Following are the list of files along with their explainations:
- GridSearchANN: A python script which is grid searching parameters for an ANN network, to be run as a script on competion's kaggle(Because of the input files)
- 1DCNNcolumnInput: A python script which is which is based on 1D convolutional NN and uses the whole sequence as input, to be run as a script on competion's kaggle(Because of the input files)
- 1DCNN100features: A python script which is which is based on 1D convolutional NN and uses the reshaped data, to be run as a script on competion's kaggle(Because of the input files)
- 2DCNN100features: A python script which is which is based on 2D convolutional NN and uses the reshaped data, to be run as a script on competion's kaggle(Because of the input files)
- 128cnn_2D_100_features: A python script which is which is a deeper 2D convolutional NN and uses the reshaped data, to be run as a script on competion's kaggle(Because of the input files)
- 2dzrandomforest: A jupyter notebook which is a combinaion of VAE and random forest with 2D latent dimension, to be run as a notebook on competion's kaggle(Because of the input files)
- 30DZrandomForest: A jupyter notebook which is a combinaion of CNN VAE and random forest with 30D latent dimension, to be run as a notebook on competion's kaggle(Because of the input files)
- 30VAEDeeperConv: A jupyter notebook which is a combinaion of a deeper CNN VAE and random forest with 30D latent dimension, to be run as a notebook on competion's kaggle(Because of the input files)
- 50DVAERandomForest: A jupyter notebook which is a combinaion of a CNN VAE and random forest with 50D latent dimension, to be run as a notebook on competion's kaggle(Because of the input files)
- LSTM6432: A jupyter notebook contains 2 layers LSTM layer(64 units followed by 32 units) and uses the reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- LSTM64322: A jupyter notebook contains 2 layers LSTM layer(64 units followed by 32 units) and uses different reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- LSTM1286432: A jupyter notebook contains 3 layers LSTM layer(128 units folled by 64 units followed by 32 units) and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- convlstm64: A jupyter notebook which is a combination of CNN(64*32) and LSTM(1 layer with 64 units) and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- conv2DLSTM6432: A jupyter notebook which is a combination of CNN(64*32) and LSTM(2 layers, 64 and 32 units) and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- conv_128_64_LSTM_64: A jupyter notebook which is a combination of CNN(128*64) and LSTM(one layer with 64 units) and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- conv_128_64_32_LSTM_64: A jupyter notebook which is a combination of CNN(1286432) and LSTM(one layer with 64 units) and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- conv2DLSTMGridSearch: A jupyter notebook which is a gridsearch of CNN+LSTM and tries different filters, kernels and depth for CNN and LSTM and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
- bidirectionalLSTM: A jupyter notebook contains 2 layers LSTM layer(64 units followed by 32 units) and first layer is bidirectional and uses reshaped data, to be run as a notebook on competion's kaggle(Because of the input files)
The public score of most of the models has been mentioned in their files