Hail frequency predictions using artificial intelligence
The disclosure includes systems and methods for identifying, using one or more processors, a plurality of properties, including the first set of properties that experienced hail damage and the second set of properties that did not experience hail damage; determining, using the one or more processors, a location and the first set of features for each property in the plurality of properties; training, using the one or more processors, a first damage frequency model; and validating, using the one or more processors, the first damage frequency model.
1. A computer implemented method comprising:
identifying, using one or more processors, a plurality of properties including a first set of properties that experienced hail damage and a second set of properties that did not experience hail damage;
determining, using the one or more processors, a location and a first set of features for each property in the plurality of properties;
wherein determining the first set of features for each property in the plurality of properties includes performing feature selection, the feature selection including repeatedly:
removing a different set of one or more features from a set of available features to create a reduced set of features;
training a new damage frequency model;
validating the new damage frequency model; and
comparing performance of the new damage frequency model to other instances of damage frequency model which used different feature sets,
wherein the first set of features is a reduced set of features.
2. The computer implemented method of claim 1 further comprising:
receiving a location;
applying a first damage frequency model trained and validated based on the first set of features based on the performance of the first damage model compared to the other instances of the damage frequency model; and
presenting a damage frequency.
3. The computer implemented method of claim 1 , wherein the first set of features comprises a building area, a vegetation density, a roof material, a roof quality, a temperature, a precipitation type or metric, an elevation, an average hail frequency, and a mean hail size.
4. The computer implemented method of claim 1 , wherein the first set of features includes one or more of a building area, a vegetation density, a roof material, a roof quality, a roof height, a roof pitch, presence of skylights, a portion of a roof covered by skylights, presence of a solar panel, a portion of the roof covered by solar panels, a number of roof facets, a roof shape, a land cover code, a temperature, a precipitation type or metric, and an elevation.
5. The computer implemented method of claim 1 , wherein the first set of properties includes properties associated with first claims data representing claims for hail damage to the first set of properties, and the second set of properties includes properties associated with second claims data representing claims for damage to the second set of properties unrelated to hail.
6. The computer implemented method of claim 1 , wherein the first set of properties includes:
properties associated with building permits for roof repair, or roof replacement, and indicating hail as a reason for repair or replacement.
7. The computer implemented method of claim 1 , wherein the second set of properties includes one or more of:
properties from one or more areas where hail is uncommon; and
properties randomly selected.
8. The computer implemented method of claim 1 , wherein validating the new damage frequency model uses data held out from training based on a time period.
9. The computer implemented method of claim 1 , wherein validating the new damage frequency model uses data held out from training based on an area associated with the data.
10. A system comprising:
a processor; and
a memory, the memory storing instructions that, when executed by the processor, cause the system to:
identify a plurality of properties including a first set of properties that experienced hail damage and a second set of properties that did not experience hail damage;
determine a location and a first set of features for each property in the plurality of properties;
wherein determining the first set of features for each property in the plurality of properties includes performing feature selection, the feature selection including repeatedly:
removing a different set of one or more features from a set of available features to create a reduced set of features;
training a new damage frequency model;
validating the new damage frequency model; and
comparing performance of the new damage frequency model to other instances of damage frequency model which used different feature sets,
wherein the first set of features is a reduced set of features.
11. The system of claim 10 , the memory further storing instructions that, when executed by the processor, cause the system to:
receive a location;
apply a first damage frequency model trained and validated based on the first set of features based on the performance of the first damage model compared to the other instances of the damage frequency model; and
present a damage frequency.
12. The system of claim 10 , wherein the first set of features comprises a building area, a vegetation density, a roof material, a roof quality, a temperature, a precipitation type or metric, an elevation, an average hail frequency, and a mean hail size.
13. The system of claim 10 , wherein the first set of features includes one or more of a building area, a vegetation density, a roof material, a roof quality, a roof height, a roof pitch, presence of skylights, a portion of roof covered by skylights, presence of a solar panel, a portion of the roof covered by solar panels, a number of roof facets, a roof shape, a land cover code, a temperature, a precipitation type or metric, and an elevation.
14. The system of claim 10 , wherein the first set of properties includes properties associated with first claims data representing claims for hail damage to the first set of properties, and the second set of properties includes properties associated with second claims data representing claims for damage to the second set of properties unrelated to hail.
15. The system of claim 10 , wherein the first set of properties includes:
properties associated with building permits for roof repair, or roof replacement, and indicating hail as a reason for repair or replacement.
16. The system of claim 10 , wherein the second set of properties includes one or more of:
properties from one or more areas where hail is uncommon; and
properties randomly selected.
17. The system of claim 10 , wherein validating the new damage frequency model uses data held out from training based on a time period.
18. The system of claim 10 , wherein validating the new damage frequency model uses data held out from training based on an area associated with the data.