Base | L | % of correct prediction of text fragment sequence | Advantage of NFIM (%) (41) | ||||||||
By unconditional probability | By 1st fragment | By 2nd fragment | By mean probability | Method 1 | Method 2 | Method 3 | Method 4 | Non-force interaction method | |||
“Russian” | 1 | 40.32 | 25.71 | 24.19 | 19.62 | 19.65 | 16.38 | 17.44 | 19.52 | 11.43 | 43.33 |
2 | 22.72 | 14.63 | 14.59 | 11.80 | 11.81 | 11.27 | 10.90 | 11.17 | 9.05 | 20.38 | |
“German” | 1 | 26.12 | 18.11 | 13.11 | 12.47 | 12.99 | 11.61 | 10.90 | 10.94 | 7.64 | 42.61 |
2 | 21.07 | 14.17 | 12.11 | 11.18 | 11.20 | 10.15 | 10.12 | 9.84 | 7.61 | 29.22 | |
“Ukrainian” | 1 | 44.41 | 27.53 | 25.18 | 23.81 | 23.91 | 18.28 | 19.63 | 17.30 | 10.14 | 70.61 |
2 | 3.37 | 2.07 | 2.15 | 1.84 | 1.84 | 1.74 | 1.73 | 1.62 | 1.50 | 7.78 | |
“English” | 1 | 33.74 | 19.32 | 18.52 | 14.78 | 14.87 | 13.40 | 12.61 | 13.28 | 8.91 | 41.42 |
2 | 65.33 | 42.61 | 41.05 | 32.79 | 32.79 | 31.63 | 29.99 | 31.10 | 24.75 | 21.15 |