| 1: Input Parameters:
|
| 2: Data: a matrix where each column represents a time series variable. |
| 3: Lag: define the Lag order. |
| 4: num: represents the number of variables. |
| 5: obs: represents the number of observations. STEP-1
|
| 6: for i to num do
|
| 7: for j to Lag do
|
| 8: Coefficient-matrix[i][j] = initial-weights |
| 9: end for
|
| 10: end for
|
|
STEP-2
|
| 11: for i to num do
|
| 12: for j to Lag do
|
| 13: X = construct Lag_matrix(data[i], Lag) |
| 14: Y = data[i][Lag+1 : obs] |
| 15: end for
|
| 16: Coefficient-matrix[i][j] = solve-coefficient (X, Y) |
| 17: end for
|
|
STEP-3
|
| 18: for i to num do
|
| 19: for t = obs+1 to obs+Forcaststep do
|
| 20: Popularity[i][t] = PredictPopularity (Coefficient-matrix[i],data[i][t-lag : t-1] |
| 21: end for
|
| 22: end for
|
|
STEP-4
|
| 23: for i to num do
|
| 24: Calculate RMSE for data[i][obs+1:obs +Forcaststep] and Popularity[i] |
| 25: end for
|
