Repo for Intern
This is an experimental project that uses ARIMA, wavelet and XGBoost to predict stock ups and downs I cannot make sure that the model can be run properly under any circumstances. (Actually even me myself having trouble with debugging...;p)
Censored some details due to some of them might have secret issues.
close_fft.py: Considering to change its name in future because the purpose of this file is beyond calculating fft.
close_ARIMA.py: modelling for ARIMA and wavelet including param estimations, difference, and wavelet trans of time-series.
parser_csv.py: simple version to read csv files and transform data into pandas.DataFrame
XGBoost_stock.py: The XGBoost model for extracting features and predictions.