Log Periodic Power Law Singularity model for financial bubble predictions. We use the Log Periodic Power Law Singularity model to build a time signal that predicts the existence of a financial bubble in the stock market.
This work is based on the research done at the Financial Crisis Observatory at ETH Zurich.
To use this code, make sure you have installed: pandas, datetime, numpy, multiprocessing, time, sklearn, random, cma, pandas_datareader.
To execute the code, define the stock index, time range, and number of fitting windows for which you want the signal to be computed in the __main__.py and then run python -m lppls. If you wish to change the qualifying conditions for the fits, go to signal/conditions_satisfied.csv and modify accordingly.
The output of the code is stored in sig.csv. It can be visualised with the python notebook visulalise.ipynb.