Filippo Neri

Agent-based modeling under partial and full knowledge learning settings to simulate financial markets

In the paper we show how L-FABS can be applied in a partial knowledge learning scenario or a full knowledge learning scenario to approximate financial time series. L-FABS combines agent-based simulation with machine learning to model the behavior of financial time series.We also discuss why Partial Knowledge and Full Knowledge learning scenario are relevant to the modeling of financial time series and how they can be used to assess the robustness of a modeling system for financial time series. In a Partial Knowledge learning setting usually only the initial conditions of the time series are provided, while in a Full Knowledge learning scenario any value of the financial time series is exploited as soon as it is available.An extensive experimental analysis of L-FABS is reported under a variety of financial time series and time frames.

Learning and predicting financial time series by combining natural computation and agent simulation

We investigate how, by combining natural computation and agent based simulation, it is possible to model financial time series. The agent based simulation can be used to functionally reproduce the structure of a financial market while the natural computation technique finds the most suitable parameter for the simulator. Our experimentation on the DJIA time series shows the effectiveness of this approach in modeling financial data. Also we compare the predictions made by our system to those obtained by other approaches.

Learning predictive models for financial time series by using agent based simulations

In this work, we discuss a computational technique to model financial time series combining a learning component with a simulation one. An agent based model of the financial market is used to simulate how the market will evolve in the short term while the learning component based on evolutionary computation is used to optimize the simulation parameters. Our experimentations on the DJIA and SP500 time series show the effectiveness of our learning simulation system in their modeling. Also we test its robustness under several experimental conditions and we compare the predictions made by our system to those obtained by other approaches. Our results show that our system is as good as, if not better than, alternative approaches to modeling financial time series. Moreover we show that our approach requires a simple input, the time series for which a model has to be learned, versus the complex and feature rich input to be given to other systems thanks to the ability of our system to adjust its parameters by learning.

A comparative study of a financial agent based simulator across learning scenarios

Integrating agent based modeling with learning results in a promising methodology to model the behavior of financial markets. We discuss here how partial and full knowledge learning setups can be combined with agent based modeling to approximate the behavior of financial time series. Partial knowledge learners operate with limited knowledge of the domain, usually only the initial conditions are used. While full knowledge learners use any domain data any time it is made available to adjust their predictions. We report in this paper an experimental study of our learning system L-FABS, introduced in previous works, in order to show how it can discover models for approximating time series working in partial knowledge and full knowledge learning scenarios.

Case Study on Modeling the Silver and Nasdaq Financial Time Series with Simulated Annealing

This paper reports a case study on modeling the SPDR Silver Trust (SLV) and Nasdaq Composite Index timeseries by using a financial agent based system using simulated annealing. We show here how adding financial information to the modeling system can significantly improve the modeling results. The learning system LFABS, previously developed by the author, will be used as a testbed for the empirical evaluation of the proposed methodology on the two case studies.