Probabilistic inventory estimation in grocery

1 . A probabilistic inventory estimation system, comprising:

one or more processors configured to:

receive a set of records pertaining to a perishable item, comprising:

shipment data pertaining to the perishable item;

transaction data pertaining to the perishable item; and

at least one user inventory observation pertaining to the perishable item;

determine a probabilistic estimate of inventory of the perishable item based at least in part on the shipment data, the transaction data, and the at least one user inventory observation pertaining to the perishable item, wherein determining the probabilistic estimate of inventory of the perishable item comprises simulating inventory trajectories of the perishable item over time at least in part by sampling shelf-life estimates of the perishable item; and

recommend an amount of the perishable item to order based at least in part on the probabilistic estimate of inventory of the perishable item; and

a memory coupled to the one or more processors and configured to provide the one or more processors with instructions.

2 . The probabilistic inventory estimation system of claim 1 , wherein determining the probabilistic estimate of inventory of the perishable item comprises determining uncertainty associated with the probabilistic estimate of inventory of the perishable item.

3 . The probabilistic inventory estimation system of claim 1 , wherein the probabilistic estimate of inventory of the perishable item is determined at least in part by performing particle filtering.

4 . The probabilistic inventory estimation system of claim 3 , wherein performing the particle filtering includes performing resampling based on the at least one user inventory observation.

5 . The probabilistic inventory estimation system of claim 4 , wherein performing the resampling includes upweighting or down-weighting a particle based at least in part on the at least one user inventory observation.

6 . The probabilistic inventory estimation system of claim 1 , wherein an expiration date of a shipment is determined based at least in part on a sampled shelf-life estimate of the perishable item.

7 . The probabilistic inventory estimation system of claim 6 , wherein the probabilistic estimate of inventory of the perishable item is determined at least in part by using the expiration date of the shipment determined based at least in part on the sampled shelf-life estimate of the perishable item and a shrink fraction associated with the shipment.

8 . A method, comprising:

receiving a set of records pertaining to a perishable item, comprising:

shipment data pertaining to the perishable item;

transaction data pertaining to the perishable item; and

at least one user inventory observation pertaining to the perishable item;

determining a probabilistic estimate of inventory of the perishable item based at least in part on the shipment data, the transaction data, and the at least one user inventory observation pertaining to the perishable item, wherein determining the probabilistic estimate of inventory of the perishable item comprises simulating inventory trajectories of the perishable item over time at least in part by sampling shelf-life estimates of the perishable item; and

recommending an amount of the perishable item to order based at least in part on the probabilistic estimate of inventory of the perishable item.

9 . The method of claim 8 , wherein determining the probabilistic estimate of inventory of the perishable item comprises determining uncertainty associated with the probabilistic estimate of inventory of the perishable item.

10 . The method of claim 8 , wherein the probabilistic estimate of inventory of the perishable item is determined at least in part by performing particle filtering.

11 . The method of claim 10 , wherein performing the particle filtering includes performing resampling based on the at least one user inventory observation.

12 . The method of claim 11 , wherein performing the resampling includes upweighting or down-weighting a particle based at least in part on the at least one user inventory observation.

13 . The method of claim 8 , wherein an expiration date of a shipment is determined based at least in part on a sampled shelf-life estimate of the perishable item.

14 . The method of claim 13 , wherein the probabilistic estimate of inventory of the perishable item is determined at least in part by using the expiration date of the shipment determined based at least in part on the sampled shelf-life estimate of the perishable item and a shrink fraction associated with the shipment.

15 . A computer program product embodied in a non-transitory computer readable storage medium and comprising computer instructions for:

receiving a set of records pertaining to a perishable item, comprising:

shipment data pertaining to the perishable item;

transaction data pertaining to the perishable item; and

at least one user inventory observation pertaining to the perishable item;

determining a probabilistic estimate of inventory of the perishable item based at least in part on the shipment data, the transaction data, and the at least one user inventory observation pertaining to the perishable item, wherein determining the probabilistic estimate of inventory of the perishable item comprises simulating inventory trajectories of the perishable item over time at least in part by sampling shelf-life estimates of the perishable item; and

recommending an amount of the perishable item to order based at least in part on the probabilistic estimate of inventory of the perishable item.