Method and system for metadata extraction for document identification
A method includes obtaining a document comprising earth property data regarding a geological region of interest, preprocessing the document to form at least one preprocessed document and determining, using a set of trained machine-learned models processing the at least one preprocessed document, a category of the document and a type of the document. The method further includes determining, using a natural language processing algorithm, metadata attributes of the document and a title of the document, and updating a database storing the document with the title, the category, the type and the metadata attributes. The method further includes identifying, by a planning module processing a query, the document from the database based on at least one of the title, the category, the type and the metadata attributes and planning a wellbore path in the geological region of interest using the earth property data comprised in the document.
1 . A method, comprising:
obtaining a document comprising earth property data regarding a geological region of interest;
preprocessing the document to form at least one preprocessed document;
determining, using a set of trained machine-learned models processing the at least one preprocessed document, a category of the document and a type of the document, wherein the category is a class of subsurface exploration methods that the document belongs to;
determining, using a natural language processing algorithm processing the at least one preprocessed document, metadata attributes of the document and a title of the document;
updating a database storing the document with the title, the category, the type and the metadata attributes;
identifying, by a planning module processing a query, the document from the database based on at least one of the title, the category, the type and the metadata attributes; and
planning a wellbore path in the geological region of interest using the earth property data comprised in the document,
wherein determining the metadata attributes of the document comprises:
generating, from the at least one preprocessed document, a set of n-grams;
obtaining from an exploration reference database, based on the category, a set of possible attributes values; and
determining the metadata attributes based on an intersection between the set of n-grams and the set of possible attributes values.
2 . The method of claim 1 , further comprising:
determining a location of a hydrocarbon reservoir in the geological region of interest using the earth property data; and
planning the wellbore path so as to cause a wellbore to penetrate the hydrocarbon reservoir based on the location.
3 . The method of claim 1 , wherein the set of trained machine-learned models comprises at least one support vector machine.
4 . The method of claim 1 , wherein determining the metadata attributes is further based on one of a set of metadata extraction exceptions and a set of metadata extraction conditions.
5 . The method of claim 1 , wherein determining the title of the document comprises:
identifying a set of sentences within the at least one preprocessed document;
determining a weight of each sentence in the set of sentences; and
identifying the title based on the weights.
6 . The method of claim 5 , wherein the weight is determined based on at least one of a list of included and excluded words and a formatting pattern of the sentence.
7 . The method of claim 5 , identifying the title based on the weights further comprises:
determining that the weight of a sentence of the set of sentences meets a predetermined criterion; and
returning the sentence as the title.
8 . The method of claim 5 , identify the title based on the weights further comprises:
determining that the weights do not meet a predetermined criterion; and
returning the metadata attributes as the title.
9 . The method of claim 1 , prior to updating the database storing the document with the title, the category, the type and the metadata attributes, the method further comprising:
determining, by a quality checking module that the title, the category, the type and the metadata attributes satisfy a set of predetermined conditions and exemptions.
10 . A system, comprising:
a machine-learned model;
a machine-readable medium storing a natural language processing (NLP) algorithm; and
a computer configured to:
obtain a document comprising earth property data regarding a geological region of interest;
preprocess the document to form at least one preprocessed document;
determine, using a set of trained machine-learned models processing the at least one preprocessed document, a category of the document and a type of the document, wherein the category is a class of subsurface exploration methods that the document belongs to;
determine, using a natural language processing algorithm processing the at least one preprocessed document, metadata attributes of the document and a title of the document;
update a database storing the document with the title, the category, the type and the metadata attributes;
identify, by a planning module processing a query, the document from the database based on at least one of the title, the category, the type and the metadata attributes; and
plan a wellbore path in the geological region of interest using the earth property data comprised in the document,
wherein the determine the metadata attributes of the document comprises:
generate, from the at least one preprocessed document, a set of n-grams;
obtain from an exploration reference database, based on the category, a set of possible attributes values; and
determine the metadata attributes based on an intersection between the set of n-grams and the set of possible attributes values.
11 . The system of claim 10 , the computer further configured to:
determine a location of a hydrocarbon reservoir in the geological region of interest using the earth property data; and
plan the wellbore path so as to cause a wellbore to penetrate the hydrocarbon reservoir based on the location.
12 . The system of claim 10 , wherein the set of trained machine-learned models comprises at least one support vector machine.
13 . The system of claim 10 , wherein determine the metadata attributes is further based on one of a set of metadata extraction exceptions and a set of metadata extraction conditions.
14 . The system of claim 10 , the computer further configured to:
identify a set of sentences within the at least one preprocessed document;
determine a weight of each sentence in the set of sentences; and
identify the title based on the weights.
15 . The system of claim 14 , wherein the weight is determined based on at least one of a list of included and excluded words and a formatting pattern of the sentence.
16 . The system of claim 14 , the computer further configured to:
determine that the weight of a sentence of the set of sentences meets a predetermined criterion; and
return the sentence as the title.
17 . The system of claim 14 , the computer further configured to:
determine that the weights do not meet a predetermined criterion; and
return the metadata attributes as the title.
18 . A non-transitory machine-readable medium comprising a plurality of machine-readable instructions executed by one or more processors, the plurality of machine-readable instructions causing the one or more processors to perform a method comprising:
obtaining a document comprising earth property data regarding a geological region of interest;
preprocessing the document to form at least one preprocessed document;
determining, using a set of trained machine-learned models processing the at least one preprocessed document, a category of the document and a type of the document, wherein the category is a class of subsurface exploration methods that the document belongs to;
determining, using a natural language processing algorithm processing the at least one preprocessed document, metadata attributes of the document and a title of the document;
updating a database storing the document with the title, the category, the type and the metadata attributes;
identifying, by a planning module processing a query, the document from the database based on at least one of the title, the category, the type and the metadata attributes; and
planning a wellbore path in the geological region of interest using the earth property data comprised in the document wherein determining the metadata attributes of the document comprises:
generating, from the at least one preprocessed document, a set of n-grams;
obtaining from an exploration reference database, based on the category, a set of possible attributes values; and
determining the metadata attributes based on an intersection between the set of n-grams and the set of possible attributes values.