add ensemble + ast transformer feature

[bagustris]

• add ensemble module: tested with: python3 -m nkululeko.ensemble tests/exp_emodb_os_knn.ini tests/exp_emodb_os_svm.ini --method mean (after running ./run_test2.sh), this module is also added to run_test2.sh
• add -overwrite argument ($2) to run_test2.sh
• add log.out and log.err as the outputs of run_test2.sh

Results of ensemble on Ravdess dataset:

os_xgb = 0.601
praat_knn = 0.500
os_praat_xgb = 0.632

ensemble: 2 classifier
mean = 0.661
sum = 0.661
max = 0.652

ensemble: 3 classifiers:
majority_vote= 0.652
mean = 0.643
max = 0.652 
sum = 0.643

run_test2.sh results

Total tests passed: 36
Total tests failed: 6
Failed modules and INI files:
nkululeko with exp_emodb_trill_test.ini
nkululeko with exp_agedb_os_mlp.ini
demo with exp_emodb_os_mlp.ini
demo with exp_agedb_os_mlp.ini
test with exp_emodb_trill_test.ini
explore with exp_agedb_explore_data.ini
Total time taken: 538 seconds

Attached log.out and log.err form test above.

log.err.txt log.out.txt