meocong / HiddenMarkovModelPredictingStock

Predicting stock price by using Hidden Markov model

HiddenMarkovModelPredictingStock

Predicting stock price by using Hidden Markov model

Requirement

• Python
• Library: numpy, Ipython notebook, matplotlib, pandas_datareader, pandas

Instruction

• HMM_predict_by_open_today.py: Hidden Markov Model using open price of a day to predict close price in days Options: model = ModelHMM(company="AAPL", day_start=day_start, day_end=day_end, n_days_previous=200, n_states=10, verbose=True, n_decimals = 3, latex = True)

  • company : Stock name of the company
  • day_start : Starting day
  • day_end : Ending day
  • n_days_previous : Number of previous days will be used for training
  • n_states : Number of hidden states
  • verbose : Printing when running or not
  • n_decimals : Number of decimals of double
  • latex : Printing matrix in latex type or not

• HMM_predict_continous_days.py: Hidden Markov Model using close price of a day to predict close price n_days_predict days after model = ModelHMM(company="AAPL", day_start=day_start, day_end=day_end, n_days_previous=100, n_states=10, n_days_predict=2, verbose=True, n_decimals = 3, latex = False)

  • n_days_predict : Number of following days will be predict using close price of day