java - Which Algorithm to choose in pattern matching time serie? -


i 've got multiple time series represented multiple list of integers (in main memory, not in database). need perform fast search, among series, find specific pattern.

example, detect pattern [ (0,1), (3,2), (4,1) ] (x,y) x = time, y = # series.

i've googled pattern detection, pattern matching, seems there thousand of algorithms, , don't quite see relation problem of time. first idea coming head, ,for every point,checking distance next point on specific train according pattern, , on.

i need directions on start, because i'm confused among theses researcher publications !

thank much

nico

specification : make multiple pass, jitter associated. example, let's take pattern define above. first pass, i'll need match exact pattern. second pass, need match pattern jitter of 1 -> [ (0,1), ( 3 +- 1, 2) (4 +- 1, 1) ], second pass -> [ (0,1), (3 +- 2, 2) (4 +- 1, 1) ] etc. jitters go five. first "event" (time,# serie) "constant" @ time.

i need add fact maximum time in time series around 100 000, may converted "bit string" of 0 , 1s.

edit : value @ specific time doesn't matter, don't have it. series of "event" @ different time & position (which serie event on). events equal. (papers detect patterns has double value etc etc, here s useless)

would possible use state machine represent problem? (0,1) can state , there, can expect next state, based on "turn" variable.


Comments

Popular posts from this blog

jquery - How can I dynamically add a browser tab? -

node.js - Getting the socket id,user id pair of a logged in user(s) -

keyboard - C++ GetAsyncKeyState alternative -