Method for providing customized online education content for rehabilitation for hearing loss patient

1 . A method by which a chatbot application executed by at least one processor of a terminal provides a chatbot for rehabilitation education for a hearing loss patient,

the method comprising:

executing the chatbot that provides hearing loss rehabilitation content that is interactive learning content for hearing rehabilitation education for the hearing loss patient;

providing a sentence-type rehabilitation content in the hearing loss rehabilitation content on the basis of the executed chatbot;

acquiring user response data regarding an audio quiz of the provided sentence-type rehabilitation content;

performing a correct/wrong processing process for determining whether or not the acquired user response data is a correct answer;

collecting and profiling learning result data obtained through the correct/wrong processing process;

training a wrong answer learning model based on the collected learning result data to output wrong answer pattern information;

training a correct answer learning model based on the collected learning result data to output correct answer probability information;

generating a customized base text by inputting the wrong answer pattern information into a text generation model;

performing a computer-implemented verification process on the generated customized base text prior to providing the content to the user, wherein the verification process comprises:

inputting the generated customized base text into the trained correct answer learning model,

calculating a correct answer probability for the generated customized base text, and

determining the generated customized base text as verified content only when the calculated correct answer probability is less than a predetermined threshold;

generating customized hearing loss rehabilitation content based on the verified customized base text; and

providing the customized hearing loss rehabilitation content to the user,

wherein the acquisition of the user response data comprises generating and storing problem pattern data for the sentence-type rehabilitation content.

2 . The method of claim 1 , wherein the sentence-type rehabilitation content is question-and-answer type learning content that asks questions based on the audio quiz.

3 . The method of claim 1 , wherein the provision of the sentence-type rehabilitation content comprises outputting the audio quiz a predetermined number of times based on a sentence, and

wherein the acquisition of the user response data comprises acquiring single user response data.

4 . The method of claim 1 , wherein the provision of the sentence-type rehabilitation content comprises outputting the audio quiz a predetermined number of times based on a plurality of sentences, and

wherein the acquisition of the user response data comprises a plurality of pieces of user response data.

5 . The method of claim 1 , wherein the provision of the sentence-type rehabilitation content comprises providing problem guidance text, an audio quiz start button, and, multiple choice selection items.

6 . The method of claim 5 , wherein the provision of the sentence-type rehabilitation content further comprises providing wrong answer guidance text, a skip button, and a restart button when the user response data is processed as a wrong answer.

7 . The method of claim 1 , wherein the correct answer learning model, when a predetermined text is input, outputs:

(i) correct answer pattern information including text structure arrangement information and grapheme combination structure information based on the predetermined text; and

(ii) correct answer probability information indicating a probability that the user will correctly answer an audio quiz corresponding to the predetermined text.

8 . The method of claim 1 , wherein the wrong answer learning model, when a predetermined text is input, outputs:

(i) wrong answer pattern information that models an analyzed type of wrong answer by grapheme, based on text structure arrangement information and grapheme combination structure information of the predetermined text; and

(ii) wrong answer probability information indicating a probability that the user will incorrectly answer an audio quiz corresponding to the predetermined text.

9 . The method of claim 1 , wherein the problem pattern data includes a content identification code, base text, and analysis data.

10 . The method of claim 9 , wherein the base text is text data that serves as the basis for forming the audio quiz provided based on the sentence-type rehabilitation content.

11 . The method of claim 9 , wherein the analysis data includes a morpheme analysis corpus list according to the base text and grapheme separation data according to the morpheme analysis corpus, as analyzing the structure of the base text.

12 . The method of claim 11 , wherein the grapheme separation data includes a plurality of grapheme-separated morphemes and positional information for each of the grapheme-separated consonants or vowels.