Neural Network data

[Guest adamskitwo]

Hi, I have been designing a neural network for some time, and have generally used data from Joes database. However, I have decided that it would be nice if I could use injuries/suspensions, and if the team has played in the last 4 days as inputs. What this would require is for some games (they dont need to be consecutive or in a certain season even), I would like to know which players missed them through injury or suspension, and if the team has played in the last 4 days. What I would need is perhaps some explicit data saying: man u 2-1 liverpool, susp: keane, butt. injured: solskaer, fortune, or some sort of data that implies that. The third attribute would be easier to get I assume just by getting the entire seasons fixture list for a few clubs in all competitions. The important factor is that I do not need all three attributes to be satisfied to input to the neural network for each game. Only one attribute can be satisfied for the neural network to have 'learned' something. Does anyone know where I can get this sort of data from? Thanks, Adam PS. One of my difficulties after will be deciding who is a key player, and who is not, and perhaps a rough guide is if the player injured is in the squad numbers 1-16. If so, they are key players. Does anyone know where I can acquire peoples squad numbers for a given year to link the importance of injuries and suspensions?

[macbeth9]

Shirt Numbers Shirt Numbers won't work. My team, Newcastle, try to rearrange short numbers each summer but players become attached to their shirt :D We have Lauren Robert at 32, and he is key, Shola is becoming more key at 23, Woodgate is 27, Bowyer 29. Robbie Elliot is 3 but won't be good enough to get a game this season. I doubt Newcatsle are much different from other clubs in this respect