Skip to content

Predict stock market pricing over 180 minutes using Black-Scholes stochastic modeling and parallel Monte-Carlo simulations.

License

Notifications You must be signed in to change notification settings

RajdeepKonwar/stockast

Repository files navigation

Stockast FOSSA Status

Stock Market Forecasting using Parallel (OpenMP) Monte-Carlo Simulations

alt text

Compile Instructions

Windows

  • Open stockast.sln
  • [Optional] Right-click Solution 'stockast' in the Solution Explorer and select Retarget solution
  • Build and run!

Linux

make

Type make clean to clean object file and executable.

Run Instructions

Windows

Simply run from Visual Studio or double-click the executable created inside x64\{config}\stockast.exe

By default, the program will try and utilize the maximum system threads available. In order to use a specific number of threads, set the environment vairable OMP_NUM_THREADS equal to the number of threads you want.

Linux

Set the number of threads to be used for computation,

export OMP_NUM_THREADS=number_of_threads

For example, export OMP_NUM_THREADS=8. Then run the program

./stockast

image

General info

  • The input file "data.csv" contains the stock-price values for 3 hours prior to run-time; this acts as the history-data and helps estimate the market volatility.
  • The output file "opt.csv" contains the output (most likely outcome) price-vector from our code. One can use Excel or gnuplot to plot the resulting line graph of the predicted stock pricing.
  • (Linux only) The script "profiling.sh" runs the parallel code from 1 to the specified number of threads. To use the script,
./profiling.sh "number_of_threads"

For example, ./profiling.sh 8.