THE quest for alpha : can artificial neural networks help?

The application of artificial neural networks (ANN) in finance is relatively new area of research. We employed ANNs that used both fundamental and technical inputs to predict future prices of widely held Australian stocks and used these predicted prices for stock portfolio selection over a 10-year period (2001-2011). We found that the ANNs generally do well in predicting the direction of stock price movements. The stock portfolios selected by the ANNs with median accuracy are able to generate positive alpha over the 10-year period. More importantly, we found that a portfolio based on randomly selected network configuration had zero chance of resulting in a significantly negative alpha but a 27% chance of yielding a significantly positive alpha. This is in stark contrast to the findings of the research on mutual fund performance where active fund managers with negative alphas outnumber those with positive alphas.

Prediction and elucidation of the population dynamics of Microcystis spp. in Lake Dianchi (China) by means of artificial neural networks

Lake Dianchi is a shallow and turbid lake, located in Southwest China. Since 1985, Lake Dianchi has experienced severe cyanabacterial blooms (dominated by Microcystis spp.). In extreme cases, the algal cell densities have exceeded three billion cells per liter. To predict and elucidate the population dynamics ofMicrocystis spp. in Lake Dianchi, a neural network based model was developed. The correlation coefficient (R 2) between the predicted algal concentrations by the model and the observed values was 0.911. Sensitivity analysis was performed to clarify the algal dynamics to the changes of environmental factors. The results of a sensitivity analysis of the neural network model suggested that small increases in pH could cause significantly reduced algal abundance. Further investigations on raw data showed that the response of Microcystis spp. concentration to pH increase was dependent on algal biomass and pH level. When Microcystis spp. population and pH were moderate or low, the response of Microcystis spp. population would be more likely to be positive in Lake Dianchi; contrarily, Microcystis spp. population in Lake Dianchi would be more likely to show negative response to pH increase when Microcystis spp. population and pH were high. The paper concluded that the extremely high concentration of algal population and high pH could explain the distinctive response of Microcystis spp. population to +1 SD (standard deviation) pH increase in Lake Dianchi. And the paper also elucidated the algal dynamics to changes of other environmental factors. One SD increase of water temperature (WT) had strongest positive relationship with Microcystis spp. biomass. Chemical oxygen demand (COD) and total phosphorus (TP) had strong positive effect on Microcystis spp. abundance while total nitrogen (TN), biological oxygen demand in five days (BOD5), and dissolved oxygen had only weak relationship with Microcystis spp. concentration. And transparency (Tr) had moderate positive relationship with Microcystis spp. concentration.

Peptide design by artificial neural networks and computer-based evolutionary search

A technique for systematic peptide variation by a combination of rational and evolutionary approaches is presented. The design scheme consists of five consecutive steps: (i) identification of a “seed peptide” with a desired activity, (ii) generation of variants selected from a physicochemical space around the seed peptide, (iii) synthesis and testing of this biased library, (iv) modeling of a quantitative sequence-activity relationship by an artificial neural network, and (v) de novo design by a computer-based evolutionary search in sequence space using the trained neural network as the fitness function. This strategy was successfully applied to the identification of novel peptides that fully prevent the positive chronotropic effect of anti-β1-adrenoreceptor autoantibodies from the serum of patients with dilated cardiomyopathy. The seed peptide, comprising 10 residues, was derived by epitope mapping from an extracellular loop of human β1-adrenoreceptor. A set of 90 peptides was synthesized and tested to provide training data for neural network development. De novo design revealed peptides with desired activities that do not match the seed peptide sequence. These results demonstrate that computer-based evolutionary searches can generate novel peptides with substantial biological activity.

Studies on the generalisation of Gaussian processes and Bayesian neural networks

The assessment of the reliability of systems which learn from data is a key issue to investigate thoroughly before the actual application of information processing techniques to real-world problems. Over the recent years Gaussian processes and Bayesian neural networks have come to the fore and in this thesis their generalisation capabilities are analysed from theoretical and empirical perspectives. Upper and lower bounds on the learning curve of Gaussian processes are investigated in order to estimate the amount of data required to guarantee a certain level of generalisation performance. In this thesis we analyse the effects on the bounds and the learning curve induced by the smoothness of stochastic processes described by four different covariance functions. We also explain the early, linearly-decreasing behaviour of the curves and we investigate the asymptotic behaviour of the upper bounds. The effect of the noise and the characteristic lengthscale of the stochastic process on the tightness of the bounds are also discussed. The analysis is supported by several numerical simulations. The generalisation error of a Gaussian process is affected by the dimension of the input vector and may be decreased by input-variable reduction techniques. In conventional approaches to Gaussian process regression, the positive definite matrix estimating the distance between input points is often taken diagonal. In this thesis we show that a general distance matrix is able to estimate the effective dimensionality of the regression problem as well as to discover the linear transformation from the manifest variables to the hidden-feature space, with a significant reduction of the input dimension. Numerical simulations confirm the significant superiority of the general distance matrix with respect to the diagonal one.In the thesis we also present an empirical investigation of the generalisation errors of neural networks trained by two Bayesian algorithms, the Markov Chain Monte Carlo method and the evidence framework; the neural networks have been trained on the task of labelling segmented outdoor images.

Neural networks would 'vote' according to Borda's Rule

Can neural networks learn to select an alternative based on a systematic aggregation of convicting individual preferences (i.e. a 'voting rule')? And if so, which voting rule best describes their behavior? We show that a prominent neural network can be trained to respect two fundamental principles of voting theory, the unanimity principle and the Pareto property. Building on this positive result, we train the neural network on profiles of ballots possessing a Condorcet winner, a unique Borda winner, and a unique plurality winner, respectively. We investigate which social outcome the trained neural network chooses, and find that among a number of popular voting rules its behavior mimics most closely the Borda rule. Indeed, the neural network chooses the Borda winner most often, no matter on which voting rule it was trained. Neural networks thus seem to give a surprisingly clear-cut answer to one of the most fundamental and controversial problems in voting theory: the determination of the most salient election method.

Predicting project performance through neural networks

Successful project delivery of construction projects depends on many factors. With regard to the construction of a facility, selecting a competent contractor for the job is paramount. As such, various approaches have been advanced to facilitate tender award decisions. Essentially, this type of decision involves the prediction of a bidderÕs performance based on information available at the tender stage. A neural network based prediction model was developed and presented in this paper. Project data for the study were obtained from the Hong Kong Housing Department. Information from the tender reports was used as input variables and performance records of the successful bidder during construction were used as output variables. It was found that the networks for the prediction of performance scores for Works gave the highest hit rate. In addition, the two most sensitive input variables toward such prediction are ‘‘Difference between Estimate’’ and ‘‘Difference between the next closest bid’’. Both input variables are price related, thus suggesting the importance of tender sufficiency for the assurance of quality production.

Generating rules with predicates, terms and variables from the pruned neural networks

Artificial neural networks (ANN) have demonstrated good predictive performance in a wide range of applications. They are, however, not considered sufficient for knowledge representation because of their inability to represent the reasoning process succinctly. This paper proposes a novel methodology Gyan that represents the knowledge of a trained network in the form of restricted first-order predicate rules. The empirical results demonstrate that an equivalent symbolic interpretation in the form of rules with predicates, terms and variables can be derived describing the overall behaviour of the trained ANN with improved comprehensibility while maintaining the accuracy and fidelity of the propositional rules.