| 摘要 |
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for predictive modeling for unintended outcomes are disclosed. In one aspect, a method includes the actions of accessing an order history that, for each of one or more past orders, indicates (i) one or more order details associated with the order, and (ii) a fulfillment outcome associated with the order. The actions further include selecting one or more particular past orders that are associated with the particular unintended order fulfillment outcome. The actions further include generating a predictive model. The actions further include receiving one or more order details associated with a subsequently received order. The actions further include providing the one or more order details as input to the predictive model. The actions further include, identifying a remedial action. The actions further include providing data indicating the remedial action. |
| 主权项 |
1. A computer-implemented method comprising:
accessing an order history that, for each of one or more past orders, indicates (i) one or more order details associated with the order, and (ii) a fulfillment outcome associated with the order, wherein the order fulfillment outcome is selected from among multiple pre-defined order fulfillment outcomes including one or more intended order fulfillment outcomes and one or more unintended order fulfillment outcomes; for each of one or more particular unintended order fulfillment outcomes, selecting one or more particular past orders that are associated with the particular unintended order fulfillment outcome; for each of the one or more particular unintended order fulfillment outcomes, generating, using the one or more order details associated with the one or more particular past orders that are associated with the particular unintended order fulfillment outcome as training data, a predictive model that is trained to estimate, based on one or more given order details associated with a given order, a likelihood that the given order will be associated with the particular unintended order fulfillment outcome; receiving one or more order details associated with a subsequently received order; for each of the predictive models, providing the one or more order details as input to the predictive model; in response to providing the one or more order details as input to the predictive model, receiving, from each of the predictive models, a likelihood that the one or more order details will be associated with the particular unintended order fulfillment outcome; determining, for each indication of the likelihood, whether the likelihood satisfies a threshold; for likelihoods that satisfy the threshold, identifying a remedial action that, when implemented, adjusts the likelihood to not satisfy the threshold; and providing, for output, data indicating the remedial action. |
