Time filter

Source Type

Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Sotoca J.M.,Jaume I University | Garcia V.,Jaume I University | Valdovinos R.M.,Valle de México University
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

The class imbalance problem has been considered a critical factor for designing and constructing the supervised classifiers. In the case of artificial neural networks, this complexity negatively affects the generalization process on under-represented classes. However, it has also been observed that the decrease in the performance attainable of standard learners is not directly caused by the class imbalance, but is also related with other difficulties, such as overlapping. In this work, a new empirical study for handling class overlap and class imbalance on multi-class problem is described. In order to solve this problem, we propose the joint use of editing techniques and a modified MSE cost function for MLP. This analysis was made on a remote sensing data . The experimental results demonstrate the consistency and validity of the combined strategy here proposed. © 2011 Springer-Verlag.


Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Garcia V.,Jaume I University | Marques A.I.,Jaume I University | Sanchez J.S.,Jaume I University | Antonio-Velazquez J.A.,Tecnologico de Estudios Superiores de Jocotitlan
Advances in Intelligent Systems and Computing | Year: 2013

In practical applications to credit risk evaluation, most prediction models often make inaccurate decisions because of the lack of sufficient default data. The challenging issue of highly skewed class distribution between defaulter and nondefaulters is here faced by means of an algorithmic solution based on cost-sensitive learning. The present study is conducted on the popular Multilayer Perceptron neural network using three misclassification cost functions, which are incorporated into the training process. The experimental results on real-life credit data sets show that the proposed cost functions to train such a neural network are quite effective to improve the prediction of examples belonging to the defaulter (minority) class. © Springer International Publishing Switzerland 2013.


Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Valdovinos R.M.,Valle de México University | Garcia V.,Jaume I University | Pacheco-Sanchez J.H.,Toluca Institute of Technology
Pattern Recognition Letters | Year: 2013

Class imbalance and class overlap are two of the major problems in data mining and machine learning. Several studies have shown that these data complexities may affect the performance or behavior of artificial neural networks. Strategies proposed to face with both challenges have been separately applied. In this paper, we introduce a hybrid method for handling both class imbalance and class overlap simultaneously in multi-class learning problems. Experimental results on five remote sensing data show that the combined approach is a promising method. © 2012 Elsevier B.V. All rights reserved.


Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Garcia V.,Autonomous University of Ciudad Juárez | Pacheco-Sanchez J.H.,Toluca Institute of Technology
Neural Processing Letters | Year: 2015

In this paper a new dynamic over-sampling method is proposed, it is a hybrid method that combines a well known over-sampling technique (SMOTE) with the sequential back-propagation algorithm. The method is based on the back-propagation mean square error (MSE) for automatically identifying the over-sampling rate, i.e., it allows only the use of necessary training samples for dealing with the class imbalance problem and avoiding to increase excessively the (neural networks) NN training time. The main aim of the proposed method is to obtain a trade-off between NN classification performance and NN training time on scenarios where the training data set represents a multi-class classification problem, it is high imbalanced and it might request a large NN training time. Experimental results on fifteen multi-class imbalanced data sets show that the proposed method is promising. © 2014, Springer Science+Business Media New York.


Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Toribio P.,Tecnologico de Estudios Superiores de Jocotitlan | Valdovinos R.M.,Valle de México University | Pacheco-Sanchez J.H.,Toluca Institute of Technology
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2012

In this paper we propose a modified back-propagation to deal with severe two-class imbalance problems. The method consists in automatically to find the over-sampling rate to train a neural network (NN), i.e., identify the appropriate number of minority samples to train the NN during the learning stage, so to reduce training time. The experimental results show that the performance proposed method is a very competitive when it is compared with conventional SMOTE, and its training time is lesser. © 2012 Springer-Verlag Berlin Heidelberg.


Fuente-Arriaga J.A.D.L.,Tecnologico de Estudios Superiores de Jocotitlan | Felipe-Riveron E.M.,National Polytechnic Institute of Mexico | Garduno-Calderon E.,Centro Oftalmologico Of Atlacomulco
Computers in Biology and Medicine | Year: 2014

This paper presents a methodology for glaucoma detection based on measuring displacements of blood vessels within the optic disc (vascular bundle) in human retinal images. The method consists of segmenting the region of the vascular bundle in an optic disc to set a reference point in the temporal side of the cup, determining the position of the centroids of the superior, inferior, and nasal vascular bundle segmented zones located within the segmented region, and calculating the displacement from normal position using the chessboard distance metric. The method was successful in 62 images out of 67, achieving 93.02% sensitivity, 91.66% specificity, and 91.34% global accuracy in pre-diagnosis. © 2014 Elsevier Ltd.


De La Fuente-Arriaga J.A.,Tecnologico de Estudios Superiores de Jocotitlan | Felipe-Riveron E.M.,National Polytechnic Institute of Mexico | Garduno-Calderon E.,Centro Oftalmologico Of Atlacomulco
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

This work presents a methodology for detecting human retina images suspect of glaucoma based on the measurement of displacement of the vascular bundle caused by the growth of the excavation or cup. The results achieved are due to the relative increase in size of the cup or excavation that causes a displacement of the blood vessel bundle to the superior, inferior and nasal optic disc areas. The method consists of the segmentation of the optic disc contour and the vascular bundle located within it, and calculation of its displacement from its normal position using the chessboard metric. The method was successful in 62 images of a total of 67, achieving an accuracy of 93.02% of sensitivity and 91.66% of specificity in the pre-diagnosis. © Springer-Verlag 2013.


Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Antonio J.A.,Tecnologico de Estudios Superiores de Jocotitlan | Valdovinos R.M.,Valle de México University | Pacheco-Sanchez J.H.,Toluca Institute of Technology
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

In this paper we study some of the most common global measures employed to measure the classifier performance on the multi-class imbalanced problems. The aim of this work consists of showing the relationship between global classifier performance (measure by global measures) and partial classifier performance, i.e., to determine if the results of global metrics match with the improved classifier performance over the minority classes. We have used five strategies to deal with the class imbalance problem over five real multi-class datasets on neural networks context. © 2013 Springer-Verlag Berlin Heidelberg.


Alejo R.,Tecnologico de Estudios Superiores de Jocotitlan | Monroy-De-Jesus J.,University of Central Mexico | Pacheco-Sanchez J.H.,Toluca Institute of Technology | Valdovinos R.M.,National Autonomous University of Mexico | And 2 more authors.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

In this work, we analyze the training samples for discovering what kind of samples are more appropriate to train the back-propagation algorithm. To do this, we propose a Gaussian function in order to identify three types of samples: Border, Safe and Average samples. Experiments on sixteen two-class imbalanced data sets where carried out, and a non-parametrical statistical test was applied. In addition, we employ the SMOTE as classification performance reference, i.e., to know whether the studied methods are competitive with respect to SMOTE performance. Experimental results show that the best samples to train the back-propagation are the average samples and the worst are the safe samples. © Springer International Publishing Switzerland 2015.


PubMed | National Polytechnic Institute of Mexico, Centro Oftalmologico Of Atlacomulco and Tecnologico de Estudios Superiores de Jocotitlan
Type: | Journal: Computers in biology and medicine | Year: 2014

This paper presents a methodology for glaucoma detection based on measuring displacements of blood vessels within the optic disc (vascular bundle) in human retinal images. The method consists of segmenting the region of the vascular bundle in an optic disc to set a reference point in the temporal side of the cup, determining the position of the centroids of the superior, inferior, and nasal vascular bundle segmented zones located within the segmented region, and calculating the displacement from normal position using the chessboard distance metric. The method was successful in 62 images out of 67, achieving 93.02% sensitivity, 91.66% specificity, and 91.34% global accuracy in pre-diagnosis.

Loading Tecnologico de Estudios Superiores de Jocotitlan collaborators
Loading Tecnologico de Estudios Superiores de Jocotitlan collaborators