Description
Comparing Machine Learning and Traditional Time Series Techniques for Financial Series Prediction
Over the past 5 years Machine Learning, Big Data and Data Science have become popular in finance, with jobs postings requiring these skills rising dramatically. However, the financial industry, like any other, is prone to fads. This year I have undertaken a research piece investigating whether Machine Learning models can outperform a model derived using 'traditional' time series prediction techniques (in this case an ARIMA model) and whether Machine Learning can provide an analyst with any additional insights into what could be driving trends observed in financial time series data. The latter suggesting that Machine Learning is another tool in the analyst's tool kit to be used alongside traditional methods rather than a replacement.