Francis Eng Hock Tay

Application of support vector machines in financial time series forecasting

This paper deals with the application of a novel neural network technique, support vector machine (SVM), in financial time series forecasting. The objective of this paper is to examine the feasibility of SVM in financial time series forecasting by comparing it with a multi-layer back-propagation (BP) neural network. Five real futures contracts that are collated from the Chicago Mercantile Market are used as the data sets. The experiment shows that SVM outperforms the BP neural network based on the criteria of normalized mean square error (NMSE), mean absolute error (MAE), directional symmetry (DS) and weighted directional symmetry (WDS). Since there is no structured way to choose the free parameters of SVMs, the variability in performance with respect to the free parameters is investigated in this study. Analysis of the experimental results proved that it is advantageous to apply SVMs to forecast financial time series.

Support vector machine with adaptive parameters in financial time series forecasting

A novel type of learning machine called support vector machine (SVM) has been receiving increasing interest in areas ranging from its original application in pattern recognition to other applications such as regression estimation due to its remarkable generalization performance. This paper deals with the application of SVM in financial time series forecasting. The feasibility of applying SVM in financial forecasting is first examined by comparing it with the multilayer back-propagation (BP) neural network and the regularized radial basis function (RBF) neural network. The variability in performance of SVM with respect to the free parameters is investigated experimentally. Adaptive parameters are then proposed by incorporating the nonstationarity of financial time series into SVM. Five real futures contracts collated from the Chicago Mercantile Market are used as the data sets. The simulation shows that among the three methods, SVM outperforms the BP neural network in financial forecasting, and there are comparable generalization performance between SVM and the regularized RBF neural network. Furthermore, the free parameters of SVM have a great effect on the generalization performance. SVM with adaptive parameters can both achieve higher generalization performance and use fewer support vectors than the standard SVM in financial forecasting.

Modified support vector machines in financial time series forecasting

This paper proposes a modified version of support vector machines, called C-ascending support vector machine, to model non-stationary financial time series. The C-ascending support vector machines are obtained by a simple modification of the regularized risk function in support vector machines, whereby the recent e-insensitive errors are penalized more heavily than the distant e-insensitive errors. This procedure is based on the prior knowledge that in the non-stationary financial time series the dependency between input variables and output variable gradually changes over the time, specifically, the recent past data could provide more important information than the distant past data. In the experiment, C-ascending support vector machines are tested using three real futures collected from the Chicago Mercantile Market. It is shown that the C-ascending support vector machines with the actually ordered sample data consistently forecast better than the standard support vector machines, with the worst performance when the reversely ordered sample data are used. Furthermore, the C-ascending support vector machines use fewer support vectors than those of the standard support vector machines, resulting in a sparser representation of solution.

Financial Forecasting Using Support Vector Machines

The use of Support Vector Machines (SVMs) is studied in financial forecasting by comparing it with a multi-layer perceptron trained by the Back Propagation (BP) algorithm. SVMs forecast better than BP based on the criteria of Normalised Mean Square Error (NMSE), Mean Absolute Error (MAE), Directional Symmetry (DS), Correct Up (CP) trend and Correct Down (CD) trend. S&P 500 daily price index is used as the data set. Since there is no structured way to choose the free parameters of SVMs, the generalisation error with respect to the free parameters of SVMs is investigated in this experiment. As illustrated in the experiment, they have little impact on the solution. Analysis of the experimental results demonstrates that it is advantageous to apply SVMs to forecast the financial time series.

Economic and financial prediction using rough sets model

Abstract A state-of-the-art review of the literature related to economic and financial prediction using rough sets model is presented, with special emphasis on the business failure prediction, database marketing and financial investment. These three applications require the accurate prediction of the future states based on the identification of patterns in the historical data. In addition, the historical data are in the format of a multi-attribute information table. All these conditions suit the rough sets model, an effective tool for multi-attribute classification problems. The different rough sets models and issues concerning the implementation of rough sets model – indicator selection, discretization and validation test, are also discussed in this paper. This paper will demonstrate that rough sets model is applicable to a wide range of practical problems pertaining to economic and financial prediction. In addition, the results show that the rough sets model is a promising alternative to the conventional methods for economic and financial prediction.

Fault diagnosis using Rough Sets Theory

Abstract The fault diagnosis on diesel engine is a difficult problem due to the complex structure of the engine and the presence of multi-excite sources. Usually, one method of fault diagnosis can only inspect one corresponding fault category. In this paper, a new method, Rough Sets Theory, is used to diagnose the valve fault for a multi-cylinder diesel engine. Through the analysis of the final reducts generated using Rough Sets Theory, it is shown that this new method is effective for valve fault diagnosis. Rough Sets analysis requires discretizing the fault condition attributes. However, in practice, some of the limits of these attributes are unknown. A new discretization method has been created and the method is found to be suitable for discretizing the attributes without a priori knowledge.

A novel micro-machining method for the fabrication of thick-film SU-8 embedded micro-channels

In this paper, a novel method to realize embedded micro-channels is presented. The presented technology is based on a direct write technique using proton beams to pattern thick-film SU-8. This proton micro-machining method allows the production of high aspect ratio and complex three-dimensional micro-structures in polymers with aspect ratios of over 100 and 20 using poly(methylmethacrylate) (PMMA) and SU-8 respectively. As the SU-8 is used as a structural material, its mechanical properties have to be characterized. For a start, the Youngs modulus of the proton beam exposed SU-8 is determined using a stylus-type load-deflection method. The second part of this paper describes the underlying theory and method used by the author to determine the Youngs modulus of the proton beam exposed SU-8. Measurements of the SU-8 micro-structures show that the Youngs modulus is dependent on the proton beam exposure dose. An exposure dose of 9.5 nC mm-2 results in an average Youngs modulus value of 4.254 GPa.

A modified Chi2 algorithm for discretization

Since the ChiMerge algorithm was first proposed by Kerber (1992), it has become a widely used and discussed discretization method. The Chi2 algorithm is a modification to the ChiMerge method. It automates the discretization process by introducing an inconsistency rate as the stopping criterion and it automatically selects the significance value. In addition, it adds a finer phase aimed at feature selection to broaden the applications of the ChiMerge algorithm. However, the Chi2 algorithm does not consider the inaccuracy inherent in ChiMerges merging criterion. The user-defined inconsistency rate also brings about inaccuracy to the discretization process. These two drawbacks are first discussed in this paper and modifications to overcome them are then proposed. By comparison, results with the original Chi2 algorithm using C4.5, the modified Chi2 algorithm, performs better than the original Chi2 algorithm. It becomes a completely automatic discretization method.

Nonlinear dielectric thin films for high-power electric storage with energy density comparable with electrochemical supercapacitors

Although batteries possess high energy storage density, their output power is limited by the slow movement of charge carriers, and thus capacitors are often required to deliver high power output. Dielectric capacitors have high power density with fast discharge rate, but their energy density is typically much lower than electrochemical supercapacitors. Increasing the energy density of dielectric materials is highly desired to extend their applications in many emerging power system applications. In this paper, we review the mechanisms and major characteristics of electric energy storage with electrochemical supercapacitors and dielectric capacitors. Three types of in-house-produced ferroic nonlinear dielectric thin film materials with high energy density are described, including (Pb_0.97La_0.02)(Zr_0.90Sn_0.05Ti_0.05)O₃ (PLZST) antiferroelectric ceramic thin films, Pb(Zn_1/3Nb_2/3)O_3-Pb(Mg_1/3Nb_2/3) O_3-PbTiO₃ (PZN-PMN-PT) relaxor ferroelectric ceramic thin films, and poly(vinylidene fluoride) (PVDF)-based polymer blend thin films. The results showed that these thin film materials are promising for electric storage with outstandingly high power density and fairly high energy density, comparable with electrochemical supercapacitors.

Measurement of longitudinal piezoelectric coefficient of thin films by a laser-scanning vibrometer

A laser scanning vibrometer (LSV) was used for the first time to measure the piezoelectric coefficient of ferroelectric thin films based on the converse piezoelectric effect. The significant advantages of the use of the LSV or this purpose were demonstrated. Several key points were discussed in order to achieve reliable and accurate results.