Add more parameters to the MLP training method

[FlorentinD]

Is your feature request related to a problem? Please describe.
In the machine learning pipelines, you can use different trainer methods when adding a model candidate, such as RandomForest or MLP.
For MLP, we could expose more parameters to configure.

Current exposed parameters are listed at https://neo4j.com/docs/graph-data-science/current/machine-learning/training-methods/mlp/.

Describe the solution you would like

We could also expose other optional parameters such as:

• Different activiation functions (in addition to Softmax)
• Dropouts
• Optimizer

These are just examples, you could also come up with others.

Additional context

Good starting points are MLPClassifierTrainConfig and MLPClassifier

If you want to work on this issue please drop a comment :)