Machine learning using multiple input data types

1. A method comprising:

obtaining, in real-time and from one or more sources, a plurality of data points associated with a specified parcel of real property;

using a machine learning model to generate a prediction from the obtained plurality of data points, the prediction indicating a likelihood that the real property will satisfy a particular parameter, the particular parameter indicating a likelihood that the real property has a title defect, and a predicted scope for the parameter, wherein the machine learning model is trained using a training set comprising a collection of data points associated with a labeled set of a plurality of real property parcels distinct from the specified parcel of real property, wherein the data points for each real property parcel comprises one or more data types, the data types including statistical information about the real property parcel and a retail history of the property, and wherein each label indicates the actual historical occurrence and value for the particular parameter for each real property parcel of the training set such that the training set includes one or more real property parcels labeled as having a title defect of specified value and one or more real property parcels labeled as not having a title defect; and

based on the prediction indicating the likelihood that the real property has a title defect and the predicted scope for the title defect, classifying the specified parcel of real property according to a determination of whether the predicted scope satisfies a threshold value.

2. The method of claim 1 , wherein the obtained plurality of data points comprises a variety of data from a variety of data sources.

3. The method of claim 2 , wherein the obtained plurality of data points is obtained from both structured data sources and unstructured data sources.

4. The method of claim 3 , wherein obtaining data points from unstructured data sources includes processing unstructured data in the unstructured data sources to identify particular types of information including performing image recognition on images contained within the unstructured data.

5. The method of claim 1 , wherein each data point is associated with a particular data type and wherein each data type is weighted in the machine learning model.

6. The method of claim 5 , wherein the weight given to each data type varies depending on a particular mix of data types input to the machine learning model.

7. The method of claim 1 , wherein values for data points input into the machine learning model for different parcels of real property are tracked to determine whether the values actually input correspond to values anticipated by the machine learning model.

8. The method of claim 1 , wherein an actual outcome with respect to the particular parameter for the specific parcel of real property is determined and compared with the prediction.

9. The method of claim 8 , wherein in response to identifying inaccuracies in one or more predictions, adjusting the machine learning model based on updated training data.

10. 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, in real-time and from one or more sources, a plurality of data points associated with a specified parcel of real property;

using a machine learning model to generate a prediction from the obtained plurality of data points, the prediction indicating a likelihood that the real property will satisfy a particular parameter, the particular parameter indicating a likelihood that the real property has a title defect, and a predicted scope for the parameter, wherein the machine learning model is trained using a training set comprising a collection of data points associated with a labeled set of a plurality of real property parcels distinct from the specified parcel of real property, wherein the data points for each real property parcel comprises one or more data types, the data types including statistical information about the real property parcel and a retail history of the property, and wherein each label indicates the actual historical occurrence and value for the particular parameter for each real property parcel of the training set such that the training set includes one or more real property parcels labeled as having a title defect of specified value and one or more real property parcels labeled as not having a title defect; and

based on the prediction indicating the likelihood that the real property has a title defect and the predicted scope for the title defect, classifying the specified parcel of real property according to a determination of whether the predicted scope satisfies a threshold value.

11. The system of claim 10 , wherein the obtained plurality of data points comprises a variety of data from a variety of data sources.

12. The system of claim 11 , wherein the obtained plurality of data points is obtained from both structured data sources and unstructured data sources.

13. The system of claim 12 , wherein obtaining data points from unstructured data sources includes processing unstructured data in the unstructured data sources to identify particular types of information including performing image recognition on images contained within the unstructured data.

14. The system of claim 10 , wherein each data point is associated with a particular data type and wherein each data type is weighted in the machine learning model.

15. The system of claim 14 , wherein the weight given to each data type varies depending on a particular mix of data types input to the machine learning model.

16. The system of claim 10 , wherein values for data points input into the machine learning model for different parcels of real property are tracked to determine whether the values actually input correspond to values anticipated by the machine learning model.

17. The system of claim 10 , wherein an actual outcome with respect to the particular parameter for the specific parcel of real property is determined and compared with the prediction.

18. The system of claim 17 , wherein in response to identifying inaccuracies in one or more predictions, adjusting the machine learning model based on updated training data.

19. 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, in real-time and from one or more sources, a plurality of data points associated with a specified parcel of real property;

using a machine learning model to generate a prediction from the obtained plurality of data points, the prediction indicating a likelihood that the real property will satisfy a particular parameter, the particular parameter indicating a likelihood that the real property has a title defect, and a predicted scope for the parameter, wherein the machine learning model is trained using a training set comprising a collection of data points associated with a labeled set of a plurality of real property parcels distinct from the specified parcel of real property, wherein the data points for each real property parcel comprises one or more data types, the data types including statistical information about the real property parcel and a retail history of the property, and wherein each label indicates the actual historical occurrence and value for the particular parameter for each real property parcel of the training set such that the training set includes one or more real property parcels labeled as having a title defect of specified value and one or more real property parcels labeled as not having a title defect; and

based on the prediction indicating the likelihood that the real property has a title defect and the predicted scope for the title defect, classifying the specified parcel of real property according to a determination of whether the predicted scope satisfies a threshold value.