Base

L

% of correct prediction of text fragment sequence

Advantage of NFIM (%) (41)

By unconditional probability

By 1^st fragment

By 2^nd 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