Machine-based prediction of visitation caused by viewing
AssigneeBillips, Inc.
InventorShawn Spooner
Life of this patent
Estimating the term…
Abstract
For machine-based prediction of visitation, a graph of paired mobile devices is formed where the edges for each pair are based on similarities in visitation and/or metadata for the devices. A machine-learned network embeds the visitation and metadata information, which is used to indicate the similarity. Since the trace data used to show access may be sparse, another machine-learned network completes the route based on routes used by similar devices. Another machine-learned network recommends effectiveness of content based on routes, the graph, metadata, and/or other information. The recommendation is based on training using counterfactual and/or other causal modeling.
Claims20
A method for machine-based prediction of visitation, the method comprising:
- acquiring first trace data showing corresponding routes previously traversed by devices over time in a region, the first trace data including unobserved portions of the corresponding routes for at least some of the devices;
- forming, by a second machine-learned network, a first graph linking pairs of the devices with edges assigned a similarity between the pairs of the devices;
- generating, by a first machine-learned network trained using deep learning, second trace data that fills-in the unobserved portions of the corresponding routes previously traversed by the at least some of the devices;
- determining, using the first trace data and/or the second trace data, whether an opportunity for obtaining first content occurred based on the corresponding routes previously traversed by the at least some of the devices;
- modeling, by a third machine-learned network, a causal effect of the first content to visits of a location from the first graph;
- and
- displaying a prediction of effectiveness of second content based on the causal effect.
A product including:
- non-transitory machine-readable media;
- and
- instructions stored on the machine-readable media, the instructions configured to, when executed, cause a processor to:
- acquire first trace data showing corresponding routes previously traversed by devices over time in a region, the first trace data including unobserved portions of the corresponding routes for at least some of the devices;
- form, by a second machine-learned network, a first graph linking pairs of the devices with edges assigned a similarity between the pairs of the devices;
- generate, by a first machine-learned network trained using deep learning, second trace data that fills-in the unobserved portions of the corresponding routes previously traversed by the at least some of the devices;
- determine, using the first trace data and/or the second trace data, whether an opportunity for obtaining first content occurred based on the corresponding routes previously traversed by the at least some of the devices;
- model, by a third machine-learned network, a causal effect of first content to visits of a location from the first graph;
- and
- display a prediction of effectiveness of second content based on the causal effect.
A system including:
- a processor;
- and
- memory including instructions configured to, when executed, cause the processor to:
- acquire first trace data showing corresponding routes previously traversed by devices over time in a region, the first trace data including unobserved portions of the corresponding routes for at least some of the devices;
- form, by a second machine-learned network, a first graph linking pairs of the devices with edges assigned a similarity between the pairs of the devices;
- generate, by a first machine-learned network trained using deep learning, second trace data that fills-in the unobserved portions of the corresponding routes previously traversed by the at least some of the devices;
- determine, using the first trace data and/or the second trace data, whether an opportunity for obtaining first content occurred based on the corresponding routes previously traversed by the at least some of the devices;
- model, by a third machine-learned network, a causal effect of first content to visits of a location from the first graph;
- and
- display a prediction of effectiveness of second content based on the causal effect.
Dependencies are inferred from the claim text.