Selected article for: "feature selection and reduction algorithm"
Author: Saban Ozturk; Umut Ozkaya; Mucahid Barstugan
Title: Classification of Coronavirus Images using Shrunken Features
  • Document date: 2020_4_6
    • ID: 2l1zw19o_52
    Snippet: Since the sAE method is supervised, it is understood that feature narrowing operation with the insufficient number of samples has failed. For this reason, the PCA algorithm, which performs feature reduction in unsupervised style, is Table 3 . Comparing these results with Table 1 and Table 2 , it is understood that the performance of the PCA algorithm is higher. Considering the proposed framework and dataset status, it is seen that the PCA algorit.....
    Document: Since the sAE method is supervised, it is understood that feature narrowing operation with the insufficient number of samples has failed. For this reason, the PCA algorithm, which performs feature reduction in unsupervised style, is Table 3 . Comparing these results with Table 1 and Table 2 , it is understood that the performance of the PCA algorithm is higher. Considering the proposed framework and dataset status, it is seen that the PCA algorithm is more suitable for feature selection. It is thought that its effect will be high, especially in the investigation of viruses such as Covid that started suddenly and in studies with a small number of data. In Table 3 , it is seen that in experiments with 260 samples, it produces more successful results than other experiments, regardless of the number of features. After over-sampling, the classification performance increases even more due to the class balance. Figure 9 shows the AUC curve of the features obtained by the PCA method.

    Search related documents:
    Co phrase search for related documents
    • AUC curve and classification performance: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20
    • AUC curve and Covid virus: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23
    • AUC curve and dataset status: 1
    • AUC curve and feature number: 1
    • AUC curve and feature reduction: 1
    • AUC curve and feature selection: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24
    • AUC curve and pca method: 1
    • AUC curve and propose framework: 1, 2, 3, 4
    • AUC curve and small number: 1, 2, 3
    • classification performance and Covid virus: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
    • classification performance and feature number: 1
    • classification performance and feature reduction: 1, 2
    • classification performance and feature selection: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24
    • classification performance and PCA algorithm: 1
    • classification performance and propose framework: 1, 2, 3, 4
    • classification performance and sae method: 1, 2
    • classification performance and small number: 1, 2, 3, 4, 5
    • classification performance increase and feature selection: 1
    • Covid virus and dataset status: 1, 2