Catálogo de publicaciones - libros

Ant colony systems comprise of a family of algorithms which draw their metaphorical inspiration from the activities and learning mechanisms of social insect societies. A key feature of these societies is their ability to promote problem-solving behaviour between individuals in the absence of a top-down control system. The algorithms can be used for both classification and optimisation purposes.

The key learning mechanisms in the PSO algorithm are driven by a metaphor of social behaviour: that good solutions uncovered by one member of a population are observed and copied by other members of the population. Of course, these learning mechanisms abound in business and other social settings. Good business strategies, good product designs, and good theories stimulate imitation and subsequent adaptation. Particle swarm algorithms have proven to be successful optimisation tools in a variety of applications, and they have clear potential for application to financial modelling.

This chapter presented an introduction to a family of algorithms inspired by an evolutionary metaphor, evolutionary algorithms. Specific EA instances were examined, namely genetic algorithms, differential evolution and genetic programming. Following two examples of how EAs can be applied for financial prediction purposes, we outlined some of the more recent developments in EC. Our presentation of evolutionary computation continues in Chap. 4 with an introduction to the grammatical evolution framework.

This chapter presented an introduction to grammatical evolution, one of the methodologies inspired by an evolutionary metaphor. In Part III of this book we present a number of case studies that illustrate the use of grammatical evolution to generate rules for index trading, intra-day trading, and foreign exchange trading; for the prediction of corporate failure; and finally for the classification of bond ratings. We now continue our exposition of biologically inspired methodologies with examples of social learning algorithms based on swarm intelligence.

The key learning mechanisms in the PSO algorithm are driven by a metaphor of social behaviour: that good solutions uncovered by one member of a population are observed and copied by other members of the population. Of course, these learning mechanisms abound in business and other social settings. Good business strategies, good product designs, and good theories stimulate imitation and subsequent adaptation. Particle swarm algorithms have proven to be successful optimisation tools in a variety of applications, and they have clear potential for application to financial modelling.

Ant colony systems comprise of a family of algorithms which draw their metaphorical inspiration from the activities and learning mechanisms of social insect societies. A key feature of these societies is their ability to promote problem-solving behaviour between individuals in the absence of a top-down control system. The algorithms can be used for both classification and optimisation purposes.

Although natural immune systems are complex and consist of a huge number of individual components, they can be considered as a distributed, selforganising system which operates in a dynamic environment. The mechanisms of natural immune systems, including their ability to distinguish between self and non-self states, and their ability to maintain a memory of previous invaders, provide a rich metaphorical inspiration for the design of patternrecognition and optimisation algorithms. In this chapter we have discussed how two of these metaphors, the negative selection process for T cells, and the clonal selection and expansion of B cells, can be applied to create a classification system and an optimisation algorithm, respectively.

The selection and implementation of a specific biologically inspired algorithm is only one component in the process of developing a complete trading system. No algorithm can compensate for poor-quality data or a poor trading system design. Earlier in the chapter, one of the major sources of information for short-term trading systems (technical indicators) was introduced. The next chapter discusses these in more detail.

Although natural immune systems are complex and consist of a huge number of individual components, they can be considered as a distributed, selforganising system which operates in a dynamic environment. The mechanisms of natural immune systems, including their ability to distinguish between self and non-self states, and their ability to maintain a memory of previous invaders, provide a rich metaphorical inspiration for the design of patternrecognition and optimisation algorithms. In this chapter we have discussed how two of these metaphors, the negative selection process for T cells, and the clonal selection and expansion of B cells, can be applied to create a classification system and an optimisation algorithm, respectively.

Prostate cancer is one of the common cancers where there is good evidence for a larger genetic component to its etiology, but the genetic models are complex. It is highly likely that the PCa predisposition genes will be polygenic and may be interacting within families. Some PCa predisposition genes are likely to be DNA repair genes (e.g., ) but these may account for only a small proportion of young cases. However, the discovery of high-risk mutations has led to the first clinical targeted screening trial based on genotype in this disease (the IMPACT study, discussed above), and this trial will serve as a basis for further targeted screening and chemoprevention trials based on genotype as further genes are identified. The lessons learned in IMPACT will be screening uptake in a high-risk male population, the psychological issues of screening men at higher risk of PCa, the utility of PSA in a higher risk population, the identification of new and better biomarkers and the clinical parameters of PCa so identified.