4
$\begingroup$

I confront to the next scenario:

Let's say I have stored data about football matches between different teams: lineups, scorers, yellow cards, and many other events.

I need to generate everyday some questions about the matches that will be played on that day. So, if I give an input of two teams, I would like a related question to be generated, based on previous data of matches between those two teams.

For example, if my input are "TeamA" and "TeamB", I would expect a question of the type:

  • "Will there be less than 2 goals scored in the match?""

  • "Will PlayerX score a goal during the match?"

Of course I expect these questions to make sense based on previous data from matches between the two given teams.

So, my questions are:

  • Would be a good solution to use AI to generate these questions? It would make sense?
  • What would be the best approach?
$\endgroup$
3
$\begingroup$

One simple approach to consider would be storing each statement as a template made in advance.

  • Will there be less/more than x goals scored in the match?

  • Will player score a goal during the match?

  • ...

The system will pick a random statement and will fill the variable fields with some statistically generated data between teamA and teamB; here you have your question.


Example: Will there be less/more than x goals scored in the match?

  • less/more fragment may be random
  • x may be the mean of the goals scored considering all the matches between teamA and teamB

Example: Will player score a goal during the match?

  • player may be a random choice between the top-goalscorer of teamA or teamB
$\endgroup$
  • 1
    $\begingroup$ It seems like a good idea, I think I'll give it a try, but for the moment I'll wait for more possible solutions before mark your answer as the good one. Thanks! $\endgroup$ – daniegarcia254 Jan 22 '17 at 13:06
  • $\begingroup$ Your welcome. Personally I tend to avoid AI if it is not strictly necessary. If I'm not mistaken, using an AI based approach would mean retraining your model daily with the data of the new matches; that's probably an overkill. We'll see, I'm curious about clever AI solutions. $\endgroup$ – Lovecraft Jan 22 '17 at 13:25
  • $\begingroup$ You can use historical data like average no. of goals in the season, player statistics, etc to generate questions that would seem more real! For example, if a team has scored on average 3 goals in the season till now, the question could be "Will there be more than 3 goals scored in the match?" Or suppose if a player A scored no goals as a striker, "Will A score a goal today?" $\endgroup$ – kiner_shah Jan 22 '17 at 15:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.