Method for analyzing document for desired content and exracting same, electronic device employing method, and non-transitory storage medium

An electronic device comprising:

• at least one processor;
• and
• a data storage storing one or more programs which when executed by the at least one processor, cause the at least one processor to:
• obtain text information of a document;
• search the text information of the document based on first keywords to determine first pages, and input the first pages into a predetermined learning model, wherein the predetermined learning model is pretrained with multiple sets of sample documents and keywords extracted from the multiple sets of sample documents, each of the multiple sets of sample documents comprises at least one of texts, tables, and pictures, and the predetermined learning model is loaded and run by the processor;
• extract second keywords from text information of the first pages based on the predetermined learning model, cluster the first keywords and the second keywords according to keyword associations to obtain clustered keywords, and extract third keywords from the clustered keywords based on the predetermined learning model;
• search the text information of the first pages based on the third keywords to determine second pages;
• determine whether the second pages meet a predetermined page standard;
• integrate the second pages in order to output integrated second pages when the second pages meet the predetermined page standard;
• and
• input the second pages into the predetermined learning model as training samples in response to the second pages fail to meet the predetermined page standard, and retrain the predetermined learning model with the training samples.

A non-transitory storage medium having stored thereon instructions that, when executed by a processor of an electronic device, causes the electronic device to perform a text extraction method, the text extraction method comprising:

• obtaining, by the processor, text information of a document;
• searching, by the processor, the text information of the document based on first keywords to determine first pages, and inputting, by the processor, the first pages into a predetermined learning model, wherein the predetermined learning model is pretrained with multiple sets of sample documents and keywords extracted from the multiple sets of sample documents, each of the multiple sets of sample documents comprises at least one of texts, tables, and pictures, and the predetermined learning model is loaded and run by the processor;
• extracting, by the processor, second keywords from text information of the first pages based on the predetermined learning model, clustering, by the processor, the first keywords and the second keywords according to keyword associations to obtain clustered keywords, and extracting, by the processor, third keywords from the clustered keywords based on the predetermined learning model;
• searching, by the processor, the text information of the first pages based on the third keywords to determine second pages;
• determining, by the processor, whether the second pages meet a predetermined page standard;
• and
• integrating, by the processor, the second pages in order to output integrated second pages when the second pages meet the predetermined page standard;
• and
• inputting, by the processor, the second pages into the predetermined learning model as training samples in response to the second pages fail to meet the predetermined page standard, and retraining, by the processor, the predetermined learning model with the training samples.