Academia.eduAcademia.edu

Figure 2 – uploaded by Nelson F R A N C I S C O F A V I L L A Ebecken

See full PDFdownloadDownload figure
Fig. 2. Voronoi diagram: a representation of the classification process

Figure 2 Voronoi diagram: a representation of the classification process

Related Figures (14)

To proceed the system’s learning, the database is divided into three subsets training representing the antibodies, testing and validating as antigens. Th initial proportion of samples adopted for each group, respectively, was 60%, 209 and 20% and then, another experiment with the proportions 70%, 20% an 10%.These proportions were adopted empirically. Both antigens and antibodie are represented by an array containing the attributes. The antigens, or test data are classified according to the closest antibody. It means that in this case, instea of a shape space with variant ¢, the model uses the Voronoi diagram [28], tha can be seen in Figure 2. In the figure, dots are the centroids of the ARBs an represent the antibodies that cover a certain group of nearest antigens.
To proceed the system’s learning, the database is divided into three subsets training representing the antibodies, testing and validating as antigens. Th initial proportion of samples adopted for each group, respectively, was 60%, 209 and 20% and then, another experiment with the proportions 70%, 20% an 10%.These proportions were adopted empirically. Both antigens and antibodie are represented by an array containing the attributes. The antigens, or test data are classified according to the closest antibody. It means that in this case, instea of a shape space with variant ¢, the model uses the Voronoi diagram [28], tha can be seen in Figure 2. In the figure, dots are the centroids of the ARBs an represent the antibodies that cover a certain group of nearest antigens.
Table 3. Results from other classifiers to cancer database Fig. 3. The total amount of antibodies (training set) for the Cancer database
Table 3. Results from other classifiers to cancer database Fig. 3. The total amount of antibodies (training set) for the Cancer database
Table 4. Results from the predictor without suppression mechanism for Pima Indian: Diabetes Database with 460 antibodies
Table 4. Results from the predictor without suppression mechanism for Pima Indian: Diabetes Database with 460 antibodies
Fig. 4. Antibodies from Cancer database after suppression and 60-40% split
Fig. 4. Antibodies from Cancer database after suppression and 60-40% split
Table 6. Results from other classifiers to diabetes database Fig. 5. The total amount of antibodies (training set) for the Diabetes database
Table 6. Results from other classifiers to diabetes database Fig. 5. The total amount of antibodies (training set) for the Diabetes database
Table 7. Results for the Iris Database with 90 antibodies without suppression Table 8. Results for the Iris Database with 10 antibodies with suppression
Table 7. Results for the Iris Database with 90 antibodies without suppression Table 8. Results for the Iris Database with 10 antibodies with suppression
Table 9. Results from other classifiers to iris database
Table 9. Results from other classifiers to iris database
Fig. 6. Antibodies from Diabetes database after suppression and 60-40% split
Fig. 6. Antibodies from Diabetes database after suppression and 60-40% split
Fig. 7. The total amount of antibodies (training set) for the Iris database
Fig. 7. The total amount of antibodies (training set) for the Iris database
Fig. 8. Antibodies from Iris database after suppression and 60-40% split
Fig. 8. Antibodies from Iris database after suppression and 60-40% split
Related topics:
Computer ScienceArtificial IntelligenceMachine LearningData classificationStatistical Classification
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved