SYSTEM AND METHOD FOR PROPERTY CONDITION ANALYSIS
Systems and methods for property condition analysis comprising: determining a measurement, optionally determining a set of property attributes, determining a condition score, optionally providing the condition score, and optionally training a condition scoring model, determining a measurement depicting a property; determining parcel data associated with the property; determining a set of attributes for the property, based on the measurement and the parcel data, using a set of attribute models; and determining a condition score based on the set of attributes using a condition scoring model.
1 . A method, comprising:
determining a measurement depicting a property;
determining parcel data associated with the property;
determining a set of attributes for the property, based on the measurement and the parcel data, using a set of attribute models; and
determining a condition score based on the set of attributes using a condition scoring model.
2 . The method of claim 1 , further comprising determining a description vector based on a property description using a separate model, wherein the condition score is further determined based on the descriptive vector using the condition scoring model.
3 . The method of claim 1 , wherein an attribute model of the set determines the respective attribute based on features extracted from the measurement.
4 . The method of claim 1 , wherein the set of attributes comprises at least one of a roof condition, a pool condition, or yard debris.
5 . The method of claim 1 , further comprising determining an explainability description associated with the condition score based on the set of attributes.
6 . The method of claim 1 , wherein the condition score is used as an input to an automated valuation model.
7 . The method of claim 1 , wherein the set of attributes comprises a paved surface condition.
8 . The method of claim 1 , wherein the measurement comprises remote exterior imagery.
9 . The method of claim 1 , wherein determining the set of attributes comprises determining a segmented measurement from the measurement based on the parcel data, wherein an attribute model of the set of attribute models determines a value for an attribute of the set of attributes based on the segmented measurement.
10 . The method of claim 1 , wherein the condition scoring model comprises a first submodel and a second submodel, wherein determining the condition score comprises:
determining the condition score using the first submodel when the set of attributes satisfies a set of conditions associated with the condition score; and
calculating the condition score using the second submodel when the set of attributes does not satisfy the set of conditions.
11 . The method of claim 10 , wherein the first submodel comprises a ruleset model or a heuristics model, and the second submodel comprises a classification model or a regression model.
12 . The method of claim 1 , wherein the measurement comprises a color image.
13 . The method of claim 1 , an attribute model of the set of attribute models is trained using different training data from the condition scoring model.
14 . A method for training a condition scoring model, comprising:
determining training data correlated with property condition for a set of training properties;
determining a training data distribution parameter for the training data; and
training the condition scoring model to predict property condition scores based on a measurement for each training property of the set of training properties, wherein a distribution parameter for the property condition scores is comparable to the training data distribution parameter.
15 . The method of claim 14 , wherein the training data excludes a property condition score for each training property of the set of training properties, wherein the training data comprises a condition indicator value for each training property of the set of training properties.
16 . The method of claim 15 , wherein the condition indicator value comprises a claim severity.
17 . The method of claim 15 , wherein the condition indicator value comprises a condition rating between C1-C6.
18 . The method of claim 14 , wherein the training data comprises at least one of inspection data, appraisal data, broker price opinion data, or historical claim data.
19 . The method of claim 14 , wherein the property condition scores are predicted by:
determining a measurement for the training property;
determining values for a set of attributes for the training property based on the measurement, using a set of attribute models; and
determining a property condition score based on the values of the set of attributes, using the condition scoring model.
20 . The method of claim 14 , wherein the measurement comprises remote exterior imagery.