Adriana Vlad, Adrian Mitrea * Estimating Conditional Probabilities and Digram Statistical Structure in Printed Romanian



Table 2.1. Letter probabilities estimation in text without blanks
Letter
i 		Probability
estimate 		Signal to
noise ratio 		Relative
error 		Confidence limits *) Relative
frequency
upper lower
E 0.1214 93.6 0.0209 0.1240 0.1189 0.1221
I 0.1058 86.6 0.0226 0.1082 0.1034 0.1056
A 0.1005 84.1 0.0233 0.1029 0.0982 0.0980
R 0.0760 72.2 0.0272 0.0781 0.0739 0.0744
N 0.0651 66.4 0.0295 0.0671 0.0632 0.0645
T 0.0621 64.8 0.0303 0.0640 0.0602 0.0629
U 0.0586 62.8 0.0312 0.0605 0.0568 0.0592
C 0.0508 58.2 0.0337 0.0525 0.0491 0.0523
L 0.0481 56.6 0.0346 0.0498 0.0464 0.0481
O 0.0422 52.8 0.0371 0.0438 0.0407 0.0424
S 0.0411 52.1 0.0376 0.0427 0.0396 0.0419
Ã0.0339 47.1 0.0416 0.0353 0.0325 0.0339
D 0.0326 46.2 0.0424 0.0340 0.0312 0.0333
P 0.0318 45.6 0.0430 0.0332 0.0304 0.0316
M 0.0287 43.2 0.0453 0.0300 0.0274 0.0285
ª0.0133 29.2 0.0671 0.0142 0.0124 0.0128
F 0.0119 27.6 0.0710 0.0128 0.0111 0.0120
Î0.0119 27.6 0.0710 0.0128 0.0111 0.0121
V 0.0115 27.2 0.0721 0.0124 0.0107 0.0113
Þ0.0110 26.6 0.0738 0.0119 0.0102 0.0115
G 0.0099 25.1 0.0779 0.0107 0.0091 0.0094
B 0.0084 23.2 0.0846 0.0091 0.0077 0.0091
Z 0.0076 22.1 0.0888 0.0083 0.0070 0.0074
Â0.0067 20.7 0.0949 0.0074 0.0061 0.0065
H 0.0038 (15.5) - - - 0.0040
J 0.0025 (12.7) - - - 0.0022
X 0.0019 (11.0) - - - 0.0021
K 0.0004 - - - - 0.0003
Y 0.0003 - - - - 0.0003
W 0.0002 - - - - 0.0003
Q 0.0000 - - - - 0.0000

*) Calculated on the whole X text (on correlated data).

49

Previous Index Next