| Summary: | Machine Learning has become increasingly popular over the past decade for use in various applications including prediction-based tasks allowing companies and industries to leverage data and make better informed decisions. This is especially true in the Financial Markets where investors and Fund Managers have been adopting Neural Network based approaches to automate their Trading Systems, commonly referred to as Algorithmic Trading. However, these Neural Networks are typically devoid of a human-like reasoning ability which makes them like a “black box”. This often limits their approach in the world of Portfolio Management where a Portfolio Manager typically wants to understand how the Neural Network arrived at the output and made the predictions.
In this project, we will explore the use of Fuzzy logic-based systems along with Neural Networks which would help increase the interpretability of the model. This combination model will be applied to different Financial Market data and future asset prices will be forecasted in a look-ahead fashion. In addition, appropriate Trend Reversal indicators such as Moving Average Convergence Divergence (MACD) will be considered to make a buy/sell decision for that asset. The model when combined with MACD, will enable us to increase the effectiveness of a vanilla indicator due to the lookahead forecasts we made.
|
|---|