Online Non-linear Prediction of Financial Time Series Patterns

We consider a mechanistic non-linear machine learning approach to learning signals in financial time series data. A modularised and decoupled algorithm framework is established and is proven on daily sampled closing time-series data for JSE equity markets. The input patterns are based on input data vectors of data windows preprocessed into a sequence of daily, weekly and monthly or quarterly sampled feature measurement changes (log feature fluctuations). The data processing is split into a batch processed step where features are learnt using a Stacked AutoEncoder (SAE) via unsupervised learning, and then both batch and online supervised learning are carried out on Feedforward Neural Networks (FNNs) using these features. The FNN output is a point prediction of measured time-series feature fluctuations (log differenced data) in the future (ex-post). Weight initializations for these networks are implemented with restricted Boltzmann machine pretraining, and variance based initializations. The validity of the FNN backtest results are shown under a rigorous assessment of backtest overfitting using both Combinatorially Symmetrical Cross Validation and Probabilistic and Deflated Sharpe Ratios. Results are further used to develop a view on the phenomenology of financial markets and the value of complex historical data under unstable dynamics.

Grantor

Department of Statistical Sciences

Year dc:date.issued

2020

Author dc:creator

• da Costa, Joel

Advisor dc:contributor.advisor

• Gebbie, Timothy

• online learning
• feedforward neural network
• restricted Boltzmann machine
• variance weight initialization
• stacked autoencoder
• pattern prediction
• JSE
• non-linear
• financial time series
• combinatorially symmetrical cross validation
• backtest overfitting
• deflated Sharpe ratio
• probabilistic Sharpe ratio

Handle dc:identifier.uri

http://hdl.handle.net/11427/32221

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/32221