Loading...

Real Time Trend Forecasting of Noisy Signal Using Deep Recurrent LSTM Network

Aghaee, Arman | 2021

354 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54076 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Vosoughi Vahdat, Bijan
  7. Abstract:
  8. Artificial neural networks are mathematical models inspired by the nervous system and brain. The types and applications of these networks are very widespread nowadays, and it seems that they can be used to track the signals well and estimate the data of the next. In this research, we try to present a model that can predict the future of the trend of noisy signals that have unpredictable behavior, or in other words, chaotic signals. Such research is also widely used in the medical sciences, including the diagnosis of epileptic seizures or heart attacks. In this research, a study with high volatility financial data has been done as an example on this issue and the proposed model tries to be used instead of a professional decision maker that forecasts the change in the trend. That is, unlike previous methods, this time instead of labeling the amount of future data, this is the time that is labeled, and the target is predicting the critical times, which are the relative minimums and maximums in the upcoming data to determine critical moments of trend change. To present the model, some important features of the data, such as RSI and ATR, are extracted using signal processing, and then, using a deep neural network of LSTM type with the Categorical Cross-Entropy cost function, decision-making networks are designed to predict the change in trend. The network has since become an automated trader to examine its actual performance on the same financial data and compare it to other models that have taken a different approach. In this comparison, the better performance of the proposed system compared to the competitors was found due to making 165% profit in a period of 6 months. Also, trading risk has been examined in this system and it has been acceptable, but competitors have not provided information about it
  9. Keywords:
  10. Artificial Intelligence ; Time Series Prediction ; Recurrent Neural Networks ; Long Short Term Memory (LSTM) ; Chaotic Signal ; Financial Signals

 Digital Object List

 Bookmark

No TOC