Ensemble-Learning

Ensemble learning to identify three diffrent types of American coal.
1st Model- cnn
2nd Model- resnet50
3rd Model- siamese network

Test Accuracy -
cnn- 88.20%
resnet50- 96.82%
siamese network- 95.19%
weighted ensemble model-97.94%

Following are the weights giving best ensemble accuracy-
cnn- 2
resnet50- 4.5
siamese network- 3.5

Weights/Accuracy plot

Name		Name	Last commit message	Last commit date
Latest commit History 32 Commits
.ipynb_checkpoints		.ipynb_checkpoints
coal_data		coal_data
modified		modified
saved_model		saved_model
README.md		README.md
augmentation_on_entire_data.ipynb		augmentation_on_entire_data.ipynb
cnn_architecture.py		cnn_architecture.py
cnn_modified.ipynb		cnn_modified.ipynb
convnet.ipynb		convnet.ipynb
data_augmentation.ipynb		data_augmentation.ipynb
ensemble.ipynb		ensemble.ipynb
finetuned_resnet.ipynb		finetuned_resnet.ipynb
resnet_architecture.py		resnet_architecture.py
resnet_final.ipynb		resnet_final.ipynb
siamese_architecture.py		siamese_architecture.py
siamese_net.ipynb		siamese_net.ipynb
test_cnn.ipynb		test_cnn.ipynb
test_resnet.ipynb		test_resnet.ipynb
test_siamese_net.ipynb		test_siamese_net.ipynb
weighted_ensemble.ipynb		weighted_ensemble.ipynb

Provide feedback