The two files are basics for applying ANN and CNN.The code consists of preprocessing of data and feeding the preprocessed data to the neural networks. The CNN has following steps that are important for any implementation:
1. Convolution Layer ( To create a feature map using various feature detectors and hence creating a convolution layer from the input images). We then apply Relu(Rectified Linear Unit) to learn the non-linear data points in the images
2. Max pooling ( To learn about the spatial invariance and distortment in the images)
3. Flattening (To convert the features in the array which can be fed to ANN)
4. Full Connection (To apply ANN to classify the images)
LSTM model implemented for sentiment analysi on IMBD dataset with 84% validation and 84% testing accuracies which includes:
1. Word Embedding
2. Apply LSTM
3. Fitting and evaluate the model