@cs.ubbcluj.ro
Department of Computer Science
Babes-Bolyai University
Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems, Multidisciplinary
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Adél Bajcsi, Anca Andreica, and Camelia Chira
SCITEPRESS - Science and Technology Publications
: Proper treatment of breast cancer is essential to increase survival rates. Mammography is a widely used, non-invasive screening method for breast cancer. A challenging task in mammogram analysis is to distinguish between tumors. In the current study, we address this problem using different feature extraction and classification methods. In the literature, numerous feature extraction methods have been presented for breast lesion classification, such as textural features, shape features, and wavelet features. In the current paper, we propose the use of shape features. In general, benign lesions have a more regular shape than malignant lesions. However, there are exceptions and in our experiments, we highlight the importance of a balanced split of these samples. Decision Tree and Random Forest methods are used for classification due to their simplicity and interpretability. A comparative analysis is conducted to evaluate the effectiveness of the classification methods. The best results were achieved using the Random Forest classifier with 96.12% accuracy using images from the Digital Dataset for Screening Mammography – DDSM.
Anikó Kopacz, Lehel Csató, and Camelia Chira
Elsevier BV
Claudia-Ioana Coste, Anca-Mirela Andreica, and Camelia Chira
SCITEPRESS - Science and Technology Publications
: Malicious links are becoming the main propagating vector for web-malware. They may lead to serious security issues, such as phishing, distribution of fake news and low-quality content, drive-by-downloads, and malicious code running. Malware link detection is a challenging domain because of the dynamics of the on-line environment, where web links and web content are always changing. Moreover, the detection should be fast and accurate enough that it will contribute to a better online experience. The present paper proposes to drive an experimental analysis on machine learning algorithms used in malicious web links detection. The algorithms chosen for analysis are Logistic Regression, Na¨ıve Bayes, Ada Boost, Gradient Boosted Tree, Linear Discriminant Analysis, Multi-layer Perceptron and Support Vector Machine with different kernel types. Our purpose is twofold. First, we compare these single algorithms run individually and calibrate their parameters. Secondly, we chose 10 models and used them in ensemble models. The results of these experiments show that the ensemble models reach higher metric scores than the individual models, improving the maliciousness prediction up to 96% precision.
Enol García González, José R. Villar, Camelia Chira, Enrique de la Cal, Luciano Sánchez, and Javier Sedano
Springer Nature Switzerland
Adél Bajcsi and Camelia Chira
Springer Nature Switzerland
Anikó Kopacz, Lehel Csató, and Camelia Chira
Springer Nature Switzerland
Stefania D. Iancu, Ramona G. Cozan, Andrei Stefancu, Maria David, Tudor Moisoiu, Cristiana Moroz-Dubenco, Adel Bajcsi, Camelia Chira, Anca Andreica, Loredana F. Leopold,et al.
Elsevier BV
Mădălina Dicu, Laura Dioşan, Anca Andreica, Camelia Chira, and Alin Cordoş
IEEE
Breast cancer is the most commonly diagnosed type of cancer. It is essential to classify patients as quickly as possible into groups with a high or low risk of cancer, to provide adequate treatment. This paper aims to address the impact of the parameters of convolutional neural networks in the binary classification of mammograms. In this paper, we treat two types of binary classification, namely: classification between normal and abnormal tissues, respectively classification between benign and malignant tumors. In the analysis, we investigate the correlation and impact of batch size and learning rate in increasing the performance of the proposed model. Following the experiments on the MIAS dataset, we concluded that for the treated problems, it is appropriate to choose a learning rate lower than 0.001. For the classification of tissues (normal/abnormal), we obtained the fact that training the model on a batch size of 32 brings the best results, namely an accuracy of 0.67, and for the classification of tumors (benign/malignant), it is more appropriate to use a batch size of 8, for which we obtained an accuracy of 0.63. For the best results configurations, we continued the experiments by investigating the impact of data augmentation. We have increased the number of training data by applying horizontal flip and rotation operations. Following these attempts, we noticed an improvement only for the tissue classification, for which we obtained an accuracy of 0.70.
Camelia Chira
IEEE
Network science is emerging as a vibrant research field with important applications in finance, biology, chemistry, physics, engineering and health. This short paper presents an overview of some challenging tasks related to the analysis of complex networks, including community detection, discovery of cycles and identification of important nodes. The solutions proposed for these important network analysis tasks engage Artificial Intelligence models and are briefly presented with an emphasis on their performance as well as the main related research questions. The analysis of financial networks is also discussed, showing the potential of using network science tools to discover financial cycles and paths.
Enol García, José R. Villar, Qing Tan, Javier Sedano, and Camelia Chira
IOS Press
Multi-robot path planning has evolved from research to real applications in warehouses and other domains; the knowledge on this topic is reflected in the large amount of related research published in recent years on international journals. The main focus of existing research relates to the generation of efficient routes, relying the collision detection to the local sensory system and creating a solution based on local search methods. This approach implies the robots having a good sensory system and also the computation capabilities to take decisions on the fly. In some controlled environments, such as virtual labs or industrial plants, these restrictions overtake the actual needs as simpler robots are sufficient. Therefore, the multi-robot path planning must solve the collisions beforehand. This study focuses on the generation of efficient collision-free multi-robot path planning solutions for such controlled environments, extending our previous research. The proposal combines the optimization capabilities of the A* algorithm with the search capabilities of co-evolutionary algorithms. The outcome is a set of routes, either from A* or from the co-evolutionary process, that are collision-free; this set is generated in real-time and makes its implementation on edge-computing devices feasible. Although further research is needed to reduce the computational time, the computational experiments performed in this study confirm a good performance of the proposed approach in solving complex cases where well-known alternatives, such as M* or WHCA, fail in finding suitable solutions.
Cristiana Moroz-Dubenco, Adél Bajcsi, Anca Andreica, and Camelia Chira
Elsevier BV
Enol García, José R. Villar, Camelia Chira, and Javier Sedano
Springer International Publishing
José R. Villar, Camelia Chira, Enrique de la Cal, Víctor M. González, Javier Sedano, and Samad B. Khojasteh
Elsevier BV
Attila Mester, Andrei Pop, Bogdan-Eduard-Mădălin Mursa, Horea Greblă, Laura Dioşan, and Camelia Chira
MDPI AG
The stability and robustness of a complex network can be significantly improved by determining important nodes and by analyzing their tendency to group into clusters. Several centrality measures for evaluating the importance of a node in a complex network exist in the literature, each one focusing on a different perspective. Community detection algorithms can be used to determine clusters of nodes based on the network structure. This paper shows by empirical means that node importance can be evaluated by a dual perspective—by combining the traditional centrality measures regarding the whole network as one unit, and by analyzing the node clusters yielded by community detection. Not only do these approaches offer overlapping results but also complementary information regarding the top important nodes. To confirm this mechanism, we performed experiments for synthetic and real-world networks and the results indicate the interesting relation between important nodes on community and network level.
Gabriela Czibula, Alexandra-Ioana Albu, Maria Iuliana Bocicor, and Camelia Chira
MDPI AG
Proteins are essential molecules, that must correctly perform their roles for the good health of living organisms. The majority of proteins operate in complexes and the way they interact has pivotal influence on the proper functioning of such organisms. In this study we address the problem of protein–protein interaction and we propose and investigate a method based on the use of an ensemble of autoencoders. Our approach, entitled AutoPPI, adopts a strategy based on two autoencoders, one for each type of interactions (positive and negative) and we advance three types of neural network architectures for the autoencoders. Experiments were performed on several data sets comprising proteins from four different species. The results indicate good performances of our proposed model, with accuracy and AUC values of over 0.97 in all cases. The best performing model relies on a Siamese architecture in both the encoder and the decoder, which advantageously captures common features in protein pairs. Comparisons with other machine learning techniques applied for the same problem prove that AutoPPI outperforms most of its contenders, for the considered data sets.
Adél Bajcsi, Anca Andreica, and Camelia Chira
Elsevier BV
Eduardo C. Paim, Ana L. C. Bazzan, and Camelia Chira
IEEE
An important problem in network science is finding relevant community structures in complex networks. A community structure is a partition of the network nodes into clusters or modules, such that each cluster is densely connected. Current community detection algorithms have time complexity, centralization, and scalability issues. In this paper, to solve this problem, we implement a multi-agent reinforcement learning algorithm that optimizes a quality metric known as modularity. We model each node of the network as an autonomous agent that can choose other nodes to form a cluster with. They receive a reward and learn a policy that maps actions to their values. Experiments on known real-world networks show results similar to other modularity optimization methods while providing answers for decentralization, data privacy, and scalability.
Samad Khojasteh, José Villar, Camelia Chira, Víctor González, and Enrique de la Cal
MDPI AG
Fall detection is a very important challenge that affects both elderly people and the carers. Improvements in fall detection would reduce the aid response time. This research focuses on a method for fall detection with a sensor placed on the wrist. Falls are detected using a published threshold-based solution, although a study on threshold tuning has been carried out. The feature extraction is extended in order to balance the dataset for the minority class. Alternative models have been analyzed to reduce the computational constraints so the solution can be embedded in smart-phones or smart wristbands. Several published datasets have been used in the Materials and Methods section. Although these datasets do not include data from real falls of elderly people, a complete comparison study of fall-related datasets shows statistical differences between the simulated falls and real falls from participants suffering from impairment diseases. Given the obtained results, the rule-based systems represent a promising research line as they perform similarly to neural networks, but with a reduced computational cost. Furthermore, support vector machines performed with a high specificity. However, further research to validate the proposal in real on-line scenarios is needed. Furthermore, a slight improvement should be made to reduce the number of false alarms.
Camelia-M. Pintea, Petrică C. Pop, and Camelia Chira
Springer Science and Business Media LLC
AbstractA well known $$\\mathcal{NP}$$ NP -hard problem called the generalized traveling salesman problem (GTSP) is considered. In GTSP the nodes of a complete undirected graph are partitioned into clusters. The objective is to find a minimum cost tour passing through exactly one node from each cluster. An exact exponential time algorithm and an effective meta-heuristic algorithm for the problem are presented. The meta-heuristic proposed is a modified Ant Colony System (ACS) algorithm called reinforcing Ant Colony System which introduces new correction rules in the ACS algorithm. Computational results are reported for many standard test problems. The proposed algorithm is competitive with the other already proposed heuristics for the GTSP in both solution quality and computational time.
Marius Joldos and Camelia Chira Technical
IEEE
The problem of community detection in complex networks is of high interest in many application domains including sociology, biology, mathematics and economy. Given a set of nodes and links between them, the aim of the problem is to find a grouping of nodes such that a strong community has dense intra-connections and sparse outside community links. In this paper, a coarse-grained evolutionary algorithm (EA) is developed to address this challenging problem. Several populations of potential solutions are evolved in parallel in an island model and periodically exchange certain individuals. Each population can be evolved by a different fitness function and several approaches to evaluate the community structure are considered in the current paper. Experiments are performed for real-world complex networks and results are analysed based on the normalized mutual information between the detected and the known community structure. Comparisons with the standard version of the EA based on different fitness functions are performed and the results confirm a good performance of the parallel EA in terms of solution quality and computational time.
Javier Sedano, Silvia González, Camelia Chira, Álvaro Herrero, Emilio Corchado, and José Ramón Villar
Oxford University Press (OUP)
In recent years, mobile devices such as smartphones, tablets and wearables have become the new paradigm of user–computer interaction. The increasing use and adoption of such devices is also leading to an increased number of potential security risks. The spread of mobile malware, particularly on popular and open platforms such as Android, has become a major concern. This paper focuses on the bad-intentioned Android apps by addressing the problem of selecting the key features of such software that support the characterization of such malware. The accurate detection and characterization of this software is still an open challenge, mainly due to its ever-changing nature and the open distribution channels of Android apps. Maximum relevance minimum redundancy and evolutionary algorithms guided by information correlation measures have been applied for feature selection on the well-known Android Malware Genome (Malgenome) dataset, attaining interesting results on the most informative features for the characterization of representative families of existing Android malware.
Javier Sedano, Camelia Chira, Silvia González, Álvaro Herrero, Emilio Corchado, and José Ramón Villar
Springer International Publishing
Cristian Jora and Camelia Chira
IEEE
The discovery of communities in complex networks is a challenging problem with various applications in the real world. Classic examples of networks include transport networks, the immune system, human brain and social networks. Given a certain grouping of nodes into communities, a good measure is needed to evaluate the quality of the community structure based on the definition that a strong community has dense intra-connections and sparse outside community links. This paper investigates several fitness functions in an evolutionary approach to community detection in complex networks. Moreover, these fitness functions are used to study dynamic networks using an extended evolutionary algorithm designed to handle changes in the network structure. Computational experiments are performed for several real-world networks which have a known community structure and thus can be evaluated. The obtained results confirm the ability of the proposed method to efficiently detect communities for both static and dynamic complex networks.
Camelia Chira, Javier Sedano, José R. Villar, Monica Camara, and Carlos Prieto
Springer Science and Business Media LLC
Camelia Chira, Ana L. C. Bazzan, and Rosaldo J. F. Rossetti
IEEE
Existing approaches to traffic assignment focus mainly on approximating the user equilibrium. However, nowadays, with the increasing number of traffic information reaching drivers, traffic authorities have a unique opportunity to try to recommend route choices that are as much as possible aligned with the system optimum. In this paper, we formulate traffic assignment as a multi-objective optimization problem and engage an evolutionary approach to find route solutions for all users in the network. The aim is to discover a good approximation of an optimal distribution of vehicles to alternative routes between their origin and destination, from the perspective of the overall system, while still considering individual needs. Several multi-objective models are defined for this purpose and integrated in a nondominated sorting genetic algorithm. Computational experiments performed support the ability of the proposed approach to detect efficient route assignments in terms of network performance, while also considering the user perspective. efficient route assignments in terms of network performance, while also considering the user perspective.