I have time series data and the goal is to predict Saturday's value based on the values of the remaining days of the week. The trick is the value have seasonality with start/end or mid of the month. So, essentially trying to formulate a function which will look like
sat_value = constant + (Mon_value * Mon_coeff) + (Tues_value * Tues_coeff)...... + start/mid/end period coeff
I tried VAR and got the coefficients but one of them is share itself. Also, I'm not sure how to formulate the above equation through VAR. So, not sure if VAR or regression would be helpful over here.
The end objective is to formulate this in excel as a function where user's can enter Sun through Friday value and then Sat value is predicted (or calculated) based on these.
Any recommendation on how to approach this? I thought about regression with keeping Sat value as dependent and remaining variables to predict but then not sure if it's a good idea with time series data.
Date Share Dy_fm_mth is_mon is_tues is_wed ..... is_sat is_mid_mth is_st_end_mth 3/01 0.5 0 1 0 0 0 0 1 3/02 0.3 1 0 1 0 0 0 0 3/03 0.6 2 0 0 1 0 0 0 3/04 0.7 3 0 0 0 0 0 0
Appreciate the help and recommendations.
https://stackoverflow.com/questions/66978637/forecasting-for-weekend-based-on-weekdays-values-in-a-multivariate-time-series April 07, 2021 at 10:04AM
没有评论:
发表评论