Normalization of unstructured catalog data

1. A computing system comprising:

a network interface configured to receive a record comprising an unstructured description of an object; and

a processor configured to

identify a type of the object from among a plurality of object types;

identify, via a first machine learning algorithm trained to identify a value of a predefined generic attribute that is applicable to each of the plurality of object types, a first raw text value that is recited within the unstructured description of the object and which describes the predefined generic attribute of the identified type of object,

identify, via a second machine learning algorithm trained to identify a value of a predefined object-specific attribute that is applicable to the type of object but not other types in the plurality of object types, a second raw text value that is recited within the unstructured description of the object and which describes the predefined object-specific attribute of the identified type of object,

modify the first and second raw text values identified via the first and second machine learning algorithms to generate normalized attribute values, and

store a structured record of the object in a structured format comprising a plurality of values of a plurality of attributes of the object from the unstructured description including the normalized attribute values for the predefined attribute of the object.

2. The computing system of claim 1 , wherein the processor is configured to identify the first raw text value of the generic attribute via a query of a database that stores hierarchical relationships of object attributes based on the identified type of the object.

3. The computing system of claim 1 , wherein the processor is configured to convert one or more words of the first raw text value into one or more normalized words to generate the normalized attribute values.

4. The computing system of claim 1 , wherein the structured format comprises a list of fields that store the plurality of values corresponding to the plurality of attributes of the object.

5. The computing system of claim 1 , wherein the processor is configured to store the plurality of values of the object in a search index that comprises information about a plurality of searchable objects.

6. The computing system of claim 1 , wherein the processor is configured to identify the first raw text value of the predefined generic attribute via execution of a machine learning algorithm that identifies values of an attribute that is universal among different types of products.

7. The computing system of claim 1 , wherein the processor is further configured to execute a query which identifies the structured record of the object as a search result of the query based on the normalized attribute values for the predefined attribute of the object.

8. The computing system of claim 1 , wherein the processor is configured to determine to search for the predefined generic attribute and the predefined object-specific attribute via a query of a hierarchical taxonomy in which attributes of an object are stored in a lowermost level of the hierarchical taxonomy.

9. A method comprising:

receiving a record comprising an unstructured description of an object;

identifying a type of the object from among a plurality of object types;

identifying, via a first machine learning algorithm trained to identify a predefined generic attribute that is applicable to each of the plurality of object types, a first raw text value that is recited within the unstructured description of the object and which describes the predefined generic attribute of the identified type of object;

identifying, via a second machine learning algorithm trained to identify a predefined object-specific attribute that is applicable to the type of object but not other types in the plurality of object types, a second raw text value that is recited within the unstructured description of the object and which describes the predefined object-specific attribute of the identified type of object;

modifying the first and second raw text values identified via the first and second machine learning algorithms to generate normalized attribute values; and

storing a structured record of the object in a structured format comprising a plurality of values of a plurality of attributes of the object from the unstructured description including the normalized attribute values for the predefined attribute of the object.

10. The method of claim 9 , wherein the identifying the first raw text value of the generic attribute comprises querying a database storing hierarchical relationships of object attributes based on the identified type of the object.

11. The method of claim 9 , wherein the modifying comprises converting one or more words of the first raw text value into one or more normalized words to generate the normalized attribute values.

12. The method of claim 9 , wherein the structured format comprises a list of fields storing the plurality of values corresponding to the plurality of attributes of the object.

13. The method of claim 9 , wherein the storing the structured record comprises storing the plurality of values of the object in a search index comprising information about a plurality of searchable objects.

14. The method of claim 9 , wherein the identifying the raw text comprises identifying the first raw text value of the predefined generic attribute via execution of a machine learning algorithm that identifies values of an attribute that is universal among a plurality of different types of objects.

15. The method of claim 9 , further comprising executing a query which identifies the structured record of the object as a search result of the query based on the normalized attribute values for the predefined attribute of the object.

16. A non-transitory computer readable medium comprising program instructions that when executed cause a computer to perform a method comprising:

receiving a record comprising an unstructured description of an object;

identifying a type of the object from among a plurality of object types;

identifying, via a first machine learning algorithm trained to identify a predefined generic attribute that is applicable to each of the plurality of object types, a first raw text value that is recited within the unstructured description of the object and which describes the predefined generic attribute of the identified type of object;

identifying, via a second machine learning algorithm trained to identify a predefined object-specific attribute that is applicable to the type of object but not other types in the plurality of object types, a second raw text value that is recited within the unstructured description of the object and which describes the predefined object-specific attribute of the identified type of object;

modifying the first and second raw text values identified via the first and second machine learning algorithms to generate normalized attribute values; and

storing a structured record of the object in a structured format comprising a plurality of values of a plurality of attributes of the object from the unstructured description including the normalized attribute values for the predefined attribute of the object.

17. The non-transitory computer readable medium of claim 16 , wherein the identifying the first raw text value of the generic attribute comprises querying a database storing hierarchical relationships of object attributes based on the identified type of the object.

18. The non-transitory computer readable medium of claim 16 , wherein the modifying comprises converting one or more words of the first raw text value into one or more normalized words to generate the normalized attribute values.

19. The non-transitory computer readable medium of claim 16 , wherein the structured format comprises a list of fields storing the plurality of values corresponding to the plurality of attributes of the object.