Paraconsistent Artificial Neural Network Applied in Breast Cancer Diagnosis Support - IFIP-AICT-459 Access content directly
Conference Papers Year : 2015

Paraconsistent Artificial Neural Network Applied in Breast Cancer Diagnosis Support

Fábio Do Amaral
  • Function : Author
  • PersonId : 996271
Jair Minoro Abe
  • Function : Author
  • PersonId : 996220
Alexandre Sandor Cadim
  • Function : Author
  • PersonId : 996272
Caique Zaneti Kirilo
  • Function : Author
  • PersonId : 996273
Carlos Arruda Baltazar
  • Function : Author
  • PersonId : 996274
Fábio Luís Pereira
  • Function : Author
  • PersonId : 996275
Hélio De Araújo
  • Function : Author
  • PersonId : 996276
Henry Costa Ungaro
  • Function : Author
  • PersonId : 996277
Lauro De Castro Tomiatti
  • Function : Author
  • PersonId : 996278
Luiz Machi Lozano
  • Function : Author
  • PersonId : 996279
Uanderson Celestino
  • Function : Author
  • PersonId : 996282
Rafael Espirito Santo
  • Function : Author
  • PersonId : 996283

Abstract

In this work, a Paraconsistent Classifier for the diagnosis of breast cancer based on the attributes of mamographic images was developed. The system uses a neural network with decision making from the final results of processing each set of attributes was created. In order to mitigate the effects of false positives diagnoses and true positives. In order to analyze the performance of the Paraconsistent Classifier, the results will be compared with the results of the following classifiers: Multi-Layer Perceptron (MLP), a dual stage classifier (ART2LDA) based on Adaptive Resonance Theory (ART) and a classifier implemented with nonlinear optimization techniques and combinatorics, associated with the classification capabilities of Radial basis Functions - (RBF-Simulated Annealing).To perform the simulations, two different databases were used. The first one, to classify calcifications, is composed of 143 samples divided into 64 benign cases and 79 malignant cases represented by form. The performances of the classifiers in discriminating benign and malignant cases are compared in terms of area under the Receiver Operating Characteristic Curve (Az). The higher the value of Az, the better the performance of the classifier.The experiments with calcifications show: Paraconsistent Classifier (Az = 0.986), MLP classifier (Az = 0.70), ART2LDA Classifier (Az = 0.696) and RBF Classifier - Simulated Annealing (Az = 0.94). For experiments with mammographic masses and tumors show: Set 1, Paraconsistent Classifier (Az = 0.939), MLP classifier (Az = 0.994), ART2LDA Classifier (Az = 0.901) and RBF Classifier - Simulated Annealing (Az = 0.912). Set 2, Paraconsistent Classifier (Az = 0.935), MLP classifier (Az = 0.994), ART2LDA Classifier (Az = 0.890) and RBF Classifier - Simulated Annealing (Az = 0.924). Set 3, Paraconsistent Classifier (Az = 0.875), MLP classifier (Az = 0.970), ART2LDA Classifier (Az = 0.850) and RBF Classifier - Simulated Annealing (Az = 0.996). Set 4, Paraconsistent Classifier (Az = 0.500), MLP classifier (Az = 0.887), ART2LDA Classifier (Az = 0.767) and RBF Classifier - Simulated Annealing (Az = 0.907). Set 5, Paraconsistent Classifier (Az = 0.929), MLP classifier (Az = 0.987), ART2LDA Classifier (Az = 0.884) and RBF Classifier - Simulated Annealing (Az = 0.998). Set 6, Paraconsistent Classifier (Az = 0.939), MLP classifier (Az = 0.982), ART2LDA Classifier (Az = 0.885) and RBF Classifier - Simulated Annealing (Az = 0.999). In the case of the Paraconsistent Classifier, the eighth experiment was composed of the total image attributes relating to mammographic masses and tumors (Az = 0.939). In the particular case of the RBF-Sort Simulated Annealing experiment with all the image attributes, it was proved its unfeasibility, due to the complexity of their algorithm, where the processing time tends to infinity for a larger number of elements. For the experiments, the Paraconsistent classifier used 20 % of the samples for the neural network training, against the total number of samples available minus one for the other classifiers.
Fichier principal
Vignette du fichier
346972_1_En_57_Chapter.pdf (618.74 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01417535 , version 1 (15-12-2016)

Licence

Attribution

Identifiers

Cite

Fábio Do Amaral, Jair Minoro Abe, Alexandre Sandor Cadim, Caique Zaneti Kirilo, Carlos Arruda Baltazar, et al.. Paraconsistent Artificial Neural Network Applied in Breast Cancer Diagnosis Support. IFIP International Conference on Advances in Production Management Systems (APMS), Sep 2015, Tokyo, Japan. pp.464-472, ⟨10.1007/978-3-319-22756-6_57⟩. ⟨hal-01417535⟩
149 View
84 Download

Altmetric

Share

Gmail Facebook X LinkedIn More