Due to recent updates, all users are required to create an Altair One account to login to the RapidMiner community. Click the Register button to create your account using the same email that you have previously used to login to the RapidMiner community. This will ensure that any previously created content will be synced to your Altair One account. Once you login, you will be asked to provide a username that identifies you to other Community users. Email us at Community with questions.
Sabermetrics / Baseball action forecasting
Contributor II
Dear everyone!
In my freetime, I'm playing for a baseball team.
Usually, there is one person showing the batter how to play the ball using a series of signs.
An example: Touches his nose, swipes over his shirt and touches his legs.
One of these signs tells the batter which action to take (hit the ball, do a bunt or whatever) and the rest of the signs is just to make it harder for the other team to know what will happen.
The "real" sign may be the first, second or any other sign in the row of signs he shows.
Sometimes, he also only shows any signs that haven't been communicated before which usually means, that the batter should play how he wants to.
To my question:
Is there a way to create an algorithm that analyses the signs that were shown, compares them to the actions that were taken by the players before and then shows a probability which action will be taken by the batter?
Of course, the user of the algorithm would have to input a series of data before like:
Signs: touch hand, touch nose, clap
Action: Bunt
Many thanks in advance!
In my freetime, I'm playing for a baseball team.
Usually, there is one person showing the batter how to play the ball using a series of signs.
An example: Touches his nose, swipes over his shirt and touches his legs.
One of these signs tells the batter which action to take (hit the ball, do a bunt or whatever) and the rest of the signs is just to make it harder for the other team to know what will happen.
The "real" sign may be the first, second or any other sign in the row of signs he shows.
Sometimes, he also only shows any signs that haven't been communicated before which usually means, that the batter should play how he wants to.
To my question:
Is there a way to create an algorithm that analyses the signs that were shown, compares them to the actions that were taken by the players before and then shows a probability which action will be taken by the batter?
Of course, the user of the algorithm would have to input a series of data before like:
Signs: touch hand, touch nose, clap
Action: Bunt
Many thanks in advance!
0
Answers
But if the coach is canny then he might be using a combination of signals.
Let's that if he has the actions:
- action 1 (touch nose)
- action 2 (piqué turn)
- action 3 (cough 3 times)
- action 4 (plie)
- action 5 (pirouette)
Actions 1,2,3,4 & 5 might mean nothing separately, but 2+4+5 together might mean that the batter should 'go long' (is that a baseball term? I've never played).
Anyway, in this case Naive Bayes might still be helpful, but association rules might be better as it would help you to visualise which combination of actions when used together signal a certain action.
Well yes, combinations would be possible, for example that he crosses his legs which might mean that all shown signs do not apply.
I guess that it would be best to let one person input the seen signs and actions as people might describe things differently.
But as the whole process of showing the sign and hitting might be relatively short, is it possible to wright down what you see in one sentence like this:?
"touches cap, claps, shakes his head and crosses his legs"
instead of putting each of the signs in separate fields?
Another processor would have to extract the relevant keywords then of course.