SYSTEM AND METHOD FOR CONTENT AND STYLE PREDICTIONS IN SPORTS
A system and method for generating a play prediction for a team is disclosed herein. A computing system retrieves trajectory data for a plurality of plays from a data store. The computing system generates a predictive model using a variational autoencoder and a neural network by generating one or more input data sets, learning, by the variational autoencoder, to generate a plurality of variants for each play of the plurality of plays, and learning, by the neural network, a team style corresponding to each play of the plurality of plays. The computing system receives trajectory data corresponding to a target play. The predictive model generates a likelihood of a target team executing the target play by determining a number of target variants that correspond to a target team identity of the target team.
1 . A computer-implemented method of generating a prediction model, the computer-implemented method comprising:
receiving, by one or more processors, tracking data corresponding to a match from a data store, wherein the tracking data includes one or more coordinates and one or more time stamps associated with at least one object or at least one actor;
generating, by the one or more processors, one or more input data sets based on the tracking data, wherein generating the one or more input data sets includes enriching the tracking data; and
learning, by the one or more processors, via a prediction model, a plurality of predicted variants for each play corresponding to the one or more input data sets.
2 . The computer-implemented method of claim 1 , the computer-implemented method further comprising:
generating, by the one or more processors, a predicted identity corresponding to each of the one or more input data sets.
3 . The computer-implemented method of claim 1 , the computer-implemented method further comprising:
reducing, by the one or more processors, a loss of the plurality of predicted variants and an input sample of one or more variants.
4 . The computer-implemented method of claim 1 , wherein the learning the plurality of predicted variants for each play corresponding to the one or more input data sets comprises:
utilizing, by the one or more processors, an optimizer to train the prediction model.
5 . The computer-implemented method of claim 1 , wherein enriching the tracking data includes:
enriching, by the one or more processors, the tracking data with additional data corresponding to a possession, a playing style, or a team identity.
6 . The computer-implemented method of claim 1 , wherein generating the one or more input data sets based on the tracking data comprises:
aligning, by the one or more processors, the at least one actor to a global template to reduce permutation noise.
7 . The computer-implemented method of claim 1 , wherein at least one of the one or more input data sets correspond to a possession of the match.
8 . A non-transitory computer-readable medium comprising one or more sequences of instructions, which, when executed by one or more processors, causes a computing system to perform operations comprising:
receiving, by the computing system, tracking data corresponding to a match from a data store, wherein the tracking data includes one or more coordinates and one or more time stamps associated with at least one object or at least one actor;
generating, by the computing system, one or more input data sets based on the tracking data, wherein generating the one or more input data sets includes enriching the tracking data; and
learning, by the computing system, via a prediction model, a plurality of predicted variants for each play corresponding to the one or more input data sets.
9 . The non-transitory computer-readable medium of claim 8 , the operations further comprising:
generating, by the computing system, a predicted identity corresponding to each of the one or more input data sets.
10 . The non-transitory computer-readable medium of claim 8 , the operations further comprising:
reducing, by the computing system, a loss of the plurality of predicted variants and an input sample of one or more variants.
11 . The non-transitory computer-readable medium of claim 8 , wherein the learning the plurality of predicted variants for each play corresponding to the one or more input data sets comprises:
utilizing, by the computing system, an optimizer to train the prediction model.
12 . The non-transitory computer-readable medium of claim 8 , wherein enriching the tracking data includes:
enriching, by the computing system, the tracking data with additional data corresponding to a possession, a playing style, or a team identity.
13 . The non-transitory computer-readable medium of claim 8 , wherein generating the one or more input data sets based on the tracking data comprises:
aligning, by the computing system, the at least one actor to a global template to reduce permutation noise.
14 . The non-transitory computer-readable medium of claim 8 , wherein at least one of the one or more input data sets correspond to a possession of the match.
15 . A computer system, comprising:
a processor; and
a memory having programming instructions stored thereon, which, when executed by the processor, causes the system to perform operations comprising:
receiving tracking data corresponding to a match from a data store, wherein the tracking data includes one or more coordinates and one or more time stamps associated with at least one object or at least one actor;
generating one or more input data sets based on the tracking data, wherein generating the one or more input data sets includes enriching the tracking data; and
learning, via a prediction model, a plurality of predicted variants for each play corresponding to the one or more input data sets.
16 . The computer system of claim 15 , the operations further comprising:
generating a predicted identity corresponding to each of the one or more input data sets.
17 . The computer system of claim 15 , the operations further comprising:
reducing a loss of the plurality of predicted variants and an input sample of one or more variants.
18 . The computer system of claim 15 , wherein the learning the plurality of predicted variants for each play corresponding to the one or more input data sets comprises:
utilizing an optimizer to train the prediction model.
19 . The computer system of claim 15 , wherein enriching the tracking data includes:
enriching the tracking data with additional data corresponding to a possession, a playing style, or a team identity.
20 . The computer system of claim 15 , wherein generating the one or more input data sets based on the tracking data comprises:
aligning the at least one actor to a global template to reduce permutation noise.