Considers the modelling of dynamic systems using biased regression and spectral methods. Provides evidence on the power of transfer function modelling for unravelling the…
Abstract
Considers the modelling of dynamic systems using biased regression and spectral methods. Provides evidence on the power of transfer function modelling for unravelling the empirical connection between endogenous and exogenous (control) variables in both regression type and spectral input‐output systems. The Multiple Input Transfer Function Noise Model – of specific value when the input variables are collinear – has previously been used to demonstrate the connection between macroeconomic forces and stock market pricing on a thin security market. Compares the adequacy of representative time and frequency domain algorithms for modelling observed data series. The estimations are done with the combined Transfer Function and Cartesian ARIMA Search algorithm of Östermark and Höglund and the CAPM/APM programs of Östermark.
Details
Keywords
Provides evidence on the power of transfer function noise modelling in explaining the empirical connection between endogenous and exogenous (control) variables in linear…
Abstract
Provides evidence on the power of transfer function noise modelling in explaining the empirical connection between endogenous and exogenous (control) variables in linear regression type input‐output systems. The multiple input transfer function noise model – of specific value when the input variables are collinear – is used to demonstrate the connection between macroeconomic forces and stock market pricing on a thin security market. Shows that the transfer function approach provides new evidence partly in conflict with previous results obtained by ordinary least squares methodology. Previous empirical evidence suggests that money supply, inflation, the level of industrial production and the psychological impact of the general index of the Stockholm Stock Exchange affects Finnish stock pricing. The problem of selecting relevant economic state variables is tackled by regressing each of the five factor time series obtained from testing the arbitrage pricing theory (see Östermark, circa 1989) on the set of tentative state variables. The economic state variables are significant explanators of stock pricing, both at the market and at the individual asset level. Only nine individual stocks are tested. Comprehensive testing of all individual stocks is left for future research.
Details
Keywords
Presents for the first time empirical evidence on the forecasting performance of multi‐layer neural nets in modelling multiple‐input vector time series processes. Compares the…
Abstract
Presents for the first time empirical evidence on the forecasting performance of multi‐layer neural nets in modelling multiple‐input vector time series processes. Compares the results produced by the neural net with those obtained by a robust VARMAX‐algorithm and a multiple‐input state space algorithm for vector‐valued time series processes. The neural net and the VARMAX‐algorithm were programmed in the C‐language and the state space algorithm was programmed in FORTRAN.
Details
Keywords
To demonstrate the scalability of the genetic hybrid algorithm (GHA) in monitoring a local neural network algorithm for difficult non‐linear/chaotic time series problems.
Abstract
Purpose
To demonstrate the scalability of the genetic hybrid algorithm (GHA) in monitoring a local neural network algorithm for difficult non‐linear/chaotic time series problems.
Design/methodology/approach
GHA is a general‐purpose algorithm, spanning several areas of mathematical problem solving. If needed, GHA invokes an accelerator function at key stages of the solution process, providing it with the current population of solution vectors in the argument list of the function. The user has control over the computational stage (generation of a new population, crossover, mutation etc) and can modify the population of solution vectors, e.g. by invoking special purpose algorithms through the accelerator channel. If needed, the steps of GHA can be partly or completely superseded by the special purpose mathematical/artificial intelligence‐based algorithm. The system can be used as a package for classical mathematical programming with the genetic sub‐block deactivated. On the other hand, the algorithm can be turned into a machinery for stochastic analysis (e.g. for Monte Carlo simulation, time series modelling or neural networks), where the mathematical programming and genetic computing facilities are deactivated or appropropriately adjusted. Finally, pure evolutionary computation may be activated for studying genetic phenomena. GHA contains a flexible generic multi‐computer framework based on MPI, allowing implementations of a wide range of parallel models.
Findings
The results indicate that GHA is scalable, yet due to the inherent stochasticity of neural networks and the genetic algorithm, the scalability evidence put forth in this paper is only indicative. The scalability of GHA follows from maximal node intelligence allowing minimal internodal communication in problems with independent computational blocks.
Originality/value
The paper shows that GHA can be effectively run on both sequential and parallel platforms. The multicomputer layout is based on maximizing the intelligence of the nodes – all nodes are provided with the same program and the available computational support libraries – and minimizing internodal communication, hence GHA does not limit the size of the mesh in problems with independent computational tasks.
Details
Keywords
Rune Höglund and Ralf Östermark
Previous evidence suggests that the relationship between different stock markets is unstable over time. In particular, the Finnish and Japanese financial economies are…
Abstract
Previous evidence suggests that the relationship between different stock markets is unstable over time. In particular, the Finnish and Japanese financial economies are interrelated and exhibit non‐linear behaviour. Presents an approximation of the influence of the Japanese stock market on the Finnish derivatives market by an adaptive recursive least squares (RLS) algorithm. The parameters are allowed to change over time through a discounting factor, thus providing a convenient means for recognizing past information to a specified degree. Following the reasoning of Bera et al. (1992), shows that the RLS algorithm is, theoretically, able to cope with conditional heteroscedasticity. Compares the results with different values on the discount factor and when choosing a suitable value the ARCH‐like effects in the residuals seem to vanish. On the other hand, some new peculiarities in the RLS residuals emerge when ARCH effects are eliminated. The results indicate that the standard RLS algorithm combined with a proper specification of the discount factor could be useful in studying relationships of this kind.
Details
Keywords
To propose a new algorithmic platform (minlp_machine) for complex mixed‐integer non‐linear programming (MINLP) problems.
Abstract
Purpose
To propose a new algorithmic platform (minlp_machine) for complex mixed‐integer non‐linear programming (MINLP) problems.
Design/methodology/approach
The platform combines features from classical non‐linear optimization methodology with novel innovations in computational techniques. The system constructs discrete search zones around non‐integer discrete‐valued variables of local solutions, which reduces the search process significantly. In complicated problems fast feasibility restoration is achieved through concentrated Hessians. The system is programmed in strict ANSI C and can be run either stand alone or as a support library for other programs. File I/O is designed to recognize possible usage in both single and parallel processor environments.
Findings
The system has been tested on Alpha and Sun mainframes and – as a support library for a Genetic Hybrid Algorithm (GHA()) – in Linux and IBM parallel supercomputer environments. The constrained problem can, for example, be solved through a sequence of first order Taylor approximations of the non‐linear constraints and occasional feasibility restoration through Hessian information of the Lagrangian of the MINLP problem, or by invoking a nonlinear solver like SQP directly in the branch and bound tree. The system has been successfully tested on a small sample of representative continuous‐valued non‐linear programming problems.
Originality/value
It is demonstrated that – through zone‐constrained search – minlp_machine() outperforms some recent competing approaches with respect to the number of nodes in the branch and bound tree.
Details
Keywords
Chaoqun Ma, Lan Liu, Junbo Wang and Jing Chen
The purpose of this paper is to examine the risk of inefficiency of China's stock index futures market by investigating the opportunity and profitability of exchange‐traded fund…
Abstract
Purpose
The purpose of this paper is to examine the risk of inefficiency of China's stock index futures market by investigating the opportunity and profitability of exchange‐traded fund (ETF) arbitrage. The explanation of behavioral risk to market efficiency is examined.
Design/methodology/approach
Based on cost‐of‐carry model, some assumptions about market efficiency were examined, and statistical tests were implemented to support the findings.
Findings
In China, borrowing and lending interest rates are quite different; dividends are small and paid in an irregular manner; and short sale cannot be used in arbitrage by all investors. It is found that the Chinese index futures market is far from efficient.
Originality/value
With reference to the empirical study, this is believed to be the first application of behavioral study to the study of market efficiency. The analysis of the statistics about Chinese index futures market and the algorithm parameters are very valuable for in‐depth understanding of the emerging markets.
Details
Keywords
The performance of Aoki’s state space algorithm and the Cartesian ARIMA search algorithm (CARlMA) of Östermark and Höglund is compared. The analysis is carried out on a set of…
Abstract
The performance of Aoki’s state space algorithm and the Cartesian ARIMA search algorithm (CARlMA) of Östermark and Höglund is compared. The analysis is carried out on a set of stock prices on the Helsinki (Finland) and Stockholm (Sweden) Stock Exchanges. Demonstrates that the Finnish and Swedish stock markets differ in predictability of stock prices. With Finnish stock data, Aoki’s state space algorithm outperforms the subset of MAPE minimizing forecasts. In contrast, with Swedish stock data, ARIMA‐models of a fairly simple structure outperform Aoki’s algorithm. The stock markets are seen to differ in complexity of time series models as well as in predictability of individual asset prices.
Details
Keywords
To discuss a new parallel algorithmic platform (minlp_machine) for complex mixed‐integer non‐linear programming (MINLP) problems.
Abstract
Purpose
To discuss a new parallel algorithmic platform (minlp_machine) for complex mixed‐integer non‐linear programming (MINLP) problems.
Design/methodology/approach
The platform combines features from classical non‐linear optimization methodology with novel innovations in computational techniques. The system constructs discrete search zones around noninteger discrete‐valued variables at local solutions, which simplifies the local optimization problems and reduces the search process significantly. In complicated problems fast feasibility restoration may be achieved through concentrated Hessians. The system is programmed in strict ANSI C and can be run either stand alone or as a support library for other programs. File I/O is designed to recognize possible usage in both single and parallel processor environments. The system has been tested on Alpha, Sun and Linux mainframes and parallel IBM and Cray XT4 supercomputer environments. The constrained problem can, for example, be solved through a sequence of first order Taylor approximations of the non‐linear constraints and feasibility restoration utilizing Hessian information of the Lagrangian of the MINLP problem, or by invoking a nonlinear solver like SQP directly in the branch and bound tree. minlp_machine( ) has been tested as a support library to genetic hybrid algorithm (GHA). The GHA(minlp_machine) platform can be used to accelerate the performance of any linear or non‐linear node solver. The paper introduces a novel multicomputer partitioning of the discrete search space of genuine MINLP‐problems.
Findings
The system is successfully tested on a small sample of representative MINLP problems. The paper demonstrates that – through concurrent nonlinear branch and bound search – minlp_machine( ) outperforms some recent competing approaches with respect to the number of nodes in the branch and bound tree. Through parallel processing, the computational complexity of the local optimization problems is reduced considerably, an important aspect for practical applications.
Originality/value
This paper shows that binary‐valued MINLP‐problems will reduce to a vector of ordinary non‐linear programming on a suitably sized mesh. Correspondingly, INLP‐ and ILP‐problems will require no quasi‐Newton steps or simplex iterations on a compatible mesh.
Details
Keywords
To solve the multi‐period portfolio management problem under transactions costs.
Abstract
Purpose
To solve the multi‐period portfolio management problem under transactions costs.
Design/methodology/approach
We apply a recently designed super genetic hybrid algorithm (SuperGHA) – an integrated optimisation system for simultaneous parametric search and non‐linear optimisation – to a recursive portfolio management decision support system (SHAREX). The parametric search machine is implemented as a genetic superstructure, producing tentative parameter vectors that control the ultimate optimisation process.
Findings
SHAREX seems to outperform the buy and hold‐strategy on the Finnish stock market. The potential of a technical portfolio system is best exploitable under favorable market conditions.
Originality/value
A number of robust engines for matrix algebra, mathematical programming and numerical calculus have been integrated with SuperGHA. The engines expand its scope as a general‐purpose algorithm for mathematical programming.