components: new feature representation containing...
Nowadays, the international scientific community of machine learning has an enormous campaign in favor creating understandable models instead black-box models. The main reason is that experts application area are showing reluctance due to cannot be understood by them, and consequently, their results difficult explained. In unsupervised problems, where have not labeled objects, obtaining explanation necessary because specialists need understand both applied model as well obtained for finding...
Hyper-heuristics are a novel tool. They deal with complex optimization problems where standalone solvers exhibit varied performance. Among such tool reside selection hyper-heuristics. By combining the strengths of each solver, this kind hyper-heuristic offers more robust However, their effectiveness is highly dependent on 'features' used to link them problem that being solved. Aiming at enhancing hyper-heuristics, in paper we propose two types transformation: explicit and implicit. The first...
Motor imagery is a complex mental task that represents muscular movement without the execution of action, involving cognitive processes motor planning and sensorimotor proprioception body. Since has similar behavior to process, it can be used create rehabilitation routines for patients with some skill impairment. However, due nature this task, its complicated. Hence, classification these signals in scenarios such as brain-computer interface systems tends have poor performance. In work, we...
Fingerprint verification has become one of the most active research areas nowadays. A key component an accurate fingerprint system is matching algorithm. An algorithm uses a robust representation. In this paper, we introduce m-triplets, new minutiae triplet representation and similarity for verification. The proposed shifts triplets to find best correspondence it rules that discard not without comparing whole representation; hence, achieving high speed. To test quality introduced similarity,...
Hyper-heuristics are powerful search methodologies that can adapt to different kinds of problems. One element paramount importance, however, is the selection module they incorporate. Traditional approaches define a set features for characterizing problem and, thus, how best solve it. However, some may vary nonlinearly as solver progresses, requiring higher resolution in specific areas feature domain. This work focuses on assessing advantage using transformations improve given consequence,...
Many of the works conducted on algorithm selection strategies—methods that choose a suitable solving method for particular problem—start from scratch since only few investigations reusable components such methods are found in literature. Additionally, researchers might unintentionally omit some implementation details when documenting strategy. This makes it difficult others to reproduce behavior obtained by an approach. To address these problems, we propose rely existing techniques Machine...
In clustering, providing an explanation of the results is important task. Pattern-based clustering algorithms return a set patterns that describe objects grouped in each cluster. The most recent proposed this approach have high computational cost stage, m aking them non suitable when huge amount are extracted from dataset. paper, we introduce algorithm for extracting small subset useful clustering. extracts collection trees generated through new induction procedure. Experimental show...
Hyper-heuristics have emerged as an important strategy for combining the strengths of different heuristics into a single method. Although hyper-heuristics been found to be successful in many scenarios, little attention has paid subsets that these methods manage and apply. In several cases, can interfere with each other harmful search. Thus, obtaining information about differences among heuristics, how they contribute search process is very important. The main contribution this paper...
In recent years, evolutionary algorithms have been found to be effective and efficient techniques train support vector machines (SVMs) for binary classification problems while multiclass neglected. This paper proposes CMOE-SVM: Cooperative Multi-Objective Evolutionary SVMs problems. CMOE-SVM enables handle via co-evolutionary optimization, by breaking down the original M-class problem into M simpler ones, which are optimized simultaneously in a cooperative manner. Furthermore, can explicitly...
Nowadays, heuristics represent a commonly used alternative to solve complex optimization problems. This, however, has given rise the problem of choosing most effective heuristic for problem. In recent years, one strategies this task been hyper-heuristics, which aim at selecting/generating wide range Most existing selection hyper-heuristics attempt recommend only instance. However, some classes problems, more than can be suitable. With premise, in paper, we address issue through an...
Cluster analysis is a data mining tool for searching patterns automatically on different types of data. However, it not always clear which clustering criterion would be the most accurate as this decision domain-dependent. This paper focuses design single-objective evolutionary algorithm that generates solutions are less biased towards one cluster structure. work's motivation starts from idea good partition induces well-trained classifier. The resulting using classifiers aims to enhance...
This paper describes the design of 2017 RedICA: Text-Image Matching (RICATIM) challenge, including dataset generation, a complete analysis results, and descriptions top-ranked developed methods. The academic challenge explores feasibility novel binary image classification scenario, where each instance corresponds to concatenation learned representations an word. Instances are labeled as positive if word is relevant for describing visual content image, negative otherwise. approach problem...