International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
In case of syntax analyzers (parsers) planned on the basis of 
context-free, non stochastic (i.e. traditional) grammars, the two | 
classes of patterns could not be distinguished: that could mean | 
the string belong to both patterns. | 
However, what we already know about probable occurrence | 
plays a meaningful role.The parsers, in many cases, are made up | 
according to a likelihood criterion. However, parsers may also | 
be built according to a further criterion, i.e. the Bayes’s theorem 
Stochastic context free grammars are more powerful in order to | 
describe languages: additional rules allow to create nested, r | 
long-distance pairwise correlations between terminal symbols. | 
In the case of complex patterns, with more articulated relations | KEY 
between sub-patterns, regular grammars (even in stochastic | 
form) are not adequate: stochastic context-free grammars, also | 
more compact, should prove more efficient, altough more | ABS” 
expensive in terms of memory and computational time. | 
A variety of Cocke-Younger-Kasami (CYK) algorithm allows | ithe 
to find out an optimal parse tree (alignment problem) for | hiehe 
stochastic context free grammars in Chomsky Normal Form; the beer 
so called “inside algorithm” allows to find out probability of a | qualit 
given sequence (string) if the grammar is previously known. | Birds 
The inside algorithm can be compared with the forward | resol 
algorithm for HMMs, the same as the CYK algorithm can be | imas 
compared with the Viterbi algorithm used for HMMs. | image 
| univei 
References 
Aho, A., Ullman, J., 1972. The theory of Parsing Translation, | 
and Compiling. Prentice Hall, Englewood Cliffs. | 
Chomsky, N., 1957. Syntactic structures. Mouton, The Hague | In dis 
Fu, K., 1974. Syntactic Methods in Pattern Recognition. high : 
Academic Press, New York. image 
Markov, A., 1913. An example of statistical investigation in the say th 
text of "Eugene Onegin". In: Proceedings of the Academy of | charge 
Sciences of St. Petersburg. | precis 
Rabiner, L. 1989. A tutorial on hidden Markov models and purpo: 
selected applications in speech recognition. In: Proceedings of | and hi 
the IEEE, 77 (2). | high-r 
Rabiner, L., Juang, B. 1993. Fundamentals of Speech | dimen 
Recognition. Prentice Hall, NJ. | algorit 
Sester, M., 1992. Automatic model acquisition by Learning. In: image 
The International Archives of the Photogrammetry, Remote | 
Sensing and Spatial Information Sciences, Whashington USA, | Recen 
Vol. XXIX, Part B3. enhan« 
Resch B. Hidden Markov Models. | resolu 
http://www.igi.tugraz.at/lehre/CI | receive 
researc 
| 1984 ( 
| domai 
engine 
| algorit 
1985), 
| 1989; 
| approa 
1995)) 
| adaptiy 
| In 200 
| geome! 
from s 
| resolut 
This al 
where 
previot 
input i 
688