Method and system for performing summarization of text

1. A method of performing summarization of text, the method comprising:

receiving, by a text summarization device, an input text, wherein the input text comprises a plurality of paragraphs;

receiving, by the text summarization device, a user-query comprising one or more tokens;

segregating, by the text summarization device, the input text into the plurality of paragraphs, wherein each of the plurality of paragraphs comprises a plurality of tokens;

creating, by the text summarization device, a plurality of paragraph-vectors, wherein each of the plurality of paragraph-vectors is representative of a paragraph of the plurality of paragraphs;

clustering, by the text summarization device, the plurality of paragraph-vectors to generate one or more clusters of paragraph-vectors;

determining, by the text summarization device, a relevant cluster of paragraph-vectors from the one or more clusters of paragraph-vectors by iteratively re-clustering the plurality of paragraph-vectors, based on an optimum number of clusters of paragraph-vectors and a degree of similarity of each cluster of paragraph-vectors with the user query using a first cosine similarity analysis, wherein the relevant cluster of paragraph-vectors is representative of a set of relevant paragraphs from the input text;

identifying, by the text summarization device, a set of relevant token-vectors from the relevant cluster of paragraph-vectors, based on a degree of similarity of each token-vector with the user query using a second cosine similarity analysis, wherein the set of relevant token-vectors is representative of a relevant excerpt from the relevant paragraphs; and

outputting, by the text summarization device, the set of relevant paragraphs corresponding to the relevant cluster of paragraph-vectors and the relevant excerpt from the relevant paragraphs.

2. The method of claim 1 , wherein creating a paragraph-vector comprises:

creating a plurality of token-vectors corresponding to the plurality of tokens belonging to each of the plurality of paragraphs; and

creating a paragraph-vector from the plurality of token-vectors based on determining a mean of the plurality of token-vectors.

3. The method of claim 2 , wherein the plurality of token-vectors is created corresponding to the plurality of tokens using at least one of a Word2vec model or a Global Vectors (GloVe) model.

4. The method of claim 1 , wherein the clustering is performed based on at least one of a hierarchical clustering model or a K-Means clustering model.

5. The method of claim 1 comprising:

verifying a number of clusters of paragraph-vectors generated.

6. The method of claim 5 , wherein verifying the number of clusters of paragraph-vectors generated comprises:

determining an optimum number of clusters of paragraph-vectors for the input text based on one of: a silhouette analysis, an elbow method, or a dendrogram;

comparing the number of clusters of paragraph-vectors generated with the optimum number of clusters of paragraph-vectors; and

verifying the number of clusters generated based on the comparison.

7. The method of claim 6 comprising:

modifying a clustering threshold, when the number of clusters of paragraph-vectors generated is more than the optimum number of clusters of paragraph-vectors.

8. A system for performing summarization of text, the system comprising:

a processor; and

a memory configured to store a plurality of processor executable instructions, wherein, upon execution by the processor, the plurality of processor executable instructions cause the processor to:

receive an input text, wherein the input text comprises a plurality of paragraphs;

receive a user-query comprising one or more tokens;

segregate the input text into the plurality of paragraphs, wherein each of the plurality of paragraphs comprises a plurality of tokens;

create a plurality of paragraph-vectors, wherein each of the plurality of paragraph-vectors is representative of a paragraph of the plurality of paragraphs;

cluster the plurality of paragraph-vectors to generate one or more clusters of paragraph-vectors;

determine a relevant cluster of paragraph-vectors from the one or more clusters of paragraph-vectors by iteratively re-clustering the plurality of paragraph-vectors, based on an optimum number of clusters of paragraph-vectors and a degree of similarity of each cluster of paragraph-vectors with the user query using a first cosine similarity analysis, wherein the relevant cluster of paragraph-vectors is representative of a set of relevant paragraphs from the input text;

identify a set of relevant token-vectors from the relevant cluster of paragraph-vectors, based on a degree of similarity of each token-vector with the user query using a second cosine similarity analysis, wherein the set of relevant token-vectors is representative of a relevant excerpt from the relevant paragraphs; and

output the set of relevant paragraphs corresponding to the relevant cluster of paragraph-vectors and the relevant excerpt from the relevant paragraphs.

9. The system of claim 8 , wherein creating a paragraph-vector comprises:

creating a plurality of token-vectors corresponding to the plurality of tokens belonging to each of the plurality of paragraphs; and

creating a paragraph-vector from the plurality of token-vectors based on determining a mean of the plurality of token-vectors.

10. The system of claim 8 , wherein the processor executable instructions, upon execution, cause the processor to:

verify a number of clusters of paragraph-vectors generated, wherein verifying the number of clusters of paragraph-vectors generated comprises:

determining an optimum number of clusters of paragraph-vectors for the input text based on one of: a silhouette analysis, an elbow method, or a dendrogram;

comparing the number of clusters of paragraph-vectors generated with the optimum number of clusters of paragraph-vectors; and

verifying the number of clusters generated based on the comparison.

11. The system of claim 10 , wherein the processor executable instructions, upon execution, cause the processor to:

modify a clustering threshold, when the number of clusters of paragraph-vectors generated is more than the optimum number of clusters of paragraph-vectors.

12. A non-transitory computer-readable medium storing computer-executable instructions for creating a process plan, the computer-executable instructions configured for:

receiving an input text, wherein the input text comprises a plurality of paragraphs; receiving a user-query comprising one or more tokens;

segregating the input text into the plurality of paragraphs, wherein each of the plurality of paragraphs comprises a plurality of tokens;

creating a plurality of paragraph-vectors, wherein each of the plurality of paragraph-vectors is representative of a paragraph of the plurality of paragraphs;

clustering the plurality of paragraph-vectors to generate one or more clusters of paragraph-vectors;

determining a relevant cluster of paragraph-vectors from the one or more clusters of paragraph-vectors by iteratively re-clustering the plurality of paragraph-vectors, based on an optimum number of clusters of paragraph-vectors and a degree of similarity of each cluster of paragraph-vectors with the user query using a first cosine similarity analysis, wherein the relevant cluster of paragraph-vectors is representative of a set of relevant paragraphs from the input text;

identify a set of relevant token-vectors from the relevant cluster of paragraph-vectors, based on a degree of similarity of each token-vector with the user query using a second cosine similarity analysis, wherein the set of relevant token-vectors is representative of a relevant excerpt from the relevant paragraphs; and

outputting the set of relevant paragraphs corresponding to the relevant cluster of paragraph-vectors and the relevant excerpt from the relevant paragraphs.

13. The non-transitory computer-readable medium of claim 12 , wherein creating a paragraph-vector comprises:

creating a plurality of token-vectors corresponding to the plurality of tokens belonging to each of the plurality of paragraphs; and

creating a paragraph-vector from the plurality of token-vectors based on determining a mean of the plurality of token-vectors.

14. The non-transitory computer-readable medium of claim 12 , wherein the computer-executable instructions further configured for:

verifying a number of clusters of paragraph-vectors generated, wherein verifying the number of clusters of paragraph-vectors generated comprises:

determining an optimum number of clusters of paragraph-vectors for the input text based on one of: a silhouette analysis, an elbow method, or a dendrogram;

comparing the number of clusters of paragraph-vectors generated with the optimum number of clusters of paragraph-vectors; and

verifying the number of clusters generated based on the comparison.

15. The non-transitory computer-readable medium of claim 14 , wherein the computer-executable instructions further configured for:

modifying a clustering threshold, when the number of clusters of paragraph-vectors generated is more than the optimum number of clusters of paragraph-vectors.