The aim of this paper is to develop a trading system based on Support Vector Machines (SVM) in or... more The aim of this paper is to develop a trading system based on Support Vector Machines (SVM) in order to use it in the S&P500 index. The data covers the period between 03/01/2000 and 30/12/2011. The inputs of the SVM are different forecasting algorithms: Relative Strength Index (RSI), Moving Average Convergence Divergence (MACD), Momentum, Bollinger Bands and the Chicago Board Options Exchange Volatility Index (VIX). A SVM Classifier has been used in order to develop the trading system with a weekly forecast. The output of the SVM is the decision making for investors. The trading system works better in bearish movement of the S&P500 than bullish movement of the S&P500.
In this paper, a result for bivariate normal distributions (Sheppard, 1899) from statistics is tr... more In this paper, a result for bivariate normal distributions (Sheppard, 1899) from statistics is transformed into a financial asset context in order to build a tool which could translate a correlation matrix into an equivalent probability matrix and vice versa. This way, the correlation coefficient parameter is more understandable in terms of joint probability of two stocks' returns, and much more useful in terms of the information it provides. We validate, empirically, our result for a sample covering the three market capitalization categories in the S&P 500 index over a ten-year period. Finally, the accuracy of this new tool is measured theoretically and some applications from the practitioners’ point of view are offered. Such applications include, for instance, the calculation of the number of trading days in a year of two stocks with the same sign of returns and how to split the average return of weighted stocks into four orthants.
Despite the ever-growing interest in trend following and a series of publications in academic jou... more Despite the ever-growing interest in trend following and a series of publications in academic journals, there is still a great shortage of theoretical results on the properties of trend following rules. Our paper fills this gap by comparing and contrasting the two most popular trend following rules, the Momentum (MOM) and Moving Average (MA) rules, from a theoretical perspective. Our approach is based on the return-based formulation of trading rules and modelling the price trends by an autoregressive return process. We provide theoretical results on the similarity between various trend following rules and the forecast accuracy of trading rules. Our results show that the similarity between the MOM and MA rules is rather high and increases with increasing trend strength. However, as compared to the MOM rule, the MA rules have a more robust forecast accuracy of the future direction of price trends. As a result, under uncertain market dynamics the MA rules tend to gain an advantage over the MOM rule. Overall, the results reported in this paper help traders to understand more deeply the properties of trend following rules as well as the differences and similarities between them.
In 1993, the Chicago Board of Options Exchange (CBOE) introduced the VXO, a volatility index base... more In 1993, the Chicago Board of Options Exchange (CBOE) introduced the VXO, a volatility index based on implied volatilities on S&P 100 index. In 2003, the CBOE changed their volatility index design and introduced the VIX in order to enhance its economic significance and to facilitate hedging. In this paper, using data from the USA and the German stock markets, we compare the forecasting capability of the volatility indexes with that of historical volatility and conditional volatility models. Following this analysis, we have studied whether it may be the case that volatility indexes forecast the realized volatilities more accurately for a different period to 30 (or 45) days, attempting to answer the question: what time horizon is the informational content of volatility indexes best adjusted for? The optimal prediction period of each volatility index (VXO, VIX, VDAX and V1X) in terms of coefficient of determination is analysed. The results identify a difference between the observed optimal forecasting period and the theoretical one. This could be explained from different perspectives such as the index's design, investor cognitive bias or overreaction.
International Journal of Trade, Economics and Finance, 2013
The goal of this research is to analyse the different results that can be achieved using Support ... more The goal of this research is to analyse the different results that can be achieved using Support Vector Machines to forecast the weekly change movement of the different simulated markets. The data cover 3000 daily close for each simulated market. The main characteristic of these markets are: high volatility, bearish movement, bullish movement and low volatility. The inputs of the SVM are the Relative Strength Index (RSI) and the Moving Average Convergence Divergence (MACD). SVM-KM is used by Matlab in order to design the algorithm. The outputs of the SVM are the degree of set membership and the market movement (bullish or bearish). The configuration for the SVM shows that results are better in high volatility markets or low volatility markets than trend markets.
The aim of this research is to analyse the different results that can be achieved using Support V... more The aim of this research is to analyse the different results that can be achieved using Support Vector Machines (SVM) to forecast the weekly change movement of the different simulated markets. The different simulated markets are developed by a GARCH model based on the S&P 500. These simulated markets are grouped by a main parameter: high volatility, bearish trend, bullish trend and low volatility. The inputs retained of the SVM are traditional technical trading rules used in quantitative analysis such as Relative Strength Index (RSI) and Moving Average Convergence Divergence (MACD) decision rules. The outputs of the SVM are the degree of set membership and the market movement (bullish or bearish). The design of the SVM algorithm has been developed by Matlab and SVM-KM. The configuration for the SVM shows that the best results are achieved in simulated markets with high volatility; also results are good in trend simulated markets.
... MIDIENDO LA VOLATILIDAD DEL MERCADO DE OPCIONES CON EL VIX Javier Giner Rubio y Sandra Morini... more ... MIDIENDO LA VOLATILIDAD DEL MERCADO DE OPCIONES CON EL VIX Javier Giner Rubio y Sandra Morini Marrero ... estudios, destaquemos los trabajos de Rubio y Salvador (1991), Peiró (1994), Martínez-Abascal (1993), Peña (1995) y Corredor y Santamaría (1996). ...
The aim of this research is to analyse the influence of the Chicago Board Options Exchange Market... more The aim of this research is to analyse the influence of the Chicago Board Options Exchange Market Volatility Index (VIX) using Support Vector Machines (SVMs) in order to forecast the weekly change in the S&P 500 index. The data covers the period between 03/01/2000 and 30/12/2011. A trading simulation is implemented so that statistical efficiency is complemented by measures of economic performance. The inputs retained are traditional technical trading rules commonly used in the analysis of equity markets such as Relative Strength Index (RSI), Moving Average Convergence Divergence (MACD), VIX and the daily return of the S&P 500. The SVM determines the best situations to buy or sell the market. The two outputs of the SVM are the movement of the market and the degree of set membership. The influence of VIX in the trading system that it has been developed is really significant when bearish periods appear. VIX allows the reduction of the Maximum DrawDown (MDD) and to increase the profit as it can be seen in the results of the research.
The Journal of the Acoustical Society of America, 2001
ABSTRACT This article is an extension of the procedure to estimate errors in ray-tracing calculat... more ABSTRACT This article is an extension of the procedure to estimate errors in ray-tracing calculations of room response proposed by Giner et al. [J. Acoust. Soc. Am. 106, 816-823 (1999)]. The previous article presented an expression to estimate the error in computing the energy reaching a receptor during a small time interval. This expression was obtained assuming that a pure ray-tracing program is used and a Monte Carlo Technique is applied. In the present work these ideas are extended in order to compute the objective acoustic parameters of a room. The corresponding equations are presented in closed form. Our results show that the number of rays involved in the analysis depends on the nature of the parameters to be evaluated. Some examples are shown in order to validate our conclusions.
The Journal of the Acoustical Society of America, 1999
ABSTRACT The Monte Carlo method is applied to the ray-tracing method to obtain an error measure o... more ABSTRACT The Monte Carlo method is applied to the ray-tracing method to obtain an error measure of the sound energy hitting a receptor during a time interval. Simple geometrical interpretations are introduced to guide the equation development. The variance measurement method introduced a few years ago as an error indicator for the ray-tracing technique is analyzed and compared with the new method. It is shown that the previous method cannot be considered either an error measure or an error estimator, and that the new one provides a consistent answer to many simulation problems. The effects of wall absorption, reverberation time, elapsed time, and integration interval in both formulas are evaluated. (C) 1999 Acoustical Society of America. [S0001-4966(99)06007-5].
In finance, getting an accurate estimation of the term structure of interest rates is essential b... more In finance, getting an accurate estimation of the term structure of interest rates is essential because this information is often used as input by other pricing financial models. In this paper, we point out the importance of selecting a suitable estimation of the term structure of interest rates. To show this fact, we use the Spanish Bond Market to estimate the initial interest rate and forward curves for one day, by using both McCulloch (1975) cubic polynomial splines, and Legendre's polynomials (Morini, 1998). We use these curves as input for pricing pure discount bonds with the Ho and Lee (1986) and Hull and White (1990) models. Then, we find the important result that using an inadequate interest rate curve affects dramatically the behaviour of the dynamic term structure models and, consequently, the estimation of the asset pricing models. JEL Classification: E43.
Support vector machines (SVMs) are a new nonparametric tool for regression estimation. We will us... more Support vector machines (SVMs) are a new nonparametric tool for regression estimation. We will use this tool to estimate the parameters of a GARCH model for predicting the conditional volatility of stock market returns. GARCH models are usually estimated using maximum likelihood (ML) procedures, assuming that the data are normally distributed. In this paper, we will show that GARCH models can be estimated using SVMs and that such estimates have a higher predicting ability than those obtained via common ML methods.
We start this paper by presenting compelling evidence of short-term momentum in the excess return... more We start this paper by presenting compelling evidence of short-term momentum in the excess returns on the S&P Composite stock price index. For the first time ever, we assume that the excess returns follow an autoregressive process of order p, AR(p), and evaluate the parameters of this process. Armed with a fairly accurate knowledge of the momentum generating process, we continue this paper by providing a number of important theoretical implications. First, we present analytical results on the profitability of longonly and long-short time-series momentum (TSMOM) strategies. Our results suggest that the long-only TSMOM strategy is profitable, while the long-short one is not. We find that over multiple periods the risk profile of the long-only TSMOM strategy resembles the risk profile of a portfolio insurance strategy. We estimate the power of the statistical test for superiority of the TSMOM strategy and find that the power is much below the acceptable level. Consequently, any empirical study tends not to reject the null hypothesis of no profitability of TSMOM strategy. Finally, we evaluate the precision of identification of the optimal number of lags in the TSMOM rule using a standard back-testing methodology and find that this precision is extremely poor. However, we demonstrate that the performance of the TSMOM rule is robust to the choice of the number of lags.
The aim of this paper is to develop a trading system based on Support Vector Machines (SVM) in or... more The aim of this paper is to develop a trading system based on Support Vector Machines (SVM) in order to use it in the S&P500 index. The data covers the period between 03/01/2000 and 30/12/2011. The inputs of the SVM are different forecasting algorithms: Relative Strength Index (RSI), Moving Average Convergence Divergence (MACD), Momentum, Bollinger Bands and the Chicago Board Options Exchange Volatility Index (VIX). A SVM Classifier has been used in order to develop the trading system with a weekly forecast. The output of the SVM is the decision making for investors. The trading system works better in bearish movement of the S&P500 than bullish movement of the S&P500.
In this paper, a result for bivariate normal distributions (Sheppard, 1899) from statistics is tr... more In this paper, a result for bivariate normal distributions (Sheppard, 1899) from statistics is transformed into a financial asset context in order to build a tool which could translate a correlation matrix into an equivalent probability matrix and vice versa. This way, the correlation coefficient parameter is more understandable in terms of joint probability of two stocks' returns, and much more useful in terms of the information it provides. We validate, empirically, our result for a sample covering the three market capitalization categories in the S&P 500 index over a ten-year period. Finally, the accuracy of this new tool is measured theoretically and some applications from the practitioners’ point of view are offered. Such applications include, for instance, the calculation of the number of trading days in a year of two stocks with the same sign of returns and how to split the average return of weighted stocks into four orthants.
Despite the ever-growing interest in trend following and a series of publications in academic jou... more Despite the ever-growing interest in trend following and a series of publications in academic journals, there is still a great shortage of theoretical results on the properties of trend following rules. Our paper fills this gap by comparing and contrasting the two most popular trend following rules, the Momentum (MOM) and Moving Average (MA) rules, from a theoretical perspective. Our approach is based on the return-based formulation of trading rules and modelling the price trends by an autoregressive return process. We provide theoretical results on the similarity between various trend following rules and the forecast accuracy of trading rules. Our results show that the similarity between the MOM and MA rules is rather high and increases with increasing trend strength. However, as compared to the MOM rule, the MA rules have a more robust forecast accuracy of the future direction of price trends. As a result, under uncertain market dynamics the MA rules tend to gain an advantage over the MOM rule. Overall, the results reported in this paper help traders to understand more deeply the properties of trend following rules as well as the differences and similarities between them.
In 1993, the Chicago Board of Options Exchange (CBOE) introduced the VXO, a volatility index base... more In 1993, the Chicago Board of Options Exchange (CBOE) introduced the VXO, a volatility index based on implied volatilities on S&P 100 index. In 2003, the CBOE changed their volatility index design and introduced the VIX in order to enhance its economic significance and to facilitate hedging. In this paper, using data from the USA and the German stock markets, we compare the forecasting capability of the volatility indexes with that of historical volatility and conditional volatility models. Following this analysis, we have studied whether it may be the case that volatility indexes forecast the realized volatilities more accurately for a different period to 30 (or 45) days, attempting to answer the question: what time horizon is the informational content of volatility indexes best adjusted for? The optimal prediction period of each volatility index (VXO, VIX, VDAX and V1X) in terms of coefficient of determination is analysed. The results identify a difference between the observed optimal forecasting period and the theoretical one. This could be explained from different perspectives such as the index's design, investor cognitive bias or overreaction.
International Journal of Trade, Economics and Finance, 2013
The goal of this research is to analyse the different results that can be achieved using Support ... more The goal of this research is to analyse the different results that can be achieved using Support Vector Machines to forecast the weekly change movement of the different simulated markets. The data cover 3000 daily close for each simulated market. The main characteristic of these markets are: high volatility, bearish movement, bullish movement and low volatility. The inputs of the SVM are the Relative Strength Index (RSI) and the Moving Average Convergence Divergence (MACD). SVM-KM is used by Matlab in order to design the algorithm. The outputs of the SVM are the degree of set membership and the market movement (bullish or bearish). The configuration for the SVM shows that results are better in high volatility markets or low volatility markets than trend markets.
The aim of this research is to analyse the different results that can be achieved using Support V... more The aim of this research is to analyse the different results that can be achieved using Support Vector Machines (SVM) to forecast the weekly change movement of the different simulated markets. The different simulated markets are developed by a GARCH model based on the S&P 500. These simulated markets are grouped by a main parameter: high volatility, bearish trend, bullish trend and low volatility. The inputs retained of the SVM are traditional technical trading rules used in quantitative analysis such as Relative Strength Index (RSI) and Moving Average Convergence Divergence (MACD) decision rules. The outputs of the SVM are the degree of set membership and the market movement (bullish or bearish). The design of the SVM algorithm has been developed by Matlab and SVM-KM. The configuration for the SVM shows that the best results are achieved in simulated markets with high volatility; also results are good in trend simulated markets.
... MIDIENDO LA VOLATILIDAD DEL MERCADO DE OPCIONES CON EL VIX Javier Giner Rubio y Sandra Morini... more ... MIDIENDO LA VOLATILIDAD DEL MERCADO DE OPCIONES CON EL VIX Javier Giner Rubio y Sandra Morini Marrero ... estudios, destaquemos los trabajos de Rubio y Salvador (1991), Peiró (1994), Martínez-Abascal (1993), Peña (1995) y Corredor y Santamaría (1996). ...
The aim of this research is to analyse the influence of the Chicago Board Options Exchange Market... more The aim of this research is to analyse the influence of the Chicago Board Options Exchange Market Volatility Index (VIX) using Support Vector Machines (SVMs) in order to forecast the weekly change in the S&P 500 index. The data covers the period between 03/01/2000 and 30/12/2011. A trading simulation is implemented so that statistical efficiency is complemented by measures of economic performance. The inputs retained are traditional technical trading rules commonly used in the analysis of equity markets such as Relative Strength Index (RSI), Moving Average Convergence Divergence (MACD), VIX and the daily return of the S&P 500. The SVM determines the best situations to buy or sell the market. The two outputs of the SVM are the movement of the market and the degree of set membership. The influence of VIX in the trading system that it has been developed is really significant when bearish periods appear. VIX allows the reduction of the Maximum DrawDown (MDD) and to increase the profit as it can be seen in the results of the research.
The Journal of the Acoustical Society of America, 2001
ABSTRACT This article is an extension of the procedure to estimate errors in ray-tracing calculat... more ABSTRACT This article is an extension of the procedure to estimate errors in ray-tracing calculations of room response proposed by Giner et al. [J. Acoust. Soc. Am. 106, 816-823 (1999)]. The previous article presented an expression to estimate the error in computing the energy reaching a receptor during a small time interval. This expression was obtained assuming that a pure ray-tracing program is used and a Monte Carlo Technique is applied. In the present work these ideas are extended in order to compute the objective acoustic parameters of a room. The corresponding equations are presented in closed form. Our results show that the number of rays involved in the analysis depends on the nature of the parameters to be evaluated. Some examples are shown in order to validate our conclusions.
The Journal of the Acoustical Society of America, 1999
ABSTRACT The Monte Carlo method is applied to the ray-tracing method to obtain an error measure o... more ABSTRACT The Monte Carlo method is applied to the ray-tracing method to obtain an error measure of the sound energy hitting a receptor during a time interval. Simple geometrical interpretations are introduced to guide the equation development. The variance measurement method introduced a few years ago as an error indicator for the ray-tracing technique is analyzed and compared with the new method. It is shown that the previous method cannot be considered either an error measure or an error estimator, and that the new one provides a consistent answer to many simulation problems. The effects of wall absorption, reverberation time, elapsed time, and integration interval in both formulas are evaluated. (C) 1999 Acoustical Society of America. [S0001-4966(99)06007-5].
In finance, getting an accurate estimation of the term structure of interest rates is essential b... more In finance, getting an accurate estimation of the term structure of interest rates is essential because this information is often used as input by other pricing financial models. In this paper, we point out the importance of selecting a suitable estimation of the term structure of interest rates. To show this fact, we use the Spanish Bond Market to estimate the initial interest rate and forward curves for one day, by using both McCulloch (1975) cubic polynomial splines, and Legendre's polynomials (Morini, 1998). We use these curves as input for pricing pure discount bonds with the Ho and Lee (1986) and Hull and White (1990) models. Then, we find the important result that using an inadequate interest rate curve affects dramatically the behaviour of the dynamic term structure models and, consequently, the estimation of the asset pricing models. JEL Classification: E43.
Support vector machines (SVMs) are a new nonparametric tool for regression estimation. We will us... more Support vector machines (SVMs) are a new nonparametric tool for regression estimation. We will use this tool to estimate the parameters of a GARCH model for predicting the conditional volatility of stock market returns. GARCH models are usually estimated using maximum likelihood (ML) procedures, assuming that the data are normally distributed. In this paper, we will show that GARCH models can be estimated using SVMs and that such estimates have a higher predicting ability than those obtained via common ML methods.
We start this paper by presenting compelling evidence of short-term momentum in the excess return... more We start this paper by presenting compelling evidence of short-term momentum in the excess returns on the S&P Composite stock price index. For the first time ever, we assume that the excess returns follow an autoregressive process of order p, AR(p), and evaluate the parameters of this process. Armed with a fairly accurate knowledge of the momentum generating process, we continue this paper by providing a number of important theoretical implications. First, we present analytical results on the profitability of longonly and long-short time-series momentum (TSMOM) strategies. Our results suggest that the long-only TSMOM strategy is profitable, while the long-short one is not. We find that over multiple periods the risk profile of the long-only TSMOM strategy resembles the risk profile of a portfolio insurance strategy. We estimate the power of the statistical test for superiority of the TSMOM strategy and find that the power is much below the acceptable level. Consequently, any empirical study tends not to reject the null hypothesis of no profitability of TSMOM strategy. Finally, we evaluate the precision of identification of the optimal number of lags in the TSMOM rule using a standard back-testing methodology and find that this precision is extremely poor. However, we demonstrate that the performance of the TSMOM rule is robust to the choice of the number of lags.
Uploads
Papers by Javier Giner