Piotr Lipinski

Evolution Strategy in Portfolio Optimization

In this paper an evolutionary algorithm to optimize a stock portfolio is presented. The method, based on Evolution Strategies, uses artificial trading experts discovered by a genetic algorithm. This approach is tested on a sample of stocks taken from the French market. Results obtained are compared with the Buy-and-Hold strategy and a stock index. Presented research extends evolutionary methods on financial economics worked out earlier for stock trading.

Evolutionary building of stock trading experts in a real-time system

This paper addresses the problem of constructing real-time stock trading expertise for financial time series. The expertise is arrived at via an evolutionary algorithm on the basis of a set of specified trading rules. As in most real-time expert systems, one of the main bottlenecks is the time constraint. In this paper, two approaches were compared using our system, Bourse-Expert, the first based on 350 trading rules, and the second based on 150 particular linear combinations of these 350 rules. Experiments carried out on real data from the Paris Stock Exchange showed that focusing on only 150 rules highly reduced the computation time without significantly reducing the quality of the expertise.

Building Risk-Optimal Portfolio Using Evolutionary Strategies

In this paper, an evolutionary approach to portfolio optimization is proposed. In the approach, various risk measures are introduced instead of the classic risk measure defined by variance. In order to build the risk-optimal portfolio, three evolutionary algorithms based on evolution strategies are proposed. Evaluations of the approach is performed on financial time series from the Warsaw Stock Exchange.

ECGA vs. BOA in discovering stock market trading experts

This paper presents two evolutionary algorithms, ECGA and BOA, applied to constructing stock market trading expertise, which is built on the basis of a set of specific trading rules analysing financial time series of recent price quotations. A few modifications of ECGA are proposed in order to reduce the computing time and make the algorithm applicable for real-time trading. In experiments carried out on real data from the Paris Stock Exchange, the algorithms were compared in terms of the efficiency in solving the optimization problem, in terms of the financial relevance of the investment strategies discovered as well as in terms of the computing time.

Evolutionary Strategies for Building Risk-Optimal Portfolios

This chapter describes an evolutionary approach to portfolio optimization. It rejects some assumptions from classic models, introduces alternative risk measures and proposes three evolutionary algorithms to solve the optimization problem. In order to validate the approach proposed, results of a number of experiments using data from the Paris Stock Exchange are presented.

Performance Measures in an Evolutionary Stock Trading Expert System

This paper addresses the problem of investment assessment and selection. A number of various performance measures are evaluated and studied. The goal of these investigations is to compare these performance measures on real-life data and to discover an optimal performance measure for selecting investment strategies in an evolutionary stock trading decision support system. Evaluations have been performed on financial time series from the Paris Stock Exchange.

Dependency mining in large sets of stock market trading rules

This paper addresses the problem of dependency mining in large sets. The first goal is to determine and reduce the dimension of data using principal component analysis. The second is to group variables into several classes using Kohonens self-organizing maps and then the K-means algorithm. Evaluations have been performed on 350 financial trading rules (variables) observed in a period of 1300 instants (observations). It was shown that the rules are strongly correlated, all of which can be reproduced from 150 generators with an accuracy of 95%. Moreover, the initial set of 350 rules was subdivided into 23 classes of similar rules.

Pattern Mining in Ultra-High Frequency Order Books with Self-Organizing Maps

This paper addresses the issue of discovering frequent patterns in order book shapes, in the context of the stock market depth, for ultra-high frequency data. It proposes a computational intelligence approach to building frequent patterns by clustering order book shapes with Self-Organizing Maps. An experimental evaluation of the approach proposed on the London Stock Exchange Rebuild Order Book database succeeded with providing a number of characteristic shape patterns and also with estimating probabilities of some typical transitions between shape patterns in the order book.

Infeasibility driven evolutionary algorithm with ARIMA-based prediction mechanism

This paper proposes an improvement of evolutionary algorithms for dynamic objective functions with a prediction mechanism based on the Autoregressive Integrated Moving Average (ARIMA) model. It extends the Infeasibility Driven Evolutionary Algorithm (IDEA) that maintains a population of feasible and infeasible solutions in order to react on changing objectives faster. Combining IDEA with ARIMA leads to a more efficient evolutionary algorithm that reacts faster to the changing objectives which profits from using information coming from the prediction mechanism and remains one time instant ahead of the original algorithm. Preliminary experiments performed on popular benchmark problems confirm that the IDEA-ARIMA outperforms the original IDEA algorithm in many cases.

A new evolutionary gene selection technique

Microarray technology allows to investigate gene expression levels by analyzing high dimensional datasets of few samples. Selection of discriminative, differentially expressed genes from such datasets is important to differentiate, prognose and understand the underlying biological processes. In this regard, the paper presents a new evolutionary gene selection method based on Student-t Stochastic Neighbor Embedding (t-SNE), Differential Evolution (DE) and Support Vector Machine (SVM). Here the underlying classification task of SVM is used as an optimization problem of DE, while t-SNE provides better ordering of genes for selection purpose. Generally, t-SNE is used to reorder the genes in such a way so that similar genes are grouped together and dissimilar genes are kept further apart. These reordered genes are then fragmented into fixed-length partitions. Thereafter, from each partition, a gene is selected randomly to encode the initial population of DE along with the combination of its weight and threshold values in order to participate in fitness computation. In the final generation of DE, a subset of genes is selected based on higher classification accuracy. The proposed technique is tested on six publicly available microarray datasets concerning various cancerous tissues of Homo sapiens and yields a potential set of genes by providing prefect or nearly perfect classification accuracy. Moreover, the superiority of the proposed technique has been demonstrated in comparison with other widely used techniques. Finally, the achieved results have also been justified by a statistical test and allowed us to draw biological conclusions through the identification of Gene Ontologies.