Emmanuel Tavares Ferreira Affonso

Dr. Emmanuel Tavares Ferreira Affonso

orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
EAEmmanuel Tavares Ferreira Affonso
Emmanuel Tavares Ferreira Affonso
GMGray Farias Moita
Gray Farias Moita
Alisson Marques SilvaAlisson Marques Silva
Alisson Marques Silva
This paper introduces a new multiclass classifier called the evolving Fuzzy Classifier (eFC). Starting its knowledge base from scratch, the eFC structure evolves based on a clustering algorithm that can add, merge, delete, or update clusters (= rules) simultaneously while providing class predictions. The procedure to add clusters uses the procrastination idea to prevent outliers from affecting the quality of learning. Two pruning mechanisms are used to maintain a concise and compact structure. In the first, redundant clusters are merged based on a similarity measure, and in the second, obsolete and unrepresentative clusters are excluded based on an inactivity strategy. The center of the clusters is adjusted based on the mean value of the attributes. The eFC model was evaluated and compared with state-of-the-art evolving fuzzy systems on 8 randomly selected data streams from the UCI and Kaggle repositories. The experimental results indicate that the eFC outperforms or is at least comparable to alternative state-of-the-art models. Specifically, the eFC achieved an average accuracy of 7% to 37% higher than the competing classifiers. The results and comparisons demonstrate that the eFC is a promising alternative for classification tasks in non-stationary environments, offering good accuracy, a compact structure, low computational cost, and efficient processing time.

17 May 2017
Verified
Conference Paper
Open access
PDF file
EAElvis Assis
Elvis Assis
EAEmmanuel Tavares Ferreira Affonso
Emmanuel Tavares Ferreira Affonso
DRDemóstenes Zegarra Rodríguez
Demóstenes Zegarra Rodríguez
RRRenata Lopes Rosa
Renata Lopes Rosa
This work describes an information system regarding to the telecommunication service with the goal to inform to the telephony service users about the operators signal quality and the users’ opinion about the telecommunication operators. The system contains information of cellular base station locations and the results are crossed with sentiment analyses of sentences extracted from a social network, which are related to the telecommunication services. The study aims to detect complaints and dissatisfaction of users about a determined service, helping to make a relation between the users’ complains extracted from social network with the number of base stations in the same geographic area of the user's social network. Therefore, the proposed information system can be used by the National Telecommunications Agency of each country for monitoring the quality-of-service of cellular network operators.

27 September 2021 in Intelligent Decision Technologies
Verified
Journal Article
Closed access
PDF file
EAEmmanuel Tavares Ferreira Affonso
Emmanuel Tavares Ferreira Affonso
Alisson Marques SilvaAlisson Marques Silva
Alisson Marques Silva
GMGray Farias Moita
Gray Farias Moita
RCRodrigo Tomas Nogueira Cardoso
Rodrigo Tomas Nogueira Cardoso
Feature Selection (FS) is currently a very important and prominent research area. The focus of FS is to identify and to remove irrelevant and redundant features from large data sets in order to reduced processing time and to improve the predictive ability of the algorithms. Thus, this work presents a straightforward and efficient FS method based on the mean ratio of the attributes (features) associated with each class. The proposed filtering method, here called MRFS (Mean Ratio Feature Selection), has only equations with low computational cost and with basic mathematical operations such as addition, division, and comparison. Initially, in the MRFS method, the average from the data sets associated with the different outputs is computed for each attribute. Then, the calculation of the ratio between the averages extracted from each attribute is performed. Finally, the attributes are ordered based on the mean ratio, from the smallest to the largest value. The attributes that have the lowest values are more relevant to the classification algorithms. The proposed method is evaluated and compared with three state-of-the-art methods in classification using four classifiers and ten data sets. Computational experiments and their comparisons against other feature selection methods show that MRFS is accurate and that it is a promising alternative in classification tasks.