Compute three-day predicted usage, waste, remaining and shortage using the actual usage and predicted usage.
compute_prediction_statistics(y, t_pred, initial_expiry_data = c(0, 0), initial_collection_data = c(60, 60, 60), start = 10)
y | is the number of units used at the current day i |
---|---|
t_pred | is the sum of the predicted number of units for days i+1, i+2, i+3 |
initial_expiry_data | the number of units that can be used the current day and the next (2-length vector) |
initial_collection_data | is the number of units to collect for days i, i+1, i+2 |
start | is when we start the clock. So for example, in this default invocation the shelf storage (initial_collection_data) was 60, 60, 60, on days 10, 11, and 12. So the evaluation of the model begins at day 11 but the decision to order collection starts on day 13. |
a list with four components, x
is the number of
units to collect, r
is a matrix of two columns
indicating remaining units usable on day i+1 and day i+2,
w
is waste, and s
is shortage.