dp - dynamic programming recogniser


dp (-r) (-a|-b|-c|-d|-e) (-i item) <token_files> file


dp is an isolated word recognizer implemented using a dynamic programming algorithm. It accepts a list of reference files ("token files") and a single unknown file and calculates the which token file is "nearest" to the unknown. The program can report the identoty of the nearest token, the nearest filename, or a table of matching scores according to selected options.

The dp program operates using a "Euclidean" metric between vectors of a "Coefficient" (CO) data set in the files (identified by the "-i item" switch). The DP algorithm usually operates without any slope constraints, but these may be incorporated with the "-r" switch.

Options and their meanings are:

-I Identify program name and version number.

-r Select a 1 to 2 ratio constraint on the length of the templates.

-a|-b|-c|-d|-e Select "-a" for nearest filename, "-b" for table of scores, "-c" for table of scores with filenames, "-d" as "-b" but sorted by score, "-e" as "-c" but sorted by score. Default is to report the "head.token" field from the best match and test files.

-i item Select input item number.


CO.xx Any Coefficient item.


1.0 - b.l.tan
2.0s - Mark Huckvale
Fri Jul 09 14:54:42 2004