I have the results of competative sporting matches in nominal sequential form for the past games (historic encounters) as the following example shows:
|Team 1||Team 2||H||H||D||D||H||D|
|Team 1||Team 2||H||H||H||D||D||H|
M1 - M5 denotes the historical encounters and FTR is the label to predict. The observations are lagged as if using time series. I was wondering the class of pattern recognition algorithums that may be sutible to handle this type of data.
Any information would be of a great help, as I am looking at different algorithums such as Conditional Random Feilds but am unsure that these are the best to tackle the problem.