is it possible to train several Neural Networks on different types of data and combine them?

Question

I want to create a NHL game predictor and have already trained one neural network on game data.

What I would like to do is train another model on player seasonal/game data and combine the two models to archive better accuracy.

Is this approach feasible? If it is, how do I go about doing it?

EDIT:

I have currently trained a neural network to classify the probability of the home team winning a game on a dataset that looks like this:

h_Won/Lost  h_metric2 h_metric3 h_metric4 a_metric2 a_metric3 a_metric4 h_team1 h_team2 h_team3 h_team4 a_team1 a_team2 a_team3 a_team4
 1            10       10         10        10         10        10      1       0        0      0         0      1        0      0
 1            10       10         10        10         10        10      1       0        0      0         0      1        0      0
 1            10       10         10        10         10        10      1       0        0      0         0      1        0      0

and so on.

I am preparing a dataset of player-data for each game that will have the shape of this:

Player     PlayerID    Won/Lost     team      opponent     metric1     metric2   
 Henke         1           1          NY          CAP         10          10

Hopefully, this new dataset will have some accuracy on if team is going to have some predictive features that are good and recognised.

Now, say I have these two trained Nural Networks and they both have an accuracy of 70% by them self. But I want to combine them both in the hopes to achieve better predictability. How is this archived? How will the test-dataset be structured?

Answer 1

The term you need is “model ensembles”, that’s the way models are combined. Pretty hard to be more specific since you don’t give a language or any other details.