We find that defining 3-dimensional matrix norms, summarizing large financial data set and quantifying market impression
with respect to several variables together are useful. We obtain a proxy for time evolution of market impression value,
and perform simulations for various model parameters. We see that using the volatility in terms of extreme values makes it
easy to evaluate volatilities when we perform simulations with the large real data set. According to the simulation results,
the Heston stochastic volatility model cannot promise much things at the long time interval since the initial information
lose their effect by the time.
It is challenging to compare the advantages and limitations of the higher order SDE-solvers. We present that there is inverse
relationship between speed of convergence of the methods and impression matrix norm (IMN) values while using the
annual volatilities having extreme values as in the line of literature. When we examine the numerical methods using IMN
in terms of their trade-offs, although SRK takes more time, it worths to prefer SRK to Euler–Maruyama method and Milstein
method just because that SRK’s lower cumulative error is important for our financial applications. On the other hand, it does
not matter which method is used for the applications having relatively low volatility cases with respect to the robustness.
We suggest Euler–Maruyama method because of its lowest cost for daily volatility usage or market situations at such low
volatility levels.