Selected article for: "amino acid and auc increase"

Author: Zheng Zhang; Sifan Ye; Aiping Wu; Taijiao Jiang; Yousong Peng
Title: Prediction of receptorome for human-infecting virome
  • Document date: 2020_2_28
  • ID: 9ruhvpbv_32
    Snippet: The RF model based on combination of 20 amino-acid frequencies achieved an AUC 0.68. When using the frequencies of k-mers with two amino acids as features, only the top N (N=1 to 400) most important features were used in the modeling to reduce over-fitting. Figure S1 showed that the AUC of the RF models increased rapidly as increase of N from 1 to 10. Then, it increased slowly or even decreased when more ten features were used. Therefore, only to.....
    Document: The RF model based on combination of 20 amino-acid frequencies achieved an AUC 0.68. When using the frequencies of k-mers with two amino acids as features, only the top N (N=1 to 400) most important features were used in the modeling to reduce over-fitting. Figure S1 showed that the AUC of the RF models increased rapidly as increase of N from 1 to 10. Then, it increased slowly or even decreased when more ten features were used. Therefore, only top ten features were used in the modeling to balance the model performance and complexity. The RF model based on top ten features achieved an AUC of 0.72.

    Search related documents:
    Co phrase search for related documents
    • amino acid and model performance: 1, 2, 3, 4
    • amino acid and rapidly increase: 1, 2, 3
    • amino acid and RF model: 1
    • auc achieve and model performance: 1, 2
    • complexity model performance and model performance: 1, 2, 3, 4
    • feature base and important feature: 1
    • feature base and model performance: 1
    • important feature and model performance: 1, 2, 3
    • model performance and rapidly increase: 1
    • model performance and RF model: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12