Financial Data Time Series Forecasting Using Neural Networks and a Comparative Study

Algorithms that are based on Machine learning & Deep learning are new methods to solve time series prediction challenges. These methods have been proved to yield better results than the traditional regression models. A study concluded that artificial Recurrent Neural Networks such as Long Short Term Memory and Autoregressive Integrated Moving Average models are top models for this application. This research reports analysis and comparison between neural networks such as Convolution Neural Network, Long Short Term Memory, Bidirectional LSTM and Autoregressive Integrated Moving Average models. Comparative analysis of the above models shows that the BiLSTM performs better than ARIMA, CNN and LSTM models.