SYSTEM AND METHOD FOR DETERMINING A TRANSIT PREDICTION MODEL
In variants, a method for predicting transit data can include, determining a set of models, training each model, determining package transit data, evaluating the set of models, selecting a model from the set of models, predicting package transit data and/or any other suitable element. In variants, the method can function to determine, select, and/or train one or more models to predict package transit (e.g., physical package delivery to a destination).
1 . A method, comprising:
receiving a request for a package associated with a prediction period;
selecting a model from a set of models trained for the prediction period, wherein each model is trained on a different set of historical transit data, and wherein the set of models comprises at least one machine learning model; and
predicting transit data for the package using the selected model.
2 . The method of claim 1 , wherein the selected model is a neural network.
3 . The method of claim 1 , further comprising:
initiating a shipment of the package; and
upon receipt of a tracking detail associated with the package, updating the predicted transit data.
4 . The method of claim 3 , wherein the tracking detail is determined based on scan event information associated with a shipping label printed for the package.
5 . The method of claim 3 , wherein the transit data is predicted based on shipment data comprising a shipping origin, wherein updating the predicted transit data comprises:
updating the shipment data, comprising updating the shipping origin to a location associated with the tracking detail; and
predicting updated transit data based on the updated shipment data.
6 . The method of claim 1 , wherein selecting the model and predicting the transit data is repeated for each leg of a multi-leg route.
7 . The method of claim 1 , wherein predicting transit data for the package using the selected model comprises predicting a delivery probability for each of a set of predetermined transit times.
8 . The method of claim 7 , further comprising determining a delivery probability for a potential transit time window, encompassing multiple transit times from the set of predetermined transit times, based on the respective delivery probabilities of the multiple transit times.
9 . The method of claim 1 , wherein the predicted transit data comprises at least one of: shipment transit time or package arrival time.
10 . The method of claim 1 , wherein the predicted transit data comprises at least one of: a shipping delay or package damage.
11 . The method of claim 1 , wherein the package is associated with a designated carrier, wherein the selected model is trained on historical transit data for packages shipped by the designated carrier.
12 . A system, comprising:
an interface configured to receive a request for a package; and
a processing system configured to:
a) select a model from a set of machine learning models, each trained on a different set of historic transit data; and
b) predict transit data for the package using the selected model.
13 . The system of claim 12 , wherein the selected model is a neural network.
14 . The system of claim 12 , wherein a) is repeated for each of a set of prediction periods.
15 . The system of claim 14 , wherein b) is performed for packages associated with requests received within a prediction period of the set of prediction periods, using the model selected for the respective prediction period.
16 . The system of claim 14 , wherein the processing system is further configured to, for each prediction period in the set of prediction periods, repeat b) for each undelivered package using the respective model.
17 . The system of claim 12 , wherein the processing system is further configured to:
initiate a shipment of the package;
receive a tracking detail associate with the shipment; and
predict updated transit data based on the tracking detail.
18 . A system of claim 12 , wherein the predicted transit data comprises a delivery probability for each of a set of predetermined transit times.
19 . The system of claim 18 , wherein the processing system is further configured to:
determine a delivery probability for each of a set of transit time windows, wherein each transit time window encompasses multiple transit times from the set of predetermined transit times, wherein the delivery probability for a transit time window is determined from the delivery probabilities for the encompassed transit times; and
return a transit time window associated with a delivery probability exceeding a threshold delivery probability.
20 . The system of claim 12 , wherein a) and b) are repeated for each leg of a multi-leg route.