Predictive machine learning models for parcels of real property
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training and applying a machine learning model. One of the methods includes the actions of obtaining a collection of data associated with a specified parcel of real property, wherein the collection of data includes one or more parameters of interest; using a machine learning model to generate a prediction from the input collection of data for each of the one or more parameters of interest, wherein the prediction for each parameter of interest comprises a likelihood value that the parameter satisfies a particular condition, and wherein the machine learning model is trained using a training set comprising a collection of data associated with a labeled set of data, the labels indicating the existence of particular parameters on the parcels; and based on the prediction, classifying each of the one or more parameters of interest.
1 . A method comprising:
obtaining, from one or more sources, a collection of data associated with a specified parcel of real property, wherein the collection of data includes one or more involuntary liens associated with the specified parcel of real property;
training a machine learning model using a training set comprising a collection of data points and labels for a set of real property parcels distinct from the specified parcel of real property, each real property parcel of the training set including information about each involuntary lien attached to the parcel, wherein the training comprises extracting features from the training set and determining associated weights such that a prediction for each involuntary lien attached to the parcel generated by the machine learning model corresponds to the known parameter values;
inputting the collection of data to the machine learning model;
using the machine learning model to generate a prediction from the input collection of data for each of the one or more involuntary liens comprising:
generating for each of the one or more involuntary liens a respective probability value that the involuntary lien has not been released from the specified parcel of real property;
for each involuntary lien:
determining whether probability value that the involuntary lien has not been released from the specified parcel of real property satisfies one or more threshold values;
determining that the involuntary lien to be released or not released based on whether or not the probability satisfies the one or more threshold values; and
using the determination for each involuntary lien as either released or not released to determine a risk tolerance for the parcel of real property and, based on the risk tolerance, determining whether to manually resolve one or more involuntary liens or to allow a title evaluation process to proceed for the parcel of real property.
2 . The method of claim 1 , wherein the obtained collection of data comprises a variety of data from a variety of data sources including a first data set for the specified parcel of real property and a second data set for one or more individuals associated with the specified parcel of real property.
3 . The method of claim 1 , wherein classifying a particular involuntary lien comprises comparing the likelihood value for the particular involuntary lien with a first threshold and a second threshold, wherein:
in response to determining that the likelihood value satisfies a first threshold, assigning a first classification to the involuntary lien,
in response to determining that the likelihood value satisfies the second threshold, assigning a second classification to the involuntary lien, and
in response to determining that the likelihood value satisfies neither the first nor the second thresholds, classifying the involuntary lien for manual evaluation.
4 . The method of claim 1 , wherein the data associated with the specified parcel of real property input to the machine learning model include dates associated with a recordation of the one or more involuntary liens and transaction data indicating dates in which ownership of the parcel changed.
5 . The method of claim 1 , wherein training the machine learning model includes learning how features of the training set affect the likelihood of the involuntary lien, wherein the feature includes one or more of an age of the involuntary lien, a type of the involuntary lien, a timing of other transactions associated with an associated parcel, and one or more changes in a status of persons associated with the parcel.
6 . A system comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
obtaining, from one or more sources, a collection of data associated with a specified parcel of real property, wherein the collection of data includes one or more involuntary liens associated with the specified parcel of real property;
training a machine learning model using a training set comprising a collection of data points and labels for a set of real property parcels distinct from the specified parcel of real property, each real property parcel of the training set including information about each involuntary lien attached to the parcel, wherein the training comprises extracting features from the training set and determining associated weights such that a prediction for each involuntary lien attached to the parcel generated by the machine learning model corresponds to the known parameter values;
inputting the collection of data to the machine learning model;
using the machine learning model to generate a prediction from the input collection of data for each of the one or more involuntary liens comprising:
generating for each of the one or more involuntary liens a respective probability value that the involuntary lien has not been released from the specified parcel of real property;
for each involuntary lien:
determining whether probability value that the involuntary lien has not been released from the specified parcel of real property satisfies one or more threshold values;
determining that the involuntary lien to be released or not released based on whether or not the probability satisfies the one or more threshold values; and
using the determination for each involuntary lien as either released or not released to determine a risk tolerance for the parcel of real property and, based on the risk tolerance, determining whether to manually resolve one or more involuntary liens or to allow a title evaluation process to proceed for the parcel of real property.
7 . The system of claim 6 , wherein the obtained collection of data comprises a variety of data from a variety of data sources including a first data set for the specified parcel of real property and a second data set for one or more individuals associated with the specified parcel of real property.
8 . The system of claim 6 , wherein classifying a particular involuntary lien comprises comparing the likelihood value for the involuntary lien with a first threshold and a second threshold, wherein:
in response to determining that the likelihood value satisfies a first threshold, assigning a first classification to the involuntary lien,
in response to determining that the likelihood value satisfies the second threshold, assigning a second classification to the involuntary lien, and
in response to determining that the likelihood value satisfies neither the first nor the second thresholds, classifying the involuntary lien for manual evaluation.
9 . The system of claim 6 , wherein the data associated with the specified parcel of real property input to the machine learning model include dates associated with a recordation of the one or more involuntary liens and transaction data indicating dates in which ownership of the parcel changed.
10 . The system of claim 6 , wherein training the machine learning model includes learning how features of the training set affect the likelihood of the involuntary lien, wherein the feature includes one or more of an age of the involuntary lien, a type of the involuntary lien, a timing of other transactions associated with an associated parcel, and one or more changes in a status of persons associated with the parcel.
11 . One or more non-transitory computer-readable storage media encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising:
obtaining, from one or more sources, a collection of data associated with a specified parcel of real property, wherein the collection of data includes one or more involuntary liens associated with the specified parcel of real property;
training a machine learning model using a training set comprising a collection of data points and labels for a set of real property parcels distinct from the specified parcel of real property, each real property parcel of the training set including information about each involuntary lien attached to the parcel, wherein the training comprises extracting features from the training set and determining associated weights such that a prediction for each involuntary lien attached to the parcel generated by the machine learning model corresponds to the known parameter values;
inputting the collection of data to the machine learning model;
using the machine learning model to generate a prediction from the input collection of data for each of the one or more involuntary liens comprising:
generating for each of the one or more involuntary liens a respective probability value that the involuntary lien has not been released from the specified parcel of real property;
for each involuntary lien:
determining whether probability value that the involuntary lien has not been released from the specified parcel of real property satisfies one or more threshold values;
determining that the involuntary lien to be released or not released based on whether or not the probability satisfies the one or more threshold values; and
using the determination for each involuntary lien as either released or not released to determine a risk tolerance for the parcel of real property and, based on the risk tolerance, determining whether to manually resolve one or more involuntary liens or to allow a title evaluation process to proceed for the parcel of real property.
12 . The non-transitory computer-readable storage media of claim 11 , wherein the obtained collection of data comprises a variety of data from a variety of data sources including a first data set for the specified parcel of real property and a second data set for one or more individuals associated with the specified parcel of real property.
13 . The non-transitory computer-readable storage media of claim 11 , wherein classifying a particular comprises comparing the likelihood value for the particular involuntary lien with a first threshold and a second threshold, wherein:
in response to determining that the likelihood value satisfies a first threshold, assigning a first classification to the involuntary lien,
in response to determining that the likelihood value satisfies the second threshold, assigning a second classification to the involuntary lien, and
in response to determining that the likelihood value satisfies neither the first nor the second thresholds, classifying the involuntary lien for manual evaluation.
14 . The non-transitory computer-readable storage media of claim 11 , wherein the data associated with the specified parcel of real property input to the machine learning model include dates associated with a recordation of the one or more involuntary liens and transaction data indicating dates in which ownership of the parcel changed.
15 . The non-transitory computer-readable storage media of claim 11 , wherein training the machine learning model includes learning how features of the training set affect the likelihood of the involuntary lien, wherein the feature includes one or more of an age of the involuntary lien, a type of the involuntary lien, a timing of other transactions associated with an associated parcel, and one or more changes in a status of persons associated with the parcel.