Keywords
Intervalo aleatorio Modelo de camino aleatorio Medidas de exactitud FOREX Minimos y máximos Redes neuronales, Accuracy measures FOREX Lows and highs Neural networks Random interval Random walk model
. Forecasting exchange rates with the iMLP: New empirical insight on one multi-layer perceptron for interval time series (ITS).
Abstract
El mercado de divisas (FX) es un mercado para convertir la moneda de un país en la de otro país. Los movimientos de los tipos de cambio al contado se observan cuidadosamente cada minuto en todo el mundo. Pero los gobiernos, los bancos y las empresas multinacionales generalmente desarrollan la toma de decisiones bajo otras frecuencias de tiempo como días, semanas, meses, trimestres o semestres. Las series de tiempo de intervalo (ITS) asignan un intervalo de valores en cada período de tiempo. Por ejemplo, los valores máximos y mínimos diarios o mensuales son ejemplos clave de ITS. Se han desarrollado varios métodos de pronóstico para ITS. Las redes neuronales han atraído investigaciones centradas en el pronóstico de divisas. El perceptrón multicapa (MLP) con una capa oculta es una de las mejores redes para pronosticar series de tiempo nítidas. Para ITS, el iMLP (intervalo MLP) se propuso como una extensión del MLP. El número y tipo de entradas y el número de neuronas en la capa oculta (15 es el número habitual) son parámetros clave para clasificar las diferentes arquitecturas de la red. Analizamos estos hiperparámetros en el desempeño de pronóstico del iMLP a través del EUR / USD en una base diaria de bajo a alto, en diferentes comportamientos como tendencia alcista, tendencia bajista o lateral; sobre diferentes medidas de precisión, incluidas las tasas de cobertura y eficiencia, e incorporando otras tasas como GBP / USD o AUD / USD. Se descarta la elección de 15 neuronas. Además, comparamos estas redes iMLP con el intervalo de caminata aleatoria y los resultados son bastante prometedores. Finalmente, concluimos que en cualquier contexto de FX, se deben considerar varias redes iMLP, lo que abre nuevas vías de investigación.
The foreign exchange market (FX) is a market for converting the currency of one country into that of another country. Spot exchange rates movements are carefully observed every minute around the world. But governments, banks and multinational companies generally develop decision-making under other frequencies of time like days, weeks, months, quarters, or semesters. Interval time series (ITS) assign an interval of values at every period of time. For example, daily or monthly lows and highs values are key examples of ITS. Several forecasting methods have been developed for ITS. Neural networks have attracted research focused on FX forecasting. The Multi-Layer Perceptron (MLP) with one hidden layer is one of the best networks for forecasting crisp time series. For ITS, the iMLP (interval MLP) was proposed as an extension of the MLP. The number and type of inputs and the number of neurons in the hidden layer (15 is the usual number) are key parameters to rank different architectures of the network. We analyze these hyperparameters in the forecasting performance of the iMLP through the EUR/USD on a low–high daily basis, on different behaviors such as uptrend, downtrend, or sideways; on different accuracy measures, including coverage and efficiency rates, and incorporating other rates such as GBP/USD or AUD/USD. The election of 15 neurons is discarded. Moreover, we compare these iMLP networks with the interval random walk and results are quite promising. Finally, we conclude that in any context of FX, several iMLP networks should be considered which opens new research avenues.
The foreign exchange market (FX) is a market for converting the currency of one country into that of another country. Spot exchange rates movements are carefully observed every minute around the world. But governments, banks and multinational companies generally develop decision-making under other frequencies of time like days, weeks, months, quarters, or semesters. Interval time series (ITS) assign an interval of values at every period of time. For example, daily or monthly lows and highs values are key examples of ITS. Several forecasting methods have been developed for ITS. Neural networks have attracted research focused on FX forecasting. The Multi-Layer Perceptron (MLP) with one hidden layer is one of the best networks for forecasting crisp time series. For ITS, the iMLP (interval MLP) was proposed as an extension of the MLP. The number and type of inputs and the number of neurons in the hidden layer (15 is the usual number) are key parameters to rank different architectures of the network. We analyze these hyperparameters in the forecasting performance of the iMLP through the EUR/USD on a low–high daily basis, on different behaviors such as uptrend, downtrend, or sideways; on different accuracy measures, including coverage and efficiency rates, and incorporating other rates such as GBP/USD or AUD/USD. The election of 15 neurons is discarded. Moreover, we compare these iMLP networks with the interval random walk and results are quite promising. Finally, we conclude that in any context of FX, several iMLP networks should be considered which opens new research avenues.
Información detallada
Materias, derechos, colecciones e identificadores