1. A method (600) of analysing a plurality of domain-specific documents, the method (600)
comprising:
determining (602), by a processor (104), one or more attributes in a first text
data extracted from the plurality of domain-specific documents based on a first set of
queries (400A) using a first generative machine learning model,
determining (604), by the processor (104), a second set of queries (400B) based
on one or more portions of the first text data corresponding to the one or more attributes
using a second generative machine learning model;
for each of a portion of the plurality of domain-specific documents:
determining (606), by the processor (104), a third set of queries (400C) based
on a domain type of the plurality of domain-specific documents and a second text data
extracted for a corresponding domain-specific document from the portion of the
plurality of domain-specific documents, using a third generative machine learning
model,
wherein the domain type is determined based on a subject matter
analysis of the plurality of domain-specific documents; and
determining (608), by the processor (104), a fourth set of queries (400D) based
on a comparison between the second set of queries (400B) and the third set of queries
(400C),
wherein the first set of queries (400A) are iteratively updated based on
the corresponding fourth set of queries (400D) determined for each of the
corresponding domain-specific document from the portion of the plurality of
domain-specific documents; and
determining (610), by the processor (104), a final set of queries corresponding to the
plurality of domain-specific documents based on the fourth set of queries (400D) determined
as about same as the second set of queries (400B).
2. The method (600) as claimed in claim 1, comprising:
analysing, by the processor (104), the plurality of domain-specific documents by:
querying, by the processor (104), a fourth generative machine learning model
using the final set of queries; and
16
displaying, by the processor (104), an answer text data output by the fourth
generative machine learning model for each query of the final set of queries based on
the first text data.
3. The method (600) as claimed in claim 1, wherein the second set of queries (400B) are
determined based on validation of the one or more portions of the first text data corresponding
to the one or more attributes based on a user input.
4. The method (600) as claimed in claim 1, wherein the first text data and the second text data
are extracted using a text extraction technique.
5. The method (600) as claimed in claim 1, wherein the comparison between the second set of
queries (400B) and the third set of queries (400C) is based on a comparison of character
embeddings and position vector of each character in a query text data of each query of the
second set of queries (400B) and the third set of queries (400C).
6. A system (100) of analysing a plurality of domain-specific documents, comprising:
a processor (104); and
a memory (106) communicably coupled to the processor, wherein the memory (106)
stores processor-executable instructions, which when executed by the processor (104), cause
the processor (104) to:
determine a first set of queries (400A) based on one or more attributes in a first text
data extracted from the plurality of domain-specific documents based on a first set of queries
(400A) using a first generative machine learning model,
determine a second set of queries (400B) based on one or more portions of the first text
data corresponding to the one or more attributes using a second generative machine learning
model;
for each of a portion of the plurality of domain-specific documents:
determine a third set of queries (400C) based on a domain type of the plurality
of domain-specific documents and a second text data extracted for a corresponding
domain-specific document from the portion of the plurality of domain-specific
documents,
wherein the domain type is determined based on a subject matter
analysis of the plurality of domain-specific documents; and
17
determine a fourth set of queries (400D) based on a comparison between the
second set of queries (400B) and the third set of queries (400C),
wherein the first set of queries (400A) are iteratively updated based on
the corresponding fourth set of queries (400D) determined for each of the
corresponding domain-specific document from the portion of the plurality of
domain-specific documents; and
determine a final set of queries corresponding to the plurality of domain-specific
documents based on the fourth set of queries (400D) determined as about same as the
second set of queries (400B).
7. The system (100) as claimed in claim 6, wherein the processor (104) is configured to:
analyze the plurality of domain-specific documents based on:
query a fourth generative machine learning model using the final set of queries;
and
display an answer text data output by the fourth generative machine learning
model for each query of the final set of queries based on the first text data.
8. The system (100) as claimed in claim 6, wherein the second set of queries (400B) are
determined based on validation of the one or more portions of the first text data corresponding
to the one or more attributes based on a user input.
9. The system (100) as claimed in claim 6, wherein the first text data and the second text data
are extracted using a text extraction technique.
10. The system (100) as claimed in claim 6, wherein the comparison between the second set of
queries (400B) and the third set of queries (400C) is based on a comparison of character
embeddings and position vector of each character in a query text data of each query of the
second set of queries (400B) and the third set of queries (400C).