Hybrid neural network

Hybrid neural network is a network with a combination of different artificial neural networks and approaches. It utilizes features of varying neural networks and approaches to achieve optimum results.

TreNet is a hybrid neural network used for trend prediction. It is composed of a long short term memory recurrent neural network (LSTM RNN) to capture dependency in historical trends, a convolutional neural network (CNN) to extract local features from local raw data and a feature fusion layer to merge and maximize features drawn from LSTM RNN and CNN.