Learning on the basis of a polynomial pattern synthesis algorithm
The proposed system may be used as an intelligent component of any relational DBMS. It should be quite useful in restructuring rapidly changing data bases and in the entry process of large bodies of information, some of it possibly not exactly specified before.
It should be possible to speed up the work of the system by using an enconding procedure for the inputs (cf./5/). In this case changes are to be made in the PSAn, since it would operate with a potentially infinite alphabet.
The system is currently been implemented on a SM-4 minicomputer (PDP-11 compatible) in LISP with the prospect to carry it over to a more powerful minicomputer, thus increasing speed and attractivity.
KeywordsInput Sample Mathematical Linguistics Pattern Language Finite Alphabet Record Structure
Unable to display preview. Download preview PDF.
- /1/.Angluin, D.: "Finding Patterns Common to a Set of Strings", J. of Computer and System Sciences 21(1980) 1Google Scholar
- /2/.Jantke, K.P.: "Polynomial Time Inference of General Pattern Languages", Proceedings of STACS '84, Paris, LINCS 166Google Scholar
- /3/.Martin, J.: "Computer Data-Base Organization", Prentice-Hall, 1977Google Scholar
- /4/.Shapiro, E.: "Inductive Inference of Theories from Facts", Res. Report 192, Dept. Comp. Sci., Yale University, 1981Google Scholar
- /5/.Shinohara, T.: Polynomial Time Inference of Extended Regular Pattern Languages", RIMS Symposia on Software Science and Engineering, Kycto, 1982, LINCS 147Google Scholar
- /6/.Winston, P.: "Artificial Intelligence", Addison-Wesley, 1977Google Scholar