SPRAAK
|
Error rate for cwr-systems. More...
Data Structures | |
struct | SprCwrCESeq |
struct | SprCwrCECost |
struct | SprCwrErr |
struct | SprCwrCEDistMat |
struct | SprCwrCEChrDiff |
Namespaces | |
xref_spr_dist_mat | |
Distance matrices | |
Typedefs | |
typedef double(* | _FuncPtr1_CWR_ERR_ )(const void *info, int ndx_ref, int ndx_cmp) |
Enumerations | |
enum | { SPR_CWR_INSERTION, SPR_CWR_DELETION, SPR_CWR_SUBSTITUTION, SPR_CWR_CORRECT, SPR_CWR_DIFF_TERMINATED } |
Functions | |
void | spr_cwr_err_free (SprCwrErr *info) |
int | spr_cwr_err_seq_expand (SprCwrCESeq *seq, int add_el) |
SprCwrCEDistMat * | spr_cwr_err_dist_free_data (SprCwrCEDistMat *dmat) |
SprCwrCEDistMat * | spr_cwr_err_dist_mat_read (const char *fname, const SprStrHashTbl *dic) |
SprCwrCEChrDiff * | spr_cwr_err_chr_diff_free (SprCwrCEChrDiff *info) |
SprCwrCEChrDiff * | spr_cwr_err_chr_diff_init (const char *options, int opt_len, const SprStrHashTbl *ndx2str) |
double | spr_cwr_err_chr_diff_func (const SprCwrCEChrDiff *info, int ndx_ref, int ndx_cmp) |
int | spr_cwr_err (SprCwrErr *info) |
Variables | |
const SprCwrErr | spr_cwr_empty_cwr_err |
const char *const | spr_cwr_err_str [] |
double(* | spr_cwr_err_nist_werr )(const void *info, int ndx_ref, int ndx_cmp) |
double(* | spr_cwr_err_dist_mat )(const void *info, int ndx_ref, int ndx_cmp) |
Alignment cost function based on (phonetic) features. More... | |
const SprCmdOptDesc | spr_cwr_od_err_chr_diff [] |
Error rate for cwr-systems.
Align a reference and test sequence given some distance metric and calculate the substitution, insertion and deletion rate. The alignment optimizes a cost (a float, primary goal) and a sub-cost (an integer, secondary goal). The sub-cost function (secondary goal) is fixed and gives preference to deletetions over insertions over substitutions. The cost function (primary goal) can be user implemented.
Currently the following cost functions (primary goals) are implemented:
a substitution (+ insertion & deletion) matrix or a hierarchical (feature-based) distance matrix
for more info see Distance matrices