I am trying to build a SARIMAX model for time series forecasting. I am not able to figure out what the period of seasonality should be. My data is daily data (exculding weekends) of stock prices for past 5 years. I have resampled it to weekly data.
the seasonal_decompose shows seasonality with different seasonal period values - 12,52
i am not able to find out what shall be the value of period in the following code
decomposition = seasonal_decompose(data['Close'], model='additive', period=?)
decomposition.plot()
plt.show()
i also did the ADFuller test and found that the number of lags is 14 and hence decided to go with period=14 in the seasonal_decompose
is this the right approach? if not, then what?
also, once i have decided what the seasonal period will be, suppose 14, then 14 will also be used while differencing, right?
for ex - will 14 be used here?
data['seasonal'] = data['Close'] - data['Close'].shift(14)
please help,
Thank you
i went with period = 14 because that was the number of lags while doing the ADfuller test
