0

I want to determine an appropriate model for daily stock index having the graph given below. It could be seen that the data has a trend but does it have seasonality too. If so what model is it i.e (additive or multiplicative ) and what would be the frequency of seasonality?

I ran a periodogram and it showed a spike only at 0. Also, its ACFs are all positive and is gradually decreasing.

Darren Tsai
  • 32,117
  • 5
  • 21
  • 51
Jor_El
  • 187
  • 8

1 Answers1

0

A very quick and easy way to handle time series is with Facebook's package called Prophet. Simply feed in your daily data in correct format, and it will make predictions. You don't need to feed any ACF/seasonality info...it should find it automatically for you.

https://cran.r-project.org/web/packages/prophet/index.html

You can also consider options like recurrent neural network using package Keras. Or consider ARIMA functions like auto.arima. I have had good results using algorithm called TBATS as well. Just try them all and choose the one that fits the best.

Corey Levinson
  • 1,553
  • 17
  • 25