Structuring of unstructured web posting
Provided are systems and methods for improving understanding and access of unstructured text-based web postings. In one example, a method may include receiving a web posting comprising unstructured text, identifying one or more explicit attributes of an activity based on text expressly recited in the unstructured text, determining one or more hidden attributes of the activity not recited in the unstructured text based on a pattern of words included in the unstructured text, converting the unstructured text into a structured object format comprising the one or more explicit attributes and the one or more determined hidden attributes, and outputting the web posting with the structured object format for display via a user interface.
1. A computing system comprising:
a memory configured to store unstructured text of a plurality of web postings; and
a processor configured to:
read a string of unstructured text of a web posting,
transform the string of unstructured text into a structured display object that comprises a normalized list of fields for attribute values,
identify an attribute value that is expressly recited in the web posting and fill-in a first field of the normalized list of fields with the identified attribute value;
execute a machine learning model on the string of unstructured text from the web posting to infer a second attribute value from the web posting that is not expressly recited in the web posting based on a pattern of words included in the string of unstructured text of the web posting and fill-in a second field of the normalized list of fields with the second attribute value inferred by the machine learning model, and
display the structured display object via a display device including the identified attribute value and the inferred second attribute value filled-in first and second fields, respectively.
2. The computing system of claim 1 , wherein the processor is configured to display names of the attribute types next to the common normalized list of fields.
3. The computing system of claim 1 , wherein the machine learning model is configured to infer the second attribute value based on one or more attribute values that are expressly recited in the string of unstructured text.
4. The computing system of claim 1 , wherein the processor is further configured to store content of attribute values included in the normalized list of fields of the structured display object in a JavaScript Object Notation (JSON) object.
5. The computing system of claim 4 , wherein the processor is further configured to determine an identifier of the web posting and store the identifier of the web posting in the JSON object.
6. The computing system of claim 1 , wherein the processor is further configured to convert the structured display object into a normalized web posting and display the normalized web posting via a website.
7. A method comprising:
storing unstructured text of a plurality of web postings;
reading a string of unstructured text of a web posting;
transforming the string of unstructured text into a structured display object that comprises a common normalized list of fields for attribute values;
identifying an attribute value that is expressly recited in the web posting and filling-in a first field of the normalized list of fields with the identified attribute value;
executing a machine learning model on the string of unstructured text from the web posting to infer a second attribute value from the web posting that is not expressly recited in the web posting based on a pattern of words included in the string of unstructured text of the web posting and filling-in a second field of the normalized list of fields with the second attribute value inferred by the machine learning model; and
displaying the structured display object via a display device including the identified attribute value and the inferred second attribute value filled-in first and second fields, respectively.
8. The method of claim 7 , wherein the method further comprises displaying names of the attribute types next to the common normalized list of fields.
9. The method of claim 7 , wherein the machine learning model is configured to infer the second attribute value based on one or more other attribute values that are expressly recited in the string of unstructured text.
10. The method of claim 7 , wherein the method further comprises storing content of attribute values included in the normalized list of fields of the structured display object in a JavaScript Object Notation (JSON) object.
11. The method of claim 10 , wherein the method further comprises determining an identifier of the web posting and storing the identifier of the web posting in the JSON object.
12. The method of claim 7 , wherein the method further comprises converting the structured display object into a normalized web posting and displaying the normalized web posting via a website.
13. A non-transitory computer-readable medium comprising instructions which when executed by a processor cause a computer to perform a method comprising:
storing unstructured text of a plurality of web postings via a computing system;
reading a string of unstructured text of a web posting;
transforming the string of unstructured text into a structured display object that comprises a common normalized list of fields for attribute values;
identifying an attribute value that is expressly recited in the web posting and filling-in a first field of the normalized list of fields with the identified attribute value;
executing a machine learning model on the string of unstructured text from the first web posting to infer a second attribute value from the web posting that is not expressly recited in the web posting based on a pattern of words included in the string of unstructured text of the web posting and filling-in a second field of the list of fields with the second attribute value inferred by the machine learning model; and
displaying the structured display object via a display device including the identified attribute value and the inferred second attribute value filled-in first and second fields, respectively.
14. The non-transitory computer-readable medium of claim 13 , wherein the method further comprises displaying names of the attribute types next to the common normalized list of fields.
15. The non-transitory computer-readable medium of claim 13 , wherein the machine learning model is configured to infer the second attribute value based on one or more other attribute values that are expressly recited in the string of unstructured text.
16. The non-transitory computer-readable medium of claim 13 , wherein the method further comprises storing content of attribute values included in the normalized list of fields of the structured display object in a JavaScript Object Notation (JSON) object.
17. The non-transitory computer-readable medium of claim 13 , wherein the method further comprises converting the structured display object into a normalized web posting and displaying the normalized web posting via a website.