Skip to main content
. 2012 Feb 24;135(3):794–806. doi: 10.1093/brain/aws001

Table 3.

Classification tables for both the optimum and conservative penalized multinomial logistic regression models

Predicted patient groups MAPT (n = 25) GRN (n = 12) C9ORF72 (n = 19) Sporadic FTD (n = 20)
Optimal model with 26 variables (overall classification 93%)
    MAPT 24 0 0 1
    GRN 0 12 0 0
    C9ORF72 0 0 17 1
    Sporadic FTD 1 0 2 18
Conservative model with 14 variables (overall classification 74%)
    MAPT 23 3 0 3
    GRN 0 7 1 0
    C9ORF72 1 1 13 4
    Sporadic FTD 1 1 5 13