KR102406560B1 - Method, apparatus and system for improving accuracy volume of outbound call for dementia test of subject based on artificial intelligence - Google Patents

Abstract

An outbound call processing method, apparatus, and system for improving the accuracy of an AI-based patient's dementia test are provided. An outbound call processing method for improving accuracy of an AI-based dementia test for a subject in the electronic device includes: transmitting a prior notice message including the dementia test schedule to a terminal of the subject; transmitting an outbound call to the terminal of the target by configuring the target customized guide message based on the artificial intelligence-based generated learning model; receiving a response to the guide message of the transmitted target-customized guide message; detecting and processing noise in the response; and terminating the outbound call transmission.

Description

{METHOD, APPARATUS AND SYSTEM FOR IMPROVING ACCURACY VOLUME OF OUTBOUND CALL FOR DEMENTIA TEST OF SUBJECT BASED ON ARTIFICIAL INTELLIGENCE}

The present invention relates to a dementia test of a subject, and more particularly, to an outbound call processing method, apparatus and system for improving the accuracy of a subject's dementia test based on artificial intelligence.

In general, dementia tests are performed in the order of a Cognitive Impairment Screening Test (CIST), a diagnostic test, and a differential test.

Screening test is a method in which test subjects directly visit dementia safety centers and public health centers in each region for the test. The diagnostic test is a method in which a neurocognitive test is performed at a dementia safety center or a convention hospital or treatment is performed by a specialist according to the results of the screening test. Lastly, the differential test is a method of finally discriminating whether dementia is present through blood tests and brain imaging at a convention hospital according to the results of the diagnostic test.

In particular, in relation to the screening test, according to the conventional method, if the test subjects directly visit the relevant institution for the screening test, the likelihood of being a dementia patient is relatively low, so there is a problem with the necessity or effectiveness of the test. In addition, there is a problem in that efficiency in terms of manpower, cost, time, etc. is lowered as a person directly answers questions in a relationship with a test subject in performing a screening test at a relevant institution. Such efficiency may be further reduced when considering the necessity or effectiveness of the aforementioned screening test itself.

Patent Document 1 relates to a method and apparatus for diagnosing dementia based on language fluency, and discloses a general test method related to a dementia test. However, despite the diverse backgrounds of dementia subjects, only the content of the dementia test is disclosed based on the language fluency, so accuracy There is a problem or a method to improve it needs to be proposed.

Korean Patent Publication No. 10-1437569 (2014.09.04)

An object of the present invention is to provide a method, apparatus and system for improving the accuracy of a test by removing noise from a response to an AI-based outbound call prior to a screening test for determining whether dementia is present.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

An outbound call processing method for improving accuracy of a dementia test for a subject based on artificial intelligence in an electronic device according to an aspect of the present invention for solving the above-mentioned problems, a prior notification message including the dementia test schedule to the terminal of the subject sending a; transmitting an outbound call to the terminal of the target by configuring the target customized guide message based on the artificial intelligence-based generated learning model; receiving a response to the guide message of the transmitted target-customized guide message; detecting and processing noise in the response; and terminating the outbound call transmission.

An artificial intelligence-based outbound call processing apparatus for improving the accuracy of a dementia test of a subject according to an aspect of the present invention includes: a memory; and a processor, wherein the processor transmits a prior notice message including the dementia test schedule to the terminal of the subject, and configures the subject-specific guidance message based on the artificial intelligence-based generated learning model to configure the subject It is possible to transmit an outbound call to a terminal of , receive a response to the guidance message of the transmitted target-customized guidance message, detect and process noise in the response, and control to terminate the outbound call transmission.

An artificial intelligence-based outbound call processing system for improving accuracy of a subject's dementia test according to an aspect of the present invention comprises: a subject terminal; and a server, wherein the server transmits a prior notice message including the dementia test schedule to the terminal of the subject, and configures the subject-specific guidance message based on the artificial intelligence-based generated learning model to configure the subject A processor for transmitting an outbound call to a terminal of a, receiving a response to the guidance message of the transmitted target-customized guidance message, detecting and processing noise in the response, and controlling to terminate the outbound call transmission can

Other specific details of the invention are included in the detailed description and drawings.

According to the present invention, it can have the following effects.

According to the present invention, it is possible to increase the accuracy and reliability of the test by appropriately removing noise in the test subject's response in the process of performing a pre-test through an artificial intelligence-based outbound call prior to a screening test for determining whether dementia is present.

Effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a block diagram illustrating an outbound call processing system for improving accuracy of an AI-based test subject for dementia according to an embodiment of the present invention.
2 is a configuration block diagram of a server according to an embodiment of the present invention.
3 is a flowchart illustrating an outbound call response processing method for improving accuracy of an AI-based dementia test according to an embodiment of the present invention.
FIG. 4 is a flowchart illustrating the noise processing method of FIG. 3 .
5 is a flowchart illustrating a response analysis processing method through partial extraction of a response according to an embodiment of the present invention.

Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only the present embodiments allow the disclosure of the present invention to be complete, and those of ordinary skill in the art to which the present invention pertains. It is provided to fully understand the scope of the present invention to those skilled in the art, and the present invention is only defined by the scope of the claims.

The terminology used herein is for the purpose of describing the embodiments and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise in the phrase. As used herein, “comprises” and/or “comprising” does not exclude the presence or addition of one or more other components in addition to the stated components. Like reference numerals refer to like elements throughout, and "and/or" includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various elements, these elements are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first component mentioned below may be the second component within the spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used herein will have the meaning commonly understood by those of ordinary skill in the art to which this invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless specifically defined explicitly.

Spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe the correlation between a component and other components. Spatially relative terms should be understood as terms including different directions of components during use or operation in addition to the directions shown in the drawings. For example, when a component shown in the drawings is turned over, a component described as "beneath" or "beneath" of another component may be placed "above" of the other component. can Accordingly, the exemplary term “below” may include both directions below and above. Components may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram illustrating an outbound call processing system for improving the accuracy of an AI-based examination for dementia test subjects according to an embodiment of the present invention, and FIG. 2 is a server 150 according to an embodiment of the present invention. is a block diagram of the

The present invention relates to a dementia test, and noise may be included in the response of the test subject due to various variables in the analysis process of the dementia test. Therefore, in the present invention, by preprocessing and removing noise in the response, it is possible to accurately analyze the response, thereby improving the accuracy and reliability of the dementia test itself.

Referring to FIG. 1 , an outbound call processing system according to an embodiment of the present invention may be configured to include a target terminal 100 and a server 150 . At this time, according to an embodiment, the outbound call processing system includes one or more components in connection with performing an operation to be described later according to the present invention in addition to the terminal to be inspected 100 and the server 150 shown in FIG. 1 . It may be added and configured. For example, in relation to transmission and reception of an outbound call to the terminal 100 by the server 150, related configurations such as a communication server and a base station may be included in the outbound call processing system according to the present invention.

At this time, various embodiments of the present invention will be described by taking the case of an outbound call as an example for better understanding of the present invention and for convenience of explanation, but the present invention is not limited thereto. Therefore, for example, even in the case of an inbound call, among the contents of the present invention, for example, contents related to response rejection, which will be described later, may be applicable in the same or similar manner.

The test subject terminal 100 is an electronic device that receives an outbound call transmitted by the server 150 , and may receive a guide message according to the outbound call transmission and transmit a response signal thereto. In this case, the response signal may include the test subject's voice. However, the response signal may include noise that can be defined in various ways other than the voice. The noise will be described later. However, the response signal may include, for example, other types of data other than the above-described voice according to the shape of the terminal, a user interface (UI) provided to the terminal, or the like, or may be configured in a combined form.

The test target terminal 100 may be any one of a TV, a monitor, a signage, a PC, a laptop computer, a tablet pc, a smart phone, a wearable device, an AI speaker, and the like. However, for the purpose of helping the understanding of the present invention and for convenience of explanation, the test subject terminal 100 is described using a smart phone as an example, but is not limited thereto.

In relation to the present invention, the test subject terminal 100 may include an interface related to reception and response of the dementia test outbound call. For such an interface, for example, a pre-test for determining whether dementia according to the present invention is performed through an outbound call, and a speaker, a microphone, etc. related to voice signal processing may be used as an example. At this time, it is sufficient that the test subject terminal 100 is connected or paired with the interface even if it has or does not directly have the interface.

Meanwhile, according to another embodiment, the test subject terminal 100 includes a display for outputting a voice of a pre-test outbound call for determining whether dementia is present as a text, and touches a response to the text output on the display. An interface that converts an input to voice and transmits the input through a voice, etc., and software/hardware for STT (Speech to Text) and TTS (Text to Speech) processing for the above processing may be provided or connected and used.

In addition, according to an embodiment, in response to the above-described outbound call for the dementia test, a method such as a gesture, a touch, etc. may be combined with a voice, and a configuration for that may be further included.

The server 150 transmits a prior notice message about the dementia test to the test subject terminal 100, and based on the artificial intelligence-based generated learning model, a first message for preventing reception of the subject and the subject based on the artificial intelligence-based learning model An outbound call can be transmitted by composing a target-customized guide message including a second message for preventing response rejection.

The server 150 may receive the target's voice responses to the first message and the second message of the target customized guide message according to the outbound call transmission, and analyze the received voice response.

The server 150 may remove noise in the voice response with reference to the analysis result.

The server 150 may terminate the outbound call transmission.

Referring to FIG. 2 , the server 150 may include a memory 210 and a processor 220 .

The memory 210 may store various training data sets related to a pre-test for AI-based dementia determination, noise, and the like, and a learning model generated using the training data.

The memory 210 is collected by the server 150, such as data related to the prior notification message for the pre-test for the dementia determination test, data related to the target customized guidance message, data related to the noise related to the response of the target customized guidance message, etc. , processing, processing, etc. various data can be stored.

The processor 220 may transmit, to the test subject terminal 100 , an outbound call including a prior notice message related to a pre-test for determining whether dementia is present and a subject-specific guide message configured based on an artificial intelligence-based learning model.

The processor 220 receives the transmitted advance notice message and the response of the outbound call, analyzes the received response, and determines whether or not to identify and remove noise in the response of the outbound call personalized guidance message and whether to pre-check. can do.

The processor 220 may terminate the transmission of the outbound call.

The server 150 may provide a part of the service related to the pre-test for determining whether dementia according to the present invention in the form of an application or a web service.

Although not shown in FIG. 1 , the server 150 may communicate with one or more external databases (DB) in order to collect various data related to a pre-test for determining whether dementia is present. The external database may include, for example, a server of a private/national institution related to domestic and foreign dementia-related medical services, a server of a medical professional institution, a database of a medical thesis service providing server, and the like.

3 is a flowchart illustrating an outbound call response processing method for improving accuracy of an AI-based dementia test according to an embodiment of the present invention. FIG. 4 is a flowchart illustrating the noise processing method of FIG. 3 . 5 is a flowchart illustrating a response analysis processing method through partial extraction within a response according to an embodiment of the present invention.

The operations of FIGS. 3 to 5 may be performed by the server 150 of FIG. 1 , in particular, the processor 220 of FIG. 2 .

First, with reference to FIG. 3 , an outbound call volume control method for improving accuracy of an AI-based test subject's dementia test according to an embodiment of the present invention will be described as follows.

In operation 11 , the server 150 may transmit a prior notice message including the dementia test schedule to the test subject terminal 100 .

In operation 12, the server 150 may configure and transmit the target-customized guide message based on the artificial intelligence-based generated learning model.

In operation 13, the server 150 may receive a response to the guide message of the transmitted personalized guide message.

In operation 14, the server 150 may determine whether noise is detected in the response.

In operation 15 , as a result of the determination in operation 14 , if noise in the response is detected, the server 150 may remove the detected noise through preprocessing.

In operation 16 , the server 150 may terminate the outbound call transmission, generate test result data and transmit it to the test subject terminal 100 .

FIG. 4 can be seen as a detailed operation of operations 14 and 15 in FIG. 3 .

Meanwhile, in FIG. 4 , for example, a voice by a third party may be determined as noise.

Referring to FIG. 4 , in operation 21 , the server 150 may identify the type of noise detected in the response.

A pre-processing operation is different according to the identified noise type.

In operation 22-1, if the identified noise type is the first type, the server 150 may analyze the response to calculate an analysis result.

In operation 23, the server 150 may measure the test environment score.

In operation 24, the server 150 may calculate a reliability value based on the measured test environment score.

In operation 25, the server 150 may determine whether the calculated reliability value is equal to or less than a threshold value.

In operation 26, as a result of the determination in operation 25, if the calculated reliability value is equal to or less than a threshold value, the server 150 may control to perform any one of evaluation criteria adjustment, score correction, and retesting.

On the other hand, in operation 22-2, if the identified noise type is the second type, the server 150 may control to reconstruct and retransmit the advance notice message for retesting or retesting scheduling.

Also, in operation 22-3, if the identified noise type is the third type, the server 150 may control the response to be preprocessed after removing the detected noise.

The server 150 needs to prevent the test result from being changed due to the influence of a third party other than the test subject. For example, the server 150 may affect the inspection accuracy when the subject to be inspected conducts the inspection while consulting or talking with a third party, or when a third party puts a hunsu next to him. You may need to be prepared.

In this regard, the server 150 may first detect whether a third party exists. To this end, the server 150 determines whether or not two or more voices are detected in the response based on the voice waveform analysis, and if the subject has an existing test history, compares the test subject's voice and the test with existing test data if the test subject has an existing test history Voices other than the subject can be discriminated and detected.

When the third party's voice is detected, the server 150 may process it as follows. However, the present invention is not limited thereto.

First, like the first type, the server 150 may calculate an examination result in consideration of an examination environment analysis and examination environment (operation 22-1). This is to prevent a case in which the test result is calculated higher or lower than the original due to the influence of the surrounding environment, that is, different from the original. Accordingly, the server 150 measures the test environment score (operation 23), calculates a reliability value in consideration of the environment (operation 24), determines whether the calculated reliability value is equal to or less than a threshold value (operation 25), and the determination If the calculated reliability value is less than or equal to the threshold value, the evaluation criteria may be adjusted, the score may be corrected, or a control may be performed to perform a retest (operation 26).

Meanwhile, in the above, as an environment that interferes with the examination progress, not only the presence of the aforementioned third party, but also a location in an inappropriate place (eg, public transportation, outside, etc.) for the examination, TV/radio sound, etc. may be considered.

Next, as in the second type, when two or more voices are detected, the server 150 may proceed with a retest or a retest schedule reservation procedure (operation 22-2).

Finally, like the third type, the server 150 may perform preprocessing for detecting and removing a third party's voice noise (eg, top, hunsu, conversation, etc.) in the response audio (operation 22-3). ).

On the other hand, when the video call is applied, the server 150 may recognize and process the presence of the aforementioned third party, whether there is an inappropriate place for the examination, etc. by using the camera function.

FIG. 5 may be performed together with or separately from FIG. 4 .

Referring to FIG. 5 , in operation 31 , the server 150 may extract a portion in the received response.

In operation 32, the server 150 may determine whether a duplicate word is included in the part in the extracted response.

If a duplicate word is included in the extracted response and the duplicate word is the first duplicate type, the server 150 may noise-process it (operation 35).

On the other hand, if the duplicated word is included in the extracted response and the duplicated word is the second duplicated type, the server 150 may apply the evaluation criteria (operation 36).

In this case, the determination of whether or not the duplicate word is included may be, for example, to distinguish noise caused by the speech habit of the dementia test subject.

In the above, the case of answering consecutively duplicated words as a duplicate answer is defined as a first duplicate type, and a case of answering non-consecutively duplicated words is defined as a second duplicate type, and the server 150 Different processing may be performed for the overlapping type and the second overlapping type.

In relation to the first duplication type, the server 150, for example, 'squirrel, squirrel, squirrel... ', it can be judged as a linguistic habit, such as muttering while thinking about the next word, so that it can be processed as noise rather than incorrect answer processing.

On the other hand, the server 150, in relation to the second overlap type, for example, 'squirrel, cow, puppy, squirrel... ', as it is a non-continuous, repeated answer without remembering the same word, it is not treated as noise, and evaluation criteria such as incorrect answer processing, deduction of points, or small points can be applied.

On the other hand, the server 150 may process a voice determined as a language habit rather than for the purpose of telling the correct answer as noise. For example, the server 150 may determine that 'I did it before' is not an incorrect answer, but corresponds to a language habit.

Meanwhile, in operation 33, the server 150 may determine whether a noun is included in the part in the extracted response.

As a result of the determination in operation 33, the server 150 may perform noise processing if a noun is not included in the part in the extracted response (operation 35). In this case, if it is not a noun in the above, the corresponding part can be viewed as a verb.

On the other hand, as a result of the determination in operation 33, if a noun is included in the part in the extracted response, the server 150 may next determine whether the noun corresponds to a dialect (operation 34). Operation 34 may be performed separately from the determination of the dialect according to the inclusion of the noun in the server 150 as follows.

The server 150 may distinguish a voice into a noun and a non-noun, and may process a non-noun voice (eg, 'something is there', 'it's difficult', etc.) as noise.

Meanwhile, in operation 34, the server 150 may determine whether a dialect is included in the extracted response.

As a result of the determination in operation 34, the server 150 may noise-process if a dialect is included in the extracted response (operation 35).

As a result of the determination in operation 34, the server 150 may apply the evaluation criteria if a dialect is not included in the extracted response (operation 36).

The server 150 can improve the accuracy of the dementia test only by preventing the test itself from raising a problem. For example, the server 150 should be able to determine the repeated speaking habit of the test subject, and should be able to identify and pre-process parts for dialects and sounds for animal names, etc.

Since the server 150 can speak different names for the same animal, plant, etc. according to age, region of origin, etc., it is possible to establish a regional dialect database to more accurately determine whether the answer is correct.

Although in FIG. 5 , if a dialect is included in the extracted response, the server 150 performs noise processing, but the present invention is not limited thereto. According to an embodiment, the server 150 collects various data on regional dialects, creates an artificial intelligence learning model using the collected data as training data, and machine learning the generated learning model to build a dialect database. can Therefore, the server 150 does not unconditionally process noise when a dialect is included in the part extracted in the response, but first determines whether data corresponding to the built dialect database is stored, and, nevertheless, the corresponding dialect cannot be interpreted If it is determined that , noise may be processed at that time.

Similar to the above-mentioned dialect, the old sound can also be processed in the manner described above.

In extracting the part in the response, when the text is extracted by performing STT processing on the response voice data, the server 150 may divide the entire text into preset sizes to define the part. For example, when there are 10 characters in one text, the server 150 may define a part by dividing the 10 characters in 2 to 3 units, for example. At this time, the portion corresponding to the space may be different in the normal voice and the text restored after STT processing. can

Meanwhile, as a result of determination in operation 33, the server 150 performs noise processing when the corresponding part is not a noun, but the present invention is not limited thereto. According to an embodiment, the server 150 pre-defines words that can be interpreted as meaningful even for verbs, collects them, and uses the collected data as training data to generate an artificial intelligence learning model, and thus the generated learning model By machine learning, it is possible to build a verb database. Therefore, the server 150 does not unconditionally process noise when a verb is included in the part extracted from the response above, but first determines whether data corresponding to the constructed verb database is stored, and if it is not included in the database, Only then can the noise be processed.

On the other hand, if the server 150 determines in operation 33 that the corresponding part is not a noun, the server 150 collects voice tone-based data for the verb and uses the collected data as training data for artificial intelligence. By creating a learning model and machine learning the learning model thus generated, a voice tone database can be built. Therefore, the server 150 does not unconditionally process noise when a verb is included in the part extracted from the response, but first determines whether data corresponding to the built-up voice tone database is stored, and if it is not included in the database Only then can noise be processed.

In addition, the server 150 may obtain location information such as GPS and motion data of the terminal 100 to be tested and use it to improve the accuracy of the dementia test. For example, a vehicle that determines a situation where the subject to be inspected is located outside such as on the street or the subject is moving through environmental information through location information and motion data (in the case of video calls, camera sensors, etc.) and responds accordingly It can detect and process various noises caused by sounds and public transportation sounds.

In general, when the STT (Speech to Text) technology is applied to the response voice, the server 150 tends to recognize the voice as a sentence. Accordingly, in consideration of this point, the server 150 may pre-process the content included in the voice audio into a noun or perform pre-processing by first extracting a part corresponding to the noun.

4 and 5 , it is possible to improve the accuracy of the test data-based dementia test according to an embodiment of the present invention. For example, the server 150 may transmit test data, receive a response, select a database suitable for a test subject, or control to remove noise that may be generated in advance.

On the other hand, the server 150 transmits the volume level setting test data including the voice test data set to the preset individual volume level to the test subject terminal 100, and sets the volume level at which the noise is minimized to improve the examination accuracy. may be

Although not shown, the server 150 may improve test accuracy by classifying voice data and non-voice data in the response and pre-processing the separated non-voice data.

The operation sequence shown in FIGS. 3 to 5 may be processed differently from the sequence shown in the corresponding drawings. For example, some operations may be performed concurrently.

The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, as a software module executed by hardware, or by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any type of computer-readable recording medium well known in the art to which the present invention pertains.

In the above, embodiments of the present invention have been described with reference to the accompanying drawings, but those of ordinary skill in the art to which the present invention pertains can realize that the present invention can be embodied in other specific forms without changing the technical spirit or essential features thereof. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

100: terminal to be tested
150 : server
210: memory
220: processor

Claims (10)

In an outbound call processing method for improving the accuracy of dementia test of subjects based on artificial intelligence in an electronic device,
transmitting a prior notice message including the dementia test schedule to the terminal of the subject;
transmitting an outbound call to the terminal of the target by configuring the target customized guide message based on the artificial intelligence-based generated learning model;
receiving a response to the guide message of the transmitted target-customized guide message;
detecting and processing noise in the response; and
Doedoe comprising the step of terminating the outbound call transmission,
The step of detecting and processing noise in the response comprises:
identifying the type of noise detected in the response;
if the identified noise type is the first type, analyzing the response to calculate an analysis result;
measuring the test environment score;
calculating a reliability value based on the measured test environment score;
determining whether the calculated reliability value is less than or equal to a threshold value; and
If the reliability value calculated as a result of the determination is less than or equal to a threshold value, comprising controlling to perform any one of evaluation criteria adjustment, score correction, or retest progress,
How to handle outbound calls. delete delete delete According to claim 1,
If the identified noise type is the second type, reconstructing and retransmitting the advance notice message for retesting or retesting scheduling progress,
How to handle outbound calls. 6. The method of claim 5,
If the identified noise type is a third type, controlling the response to be preprocessed after removing the detected noise,
How to handle outbound calls. 7. The method of claim 6,
further comprising extracting a portion in the received response,
How to handle outbound calls. 8. The method of claim 7,
Control that the part in the extracted response is a duplicate word type, and noise processing or evaluation criteria is applied according to the type,
If the part in the extracted response does not correspond to a noun, it is controlled to be noise-processed,
Controlling that the part in the extracted response corresponds to a noun, but noise processing or evaluation criteria is applied depending on whether it is a dialect,
How to handle outbound calls. In the outbound call processing device for improving the accuracy of dementia test of subjects based on artificial intelligence,
Memory; and
A processor comprising:
Transmitting a prior notice message including the dementia test schedule to the subject's terminal, composing the subject customized guide message based on the artificial intelligence-based generated learning model, and sending an outbound call to the subject's terminal, Control to receive a response to the guidance message of the transmitted target-customized guidance message, detect and process noise in the response, and terminate the outbound call transmission,
When detecting and processing noise in the response, the type of noise detected in the response is identified, and if the identified noise type is the first type, the response is analyzed to calculate an analysis result, and the test environment score is measured calculating a reliability value based on the measured test environment score, and if the calculated reliability value is less than or equal to a threshold value, controlling to perform any one of evaluation criteria adjustment, score correction, or retesting;
Outbound call processing unit. In the outbound call processing system for improving the accuracy of dementia test of subjects based on artificial intelligence,
subject terminal; and
Including a server, wherein the server,
Transmitting a prior notice message including the dementia test schedule to the subject's terminal, composing the subject customized guide message based on the artificial intelligence-based generated learning model, and sending an outbound call to the subject's terminal, A processor for receiving a response to the guidance message of the transmitted target-customized guidance message, detecting and processing noise in the response, and controlling to terminate the outbound call transmission,
The processor, when detecting and processing noise in the response, identifies a type of noise detected in the response, and if the identified noise type is a first type, analyzes the response to calculate an analysis result, and an inspection environment To calculate a reliability value based on the measured test environment score by measuring a score, and if the calculated reliability value is less than or equal to a threshold value, controlling to perform any one of evaluation criteria adjustment, score correction, or retest progress,
Outbound call handling system.

Images