Extension :
Higher level of spatio-temporal reasoning can be used to determine the
team to which the players belong, by using only the spatial position data.
More modules can be added to detect more number of events thus providing
a more holistic view of the match. A much higher level of learning can
be done to "suggest" some sequence of events which are likely to be very
successful rather than only follow the team's strategy. Thus a pass can
be suggested which is not a good pass in terms of difficulty but which
is very good in terms of striking a goal.
For the source code of this project CLICK HERE