From 0da1f6de1b33bbff5cb99b1938bb07d050479f10 Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Wed, 14 Dec 2011 21:02:50 -0800 Subject: random incomplete metric stuff, including string subsequence kernel impl --- utils/kernel_string_subseq.h | 51 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 51 insertions(+) create mode 100644 utils/kernel_string_subseq.h (limited to 'utils/kernel_string_subseq.h') diff --git a/utils/kernel_string_subseq.h b/utils/kernel_string_subseq.h new file mode 100644 index 00000000..516e8b89 --- /dev/null +++ b/utils/kernel_string_subseq.h @@ -0,0 +1,51 @@ +#ifndef _KERNEL_STRING_SUBSEQ_H_ +#define _KERNEL_STRING_SUBSEQ_H_ + +#include +#include +#include + +template +float ssk(const T* s, const size_t s_size, const T* t, const size_t t_size, const float lambda) { + assert(N > 0); + boost::multi_array kp(boost::extents[N + 1][s_size + 1][t_size + 1]); + const float l2 = lambda * lambda; + for (unsigned j = 0; j < s_size; ++j) + for (unsigned k = 0; k < t_size; ++k) + kp[0][j][k] = 1.0f; + for (unsigned i = 0; i < N; ++i) { + for (unsigned j = 0; j < s_size; ++j) { + float kpp = 0.0f; + for (unsigned k = 0; k < t_size; ++k) { + kpp = lambda * (kpp + lambda * (s[j]==t[k]) * kp[i][j][k]); + kp[i + 1][j + 1][k + 1] = lambda * kp[i + 1][j][k + 1] + kpp; + } + } + } + float kn = 0.0f; + for (int i = 0; i < N; ++i) + for (int j = 0; j < s_size; ++j) + for (int k = 0; k < t_size; ++k) + kn += l2 * (s[j] == t[k]) * kp[i][j][k]; + return kn; +} + +template +float ssk(const std::vector& s, const std::vector& t, const float lambda) { + float kst = ssk(&s[0], s.size(), &t[0], t.size(), lambda); + if (!kst) return 0.0f; + float kss = ssk(&s[0], s.size(), &s[0], s.size(), lambda); + float ktt = ssk(&t[0], t.size(), &t[0], t.size(), lambda); + return kst / std::sqrt(kss * ktt); +} + +template +float ssk(const std::string& s, const std::string& t, const float lambda) { + float kst = ssk(&s[0], s.size(), &t[0], t.size(), lambda); + if (!kst) return 0.0f; + float kss = ssk(&s[0], s.size(), &s[0], s.size(), lambda); + float ktt = ssk(&t[0], t.size(), &t[0], t.size(), lambda); + return kst / std::sqrt(kss * ktt); +} + +#endif -- cgit v1.2.3