N | number of states |
T | number of observations |
S = {s1, ..., si, ... , sN} | states, where i = 1...N |
O = {o1, ..., ok, ... , oT} | observations, where k = 1...T |
A = {a11, ..., aij, ... , aNN} | state transition probability distribution, where aij constructed as follows: i = 1...N, j = 1...N, i - represents start state, j - represents end state. |
B = {b11, ... , bik, ... , bNT} | state emission probability distribution, where bik constructed as follows: i = 1...N, k = 1...T, i - represents state, k - represents observation. |
Π = {π1, ..., πi, ..., πN} | initial state probability distribution, where i = 1...N |
Ω = {ω1, ..., ωi, ..., ωN} | terminal state probability distribution, where i = 1...N |
λ = (A, B, π) | HMM model |
Below you can see visualised notation:
Diagram 1. HMM model example |
No comments:
Post a Comment