Improve model accuracy by 20%

Hey experts,
I have attached my machine-learning models here. I was hoping you could improve it by at least 20%.
FTR_predictor
--------------------
Training Data:
--------------------
F1 Score:0.590236289594967
Accuracy:0.590236289594967
--------------------
Test Data:
--------------------
F1 Score:0.5980883022303141
Accuracy:0.5980883022303141
--------------------


Over_1.5_predictor
--------------------
Training Data:
--------------------
F1 Score:0.7743178326708772
Accuracy:0.7743178326708772
--------------------
Test Data:
--------------------
F1 Score:0.774920345926263
Accuracy:0.7749203459262631
--------------------


Over_2.5_predictor
--------------------
Training Data:
--------------------
F1 Score:0.7104050330415275
Accuracy:0.7104050330415275
--------------------
Test Data:
--------------------
F1 Score:0.6949817933545744
Accuracy:0.6949817933545744
--------------------


HT_0.5_predictor
--------------------
Training Data:
--------------------
F1 Score:0.7450559633251238
Accuracy:0.7450559633251238
--------------------
Test Data:
--------------------
F1 Score:0.7411242603550295
Accuracy:0.7411242603550295
--------------------


HTR_predictor
--------------------
Training Data:
--------------------
F1 Score:0.5843839157258163
Accuracy:0.5843839157258163
--------------------
Test Data:
--------------------
F1 Score:0.5844333181611289
Accuracy:0.5844333181611289
--------------------

Budget: 20$
TImeline: Soonest possible time