An Intelligent Market Model that Incorporates Individual Trader’s Behavior to Simulate Forex Trading Data

A speculator in the foreign exchange (Forex) market interacts in an environment where other participants as well as their strategies are unobserved. This constitutes a challenging interaction environment that generates complex time series data that may be difficult to predict. Technical analysts’ use of only historic Forex data, disregarding individuals’ trading experience, could be limiting in market prediction. Individuals’ trading experience data is protected. The historic Forex data in combination with individuals’ trading experience can constitute a knowledge base for training models to support Forex price prediction. We use game-theoretic analysis to build a model that considers individual trader traits and generates Forex data. Exponential moving average and magnitude of price deviations from open value for clusters of regular time intervals are compared for both real and generated data. The results show similarity and consistent behavior for the two sets of data with respect to trend, volatility clustering, seasonality and periodicity. Furthermore, stationarity and temporal structure analysis using the augmented Dickey–Fuller (ADF) test and the autocorrelation function (ACF) yielded meaningful results, indicating comparable patterns in both datasets. Besides, model results demonstrate that traders in the Forex market operate in an information-skewed environment. Further experiments are required to refine the model data, optimize the intelligent market score and subsequently apply the results in Forex market analysis.