Evaluation score filtering method using mismatch

Abstract

The present disclosure relates to a method for filtering evaluation scores by using a mismatch between evaluation content and the evaluation score. According to some embodiments of the present disclosure, an evaluation score filtering method using a mismatch may be provided, the evaluation score filtering method comprising the steps of: acquiring first evaluation content and a first evaluation score related to the first evaluation content; acquiring a second evaluation score by performing sentiment analysis on the first evaluation content; determining whether a mismatch has occurred between the first evaluation score and the second evaluation score; if it is determined that the mismatch has occurred, adjusting the first evaluation score; and storing the adjusted first evaluation score.

Description

Evaluation score filtering method using mismatch

[0001] The present disclosure relates to a method for filtering evaluation scores by utilizing a mismatch between evaluation content and an evaluation score. More specifically, the present disclosure relates to a method for performing sentiment analysis on evaluation content input from a user terminal to derive a sentiment score, and adjusting the evaluation score if a mismatch occurs between the derived sentiment score and the evaluation score. Furthermore, the present disclosure relates to a method for improving the discriminatory power of the evaluation score by deleting or adjusting the evaluation score of the user terminal in the event of a mismatch between the location of the user terminal and the location where the service is provided, thereby ensuring a high level of reliability in the evaluation score of the service.

[0002] A rating score is a numerical or other representation of a user's experience and satisfaction with a product, service, or content. A rating score can serve as valuable information for other users or consumers when choosing a product or service. For example, user reviews, star ratings, and satisfaction levels can be included in a rating score.

[0003] Evaluation content refers to users expressing their experiences and satisfaction with products, services, or content using text, audio, or images, rather than numbers. Evaluation content can serve as information to confirm actual evaluations when evaluation scores are difficult to interpret. For example, a caregiver matching platform might award a maximum score of 5 points and a one-line review stating, "You worked hard, but there were some inconveniences." While the evaluation score may not reflect the actual evaluation, the one-line review can be considered to reflect the actual evaluation.

[0004] Sentiment analysis is a branch of natural language processing (NLP) that analyzes text content to identify and classify the emotions and opinions contained within it. Sentiment analysis can determine whether a text is positive, negative, or neutral. Sentiment analysis is often used to assess customer satisfaction. For example, if a service user writes a review, a computing device can perform sentiment analysis on the review to determine whether the user is satisfied or dissatisfied with the service.

[0005] Meanwhile, evaluation content can be quantified using sentiment analysis. For example, if a caregiver matching platform includes a one-line review saying, "You worked hard, but there were some inconveniences," sentiment analysis could assign a score of 1 for "Friendliness," -1 for "Skill," and 0 for "Communication."

[0006] The quantified sentiment analysis of evaluation content can be used to adjust evaluation scores. This adjustment ensures that evaluation scores reflect actual evaluations and increases the reliability of the evaluations.

[0007] The background technology described above is technical information that the inventor possessed for the purpose of deriving the present disclosure or acquired during the process of deriving the present disclosure, and cannot necessarily be said to be technology known to the general public prior to the filing of the present disclosure.

[0008] A technical problem to be achieved through some embodiments of the present disclosure is to provide a method and device for performing sentiment analysis on evaluation content and adjusting an evaluation score related to the evaluation content.

[0009] The problem to be solved through some embodiments of the present disclosure is to provide a method and device for performing sentiment analysis on evaluation content to produce a sentiment score reflecting an actual evaluation.

[0010] Another technical task to be achieved through some embodiments of the present disclosure is to create a sentiment dictionary specialized for a specific domain and to provide a method for performing sentiment analysis specialized for a specific domain.

[0011] Another technical challenge to be achieved through some embodiments of the present disclosure is a method for sentiment analysis of data other than text data to make the evaluation score reflect the actual evaluation score.

[0012] Another technical problem to be achieved through some embodiments of the present disclosure is a method for filtering evaluation scores using the location of a service evaluator and the location of a service provider.

[0013] The technical problems of the present disclosure are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art of the present disclosure from the description below.

[0014] According to one embodiment of the present disclosure, a method for filtering evaluation scores using mismatch is provided. The method for filtering evaluation scores according to the present embodiment may include the steps of: obtaining a first evaluation content and a first evaluation score related to the first evaluation content; performing a sentiment analysis on the first evaluation content to obtain a second evaluation score; determining whether a mismatch occurs between the first evaluation score and the second evaluation score; adjusting the first evaluation score if the occurrence of the mismatch is determined; and storing the adjusted first evaluation score.

[0015] In one embodiment, the first evaluation content may be evaluation content for a matched caregiver input through a caregiver matching platform.

[0016] In one embodiment, the step of analyzing the sentiment and obtaining a second evaluation score comprises:

[0017] The method may include a step of sentiment analysis of at least one of the personal information of the patient of the caregiver, the personal information of the user of the caregiver, the consultation content of the user, and the user's access record to the caregiver matching platform, together with the first evaluation content.

[0018] In one embodiment, the step of obtaining the first evaluation score may further include a step of adjusting the first evaluation score by considering at least one of an evaluation scoring tendency according to a relationship between a user of the caregiver matching platform and the patient, an evaluation scoring tendency according to the difficulty of caregiving for the patient, and an evaluation scoring tendency according to the health status of the patient.

[0019] In one embodiment, the sentiment analysis may be a sentiment analysis using a sentiment lexicon generated by obtaining a correlation coefficient between a first word and a first sentiment using a plurality of nursing evaluation contents and a plurality of nursing evaluation scores related to the plurality of nursing evaluation contents, calculating a sentiment score of the first word proportional to the correlation coefficient if the correlation coefficient of the first sentiment is greater than or equal to a threshold, and calculating a sentiment score of the first word as 0 if the correlation coefficient of the first sentiment is less than the threshold.

[0020] According to another embodiment of the present disclosure, the evaluation score filtering method according to the present embodiment may include the steps of: obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching through a service matching platform, and a first evaluation score related to the first evaluation content; obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; determining whether there is a mismatch between the geographic location information of the user terminal and the service provision location; and deleting the first evaluation content and the first evaluation score when it is determined that the mismatch has occurred.

[0021] In one embodiment, the step of determining whether there is a mismatch may include the step of determining geographic location information of the user terminal by using the access location of the user terminal to the service match platform during the service provision period of the first matching.

[0022] According to another embodiment of the present disclosure, an evaluation score filtering method according to the present embodiment may include the steps of: obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform, and a first evaluation score related to the first evaluation content; obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; determining whether a mismatch occurs between the geographic location information of the user terminal and the service provision location; determining whether a user of the user terminal cannot access the service provision location if it is determined that the mismatch does not occur; and deleting the first evaluation content and the first evaluation score if it is determined that the user cannot access the service provision location.

[0023] According to another embodiment of the present disclosure, an evaluation score filtering method according to the present embodiment may include a step of obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching through a service matching platform and a first evaluation score related to the first evaluation content, a step of obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information, a step of determining whether there is a mismatch between the geographic location information of the user terminal and the service provision location, a step of adjusting the first evaluation score when the occurrence of the mismatch is determined, and a step of storing the adjusted first evaluation score.

[0024] FIG. 1 is a configuration diagram of an evaluation score filtering system using mismatch according to one embodiment of the present disclosure.

[0025] FIG. 2 is a specific configuration diagram related to a user terminal in an evaluation score filtering system using mismatch according to one embodiment of the present disclosure.

[0026] FIG. 3 is a specific configuration diagram related to a server and a database in an evaluation score filtering system using mismatch according to one embodiment of the present disclosure.

[0027] FIG. 4 is a flowchart of a method for filtering evaluation scores using mismatch according to one embodiment of the present disclosure.

[0028] Figure 5 is an example of a step of obtaining a care evaluation score and care evaluation content described with reference to Figure 4.

[0029] Figure 6 is an example of a step of sentiment analysis and calculation of a sentiment score for the care evaluation content described with reference to Figure 4.

[0030] Figure 7 is an example of a method for sentiment analysis of care evaluation content described with reference to Figure 4.

[0031] FIG. 8 is an example of a method for sentiment analysis of data not input by a user in a method for filtering evaluation scores using mismatch according to one embodiment of the present disclosure.

[0032] Figure 9 is an example of a step for determining whether there is a mismatch between the emotional score calculated by sentiment analysis of the care evaluation content described with reference to Figure 4 and the care evaluation score.

[0033] Figure 10 is an example of a step for adjusting the nursing evaluation score described with reference to Figure 4.

[0034] FIG. 11 is a flowchart of a method for adjusting a nursing evaluation score using a scoring tendency in a method for filtering evaluation scores using mismatch according to an embodiment of the present disclosure.

[0035] Figure 12 is an example of scoring tendency data described with reference to Figure 11.

[0036] Figure 13 is an example of a step of first adjusting the nursing evaluation score using the scoring tendency described with reference to Figure 11.

[0037] Figure 14 is an example of a step of performing sentiment analysis using a sentiment dictionary specialized in the care domain in the sentiment analysis described with reference to Figure 4.

[0038] FIG. 15 is a flowchart of an evaluation score method for deleting a care evaluation score by utilizing a mismatch between a caregiver matching application access location and a care service provision location according to one embodiment of the present disclosure.

[0039] FIG. 16 is an example of a step for determining a caregiver matching application connection location by considering the duration of care service provision in one embodiment.

[0040] FIG. 17 is a flowchart of an evaluation score method for adjusting a care evaluation score by utilizing a mismatch between a caregiver matching application access location and a care service provision location according to one embodiment of the present disclosure.

[0041] FIG. 18 is an example of a step of adjusting a care evaluation score by utilizing a mismatch between a caregiver matching application access location and a care service provision location in one embodiment.

[0042] FIG. 19 is a hardware configuration diagram of a computing device that may be used as a component in some embodiments of the present disclosure.

[0043] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the attached drawings. Advantages and features of the embodiments of the present disclosure, and methods for achieving them, will become clear with reference to the embodiments described in detail below together with the attached drawings. However, the technical idea of ​​the present disclosure is not limited to the following embodiments and may be implemented in various different forms. The following embodiments are provided only to complete the technical idea of ​​the present disclosure and to fully inform those skilled in the art of the embodiments of the present disclosure of the scope of the present disclosure, and the technical idea of ​​the present disclosure is defined only by the scope of the claims.

[0044] When designating components in each drawing, it should be noted that identical components are assigned identical reference numerals whenever possible, even if they appear on different drawings. Furthermore, when describing the embodiments of this specification, detailed descriptions of related known configurations or functions are omitted if they are deemed to obscure the main point.

[0045] Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used in a sense commonly understood by those of ordinary skill in the art to which the embodiments of this specification pertain. Furthermore, terms defined in commonly used dictionaries shall not be interpreted ideally or excessively unless explicitly and specifically defined otherwise. The terminology used herein is for the purpose of describing embodiments and is not intended to limit the embodiments of this specification. In this specification, the singular also includes the plural unless specifically stated otherwise.

[0046] Additionally, terms such as first, second, A, B, (a), (b), etc. may be used to describe components of the embodiments of the present specification. These terms are only intended to distinguish the components from other components, and the nature, order, or sequence of the components are not limited by the terms. When it is described that a component is "connected," "coupled," or "connected" to another component, it should be understood that the component may be directly connected or connected to the other component, but another component may also be "connected," "coupled," or "connected" between each component.

[0047] Hereinafter, some embodiments are described in detail with reference to the attached drawings.

[0048] The method for filtering evaluation scores using mismatches according to some embodiments of the present disclosure and the device applying the method can be applied to any field in which a user (or guardian) inputs evaluation content and evaluation scores for a product provider (or caregiver). Here, the products to be provided by the product provider are not limited to goods and may also include services.

[0049] For example, the present disclosure may be applied to a field in which a patient's guardian inputs a review and rating of the patient's caregiving service, and the review is sentiment-analyzed and the rating is adjusted. For another example, the present disclosure may be applied to a field in which a patient inputs a review and rating of the patient's caregiving service, and the review is sentiment-analyzed and the rating is adjusted. For convenience of explanation, the present disclosure will be described below based on an example in which a patient's guardian inputs a review and rating of the patient's caregiver, and the review is sentiment-analyzed and the rating is adjusted. However, it should be noted that although the present disclosure is described based on this example, the scope of the present disclosure is not limited to the examples described below.

[0050] Referring to FIGS. 1 to 4, the configuration and operation of an evaluation score filtering system using mismatch according to one embodiment of the present disclosure are described.

[0051] First, a description will be given with reference to Fig. 1, which is a configuration diagram of an evaluation score filtering system using mismatch according to the present embodiment. As illustrated in Fig. 1, the evaluation score filtering system using mismatch according to the present embodiment may include a user terminal (1000), a network (2000), a server (3000), and a database (4000).

[0052] For a specific example, a caregiver (hereinafter referred to as a "caregiver") of a patient who is a user can input a review and rating of a caregiver (hereinafter referred to as a "caregiver") who is a product provider as care evaluation content (1100) and a care evaluation score (1200) in a caregiver matching application of a user terminal (1000). In addition, the input review and rating can be transmitted to a server (3000) via a network (2000). A sentiment analysis unit (3100) of the server (3000) can perform sentiment analysis on the review to calculate a sentiment score. In addition, a mismatch determination unit (3200) of the server can determine whether a mismatch occurs between the rating and the sentiment score. In addition, a score filtering unit (3300) of the server can adjust the rating for the care service so that the mismatch does not occur when the mismatch occurs. Additionally, the rating for the adjusted care service may be stored in the database (4000) as an adjusted care evaluation score (4300). Additionally, the rating for the adjusted caregiver may be displayed on the user terminal (1000).

[0053] For another example, the user terminal may transmit the caregiver matching application access location of the guardian and the patient's care location, which is the service provision location, to the server (3000) and database (4000) via the network (2000) as data (1300) not entered by the user. The mismatch determination unit (3200) of the server (3000) may determine whether a mismatch has occurred between the guardian's application access location and the care location. If the mismatch has occurred, the score filtering unit (3300) of the server (3000) may adjust or delete the rating for the care service so that the mismatch does not occur.

[0054] Reviews and ratings for the above care services are provided for your convenience and may be included in the care evaluation content and care evaluation score. Hereinafter, reviews and ratings will be referred to as "care evaluation content" and "care evaluation score," respectively.

[0055] In FIG. 1, an example is shown in which a user terminal (1000), a server (3000), and a database (4000) are connected to a network (2000), but this is only for convenience of understanding, and the number of user terminals (1000), servers (3000), and databases (4000) that can be connected to the network may vary.

[0056] Hereinafter, the components illustrated in Fig. 1 will be described in more detail with reference to Figs. 2 and 3.

[0057] As illustrated in FIG. 2, the user terminal (1000) can transmit care evaluation content (1100) and care evaluation score (1200) input by the user and data (1300) not input by the user to the server (3000) and database (4000) via the network (2000).

[0058] The care evaluation content (1100) may include text contents, audio contents, and image contents related to the care evaluation.

[0059] For specific examples, the text content may include sentences in which a caregiver evaluates a caregiver, reviews of care services, and one-line comments.

[0060] Additionally, the voice content may include a history of phone calls between a guardian and a caregiver, a history of consultations between a guardian and a caregiver, and voice content of a patient and a caregiver collected through voice recognition technology.

[0061] The above nursing evaluation content is not limited to these embodiments and may include all types of content that a user can input.

[0062] Next, the above-mentioned care evaluation score (1200) may include a score for evaluating the caregiver, a score for the care service, etc. In addition, the above-mentioned care evaluation score (1200) may have a score entered for each of one or more items.

[0063] For example, the care evaluation score may be a score consisting of an integer between 1 and 5 for evaluating the caregiver. Here, a higher number of the score may indicate a better evaluation of the caregiver. In addition, the care evaluation score may be a score consisting of an integer between 1 and 5 entered for each item including kindness, responsibility, cleanliness, communication, and skill. Here, the guardian may enter a high number for the kindness item and a low number for the skill item for a caregiver who is kind but not very skilled.

[0064] The above nursing evaluation score is not limited to these embodiments and may include any type of score that a user can input.

[0065] Next, the data (1300) not entered by the user may include text data, voice data, image data, and context data related to the nursing evaluation. Here, context data refers to data occurring in a specific situation or environment.

[0066] For example, data not entered by the user may include the personal information of the guardian (name, age, gender, etc.), the relationship between the guardian and the patient (mother and son, mother and daughter, father and son, father and daughter, husband and wife, siblings, etc.), the guardian's access records to the caregiver matching platform and application (access location, access time, access time, etc.), the patient's personal information (name, age, gender, etc.), the patient's physical characteristics (height, weight, etc.), the patient's symptoms (dementia, sleep disorder, bedsores, paralysis, ostomy, presence of infectious disease and details of infectious disease, etc.), the patient's health status (good, average, bad, etc.), the type of nursing service that should be provided to the patient (whether suction is needed, whether feeding is needed, whether mobility assistance is needed, whether toilet mobility assistance is needed, etc.), the difficulty level of the patient's nursing service (very easy, easy, average, difficult, very difficult, etc.), and the patient's nursing location, etc.

[0067] The above data not entered by the user is not limited thereto and may include all types of content related to caregiver evaluation that the user terminal can obtain without user input.

[0068] As illustrated in FIG. 3, the server (3000) may include a sentiment analysis unit (3100), a mismatch determination unit (3200), a score filtering unit (3300), a sentiment pre-adjustment unit (3400), and a score analysis unit (3500).

[0069] The sentiment analysis unit (3100) can perform sentiment analysis on the care evaluation content (1100) received from the user terminal and calculate a sentiment score of the caregiver evaluation content. Here, the sentiment score refers to a numerical value indicating the emotion of the content. For example, if the sentiment analysis unit analyzes the sentence "Thank you" and determines that it is very positive, the sentiment analysis unit can numerically assign a sentiment score of the sentence "Thank you" as +2.

[0070] The mismatch determination unit (3200) can determine whether a mismatch has occurred between the care evaluation score (1200) received from the user terminal and the care evaluation content (1100) received from the user terminal by performing a sentiment analysis on the calculated sentiment score and the care evaluation score. In addition, the mismatch determination unit can determine whether a mismatch has occurred between the caregiver matching application access location of the guardian and the patient's care location among the data received from the user terminal that was not input by the user.

[0071] In one embodiment, a mismatch may indicate a discrepancy between the caregiver evaluation score and the emotional score. For example, if a caregiver enters a maximum score of 5 for their evaluation of the caregiver and then writes a one-line comment, "You're working hard, but there were some areas that were dissatisfactory," a mismatch can be considered to have occurred because the evaluation score and the one-line comment do not match.

[0072] In one embodiment, a mismatch may be defined as a case where the distance between the caregiver's access location on the caregiver matching application and the caregiving location exceeds a threshold. For example, if the caregiver accesses the caregiver matching application from a location more than 100 km away from the patient's caregiving location, a mismatch between the caregiver's access location and the caregiving location can be considered to have occurred.

[0073] The score filtering unit (3300) can adjust or delete the care evaluation score if the mismatch determination unit determines that a mismatch has occurred. For example, if a caregiver inputs the maximum score of 5 as the care evaluation score and writes a one-line comment such as, "You're working hard, but there were some unsatisfactory aspects," and thus a mismatch has occurred, the score filtering unit can reduce the care evaluation score entered by the caregiver to less than 5. For another example, if a caregiver inputs the care evaluation score and care evaluation content at a location 100 km away from the patient's care location, and thus a mismatch has occurred, the score filtering unit can delete the care evaluation score.

[0074] The sentiment dictionary adjustment unit (3400) can adjust the sentiment dictionary content (4100) of the sentiment analysis performed by the sentiment analysis unit. Here, the sentiment dictionary (Sentiment Lexicon) may be a database that matches words with the emotions they represent to analyze and understand the emotional meaning of text content. The sentiment dictionary may need to be specialized for a specific domain. For example, while "proudly" generally corresponds to a positive emotion, in the caregiving domain, "proudly" may correspond to a negative emotion. For example, if a guardian writes in a review of a caregiver, "Proudly asked for money without doing any work," "proudly" may correspond to a negative emotion.

[0075] The score analysis unit (3500) can extract specific care evaluation scoring tendency data according to specific care evaluation content and store it in the database by using multiple care evaluation contents previously stored in the database (4000) and care evaluation scores related to the care evaluation contents. Here, the scoring tendency may refer to the tendency of evaluation users to input scores. Specifically, the scoring tendency may refer to the tendency of individual users, which are a specific group, to be related to a scoring method under a certain standard or situation. For example, the higher the difficulty of nursing a patient, the more likely the caregiver may be to input a lower care evaluation score. This tendency may be referred to as a scoring tendency for caregivers of patients with high care difficulty.

[0076] As illustrated in FIG. 3, the database (4000) may include emotional dictionary data (4100), scoring tendency data (4200), and adjusted caregiver rating scores (4300).

[0077] The emotional dictionary data (4100) may be data used as a basis for emotional analysis when the emotional analysis unit (3100) of the server (3000) performs emotional analysis. In addition, the emotional dictionary data may be adjusted by the emotional dictionary adjustment unit (3400) of the server (3000) to reflect previously stored nursing evaluation content and a nursing evaluation score related to the nursing evaluation content. In addition, the emotional dictionary data may include data that corresponds words related to the nursing evaluation content with the emotions expressed by the words in order to analyze and understand the emotional meaning of the nursing evaluation content.

[0078] The scoring tendency data (4200) may include scoring tendency data derived by analyzing nursing evaluation content stored in advance by the score analysis unit (3500) of the server (3000) and nursing evaluation scores related to the nursing evaluation content. In addition, the scoring tendency data may include data used for primarily adjusting the nursing evaluation score (1200) received from the user terminal by the score filtering unit (3600) of the server (3000). Here, primarily adjusting the nursing evaluation score may mean adjusting the nursing evaluation score (1200) to a certain range by using the previously stored scoring tendency data corresponding to the nursing evaluation content received from the user terminal or data not input by the user.

[0079] For example, scoring tendency data may represent scoring tendencies based on a user's rating history. If a user tends to mostly enter ratings between 3 and 5, this could be interpreted as a scoring tendency where the lowest rating is 3, the middle rating is 4, and the highest rating is 5.

[0080] For another example, scoring tendency data may represent scoring tendencies based on the caregiver-patient relationship. In a mother-child relationship, caregivers may have difficulty satisfying caregiving services, and thus may tend to enter more extreme scores than in a child-mother relationship. In a child-mother relationship, a caregiver may enter a score of 3 for the caregiver, and in a mother-child relationship, a caregiver may enter a score of 1 for the caregiver, resulting in the same evaluation score.

[0081] For another example, scoring tendency data may represent scoring tendencies based on the difficulty of caregiving. The higher the difficulty of caregiving, the more difficult it may be for caregivers to be satisfied with caregiving services, and thus, the lower the caregiving evaluation score may tend to be. Specifically, if caregivers tend to only enter a number between 1 and 3 when providing caregiving services, this could be interpreted as a scoring tendency where 1 is the lowest score, 2 is the middle score, and 3 is the highest score.

[0082] For another example, scoring trend data may represent scoring trends based on a patient's health status. The more critical a patient's health status is, the more difficult it is for caregivers to be satisfied with caregiving services, which may lead to lower evaluation scores. Specifically, if caregivers with critical health conditions tend to only input numbers between 1 and 3 when evaluating caregivers, this can be interpreted as a scoring trend where 1 represents the lowest score, 2 represents the middle score, and 3 represents the highest score.

[0083] The adjusted care evaluation score (4300) may be a care evaluation score adjusted by the score filtering unit (3300) of the server (3000). Furthermore, the adjusted care evaluation score may be used to calculate the average care evaluation score of the corresponding caregiver. For example, if a caregiver inputs the maximum score of 5 points along with a one-line comment, "You're working hard, but there are some areas where I'm dissatisfied," the score adjustment unit may adjust the care evaluation score from 5 points to 4 points. Here, the adjusted 4 points may be the adjusted care evaluation score.

[0084] So far, embodiments of components of an evaluation score filtering system utilizing mismatches have been described. Below, FIGS. 4 through 10 illustrate each step of a method for evaluating score filtering using mismatches according to another embodiment of the present disclosure.

[0085] In the following FIGS. 4 to 10, the explanation will continue assuming that each step of the methods is performed by the user terminal (1000), network (2000), server (3000), and database (4000) illustrated in FIGS. 1 and 3. However, for convenience of explanation, the description of the operating entity of each step included in the present disclosure may be omitted.

[0086] FIG. 4 is a flowchart illustrating a method for filtering evaluation scores using mismatches that can be performed in a user terminal, a network, and a server according to some embodiments of the present disclosure.

[0087] In step S100, the user terminal obtains care evaluation content and a care evaluation score. Next, the server receives and obtains the care evaluation content and the care evaluation score from the user terminal.

[0088] For a specific example, as illustrated in FIG. 5, a review (1101) by guardian X regarding caregiver A, such as "She is a really kind and good person who tries hard from the patient's perspective," and an item-by-item score (1201) such as "Kindness 5 points, responsibility 5 points, cleanliness 5 points, communication 5 points, and skill 5 points" can be entered on a user terminal (1001). Here, the review (1101) can be included in the care evaluation content (1100), and the item-by-item score (1201) can be included in the care evaluation score (1200).

[0089] Returning to Figure 4, the server can perform sentiment analysis on the care evaluation content through the sentiment analysis unit (3100) (S200).

[0090] For example, as illustrated in FIG. 6, the sentiment analysis unit (3100) can sentimentally analyze the review (1101) entered by guardian X. The sentiment analysis can be performed based on the sentiment dictionary content (4100) acquired from the database (4000). After completing the sentiment analysis, the sentiment analysis unit (3100) can calculate an item-by-item sentiment score (1111) corresponding to the review (1101) entered by guardian X. Here, the item-by-item sentiment score (1111) can be a score calculated by sentimentally analyzing the review (1101) entered by guardian X by dividing it into each of the emotions of kindness, responsibility, cleanliness, communication, and skill.

[0091] In one embodiment, as illustrated in FIG. 7, the sentiment analysis may include extracting words from text content, such as sentences, and calculating a sentiment score by summing the corresponding words. The method for extracting words may include extracting all duplicate words from the sentence.

[0092] For a specific example, FIG. 7 illustrates a method (3101) for extracting words that allow for word duplication. If a guardian writes a review saying, "Thank you, thank you, thank you so much," word extraction that allows for word duplication would be "Thank you, thank you, thank you so much, thank you."

[0093] In one embodiment, as illustrated in FIG. 7, the method for extracting the word may include extracting only the word with the largest absolute value of sentiment score and the longest word length among words with overlapping positions in the sentence.

[0094] For a specific example, FIG. 7 illustrates an extraction method (3102) that does not allow word position overlap. If a guardian writes a review saying, "You are so kind," the overlapping words in the positions of "You are so kind" may be "잘, 해, 할해, 주십니다, 할해십니다" in "잘하는곳." Here, considering the emotional scores in the emotional dictionary (4101), the words with the largest absolute values ​​of emotional scores are "잘해" and "잘해십니다.", and among them, the word with the longest word length is "잘해." Therefore, extracting the word with the largest absolute value of emotional score and the longest word length without allowing position overlap is to extract two words, such as "You are, so kind."

[0095] In one embodiment, the sentiment analysis may include a method of extracting text content from speech content, extracting words from the text content, and calculating a sum of scores of a sentiment dictionary corresponding to the words.

[0096] In one embodiment, the sentiment analysis may include a method of dividing the image content itself into categories such as kindness, responsibility, cleanliness, communication, and skill, and then analyzing the content to derive a sentiment score. The image content may include photographic content, graphic content, and the like.

[0097] In one embodiment, the sentiment analysis may include a method of calculating a sentiment score by sentiment-analyzing data that is not entered by a user other than the care evaluation content.

[0098] For example, as illustrated in FIG. 8, the user terminal (1000) can transmit the patient's personal information (1302-1), the guardian's personal information (1302-2), the guardian's consultation content (1302-3), and the guardian's access record (1302-4) among the data (1300) not entered by the user, together with the nursing evaluation content (1202), to the server (3000). The server's sentiment analysis unit (3100) can perform sentiment analysis on the nursing evaluation content, the patient's personal information, the guardian's personal information, the guardian's consultation content, and the guardian's access record together, and can calculate a sentiment score (1112).

[0099] The above sentiment analysis is not limited to these embodiments, and may include any type of sentiment analysis capable of producing a sentiment score from any form of content relevant to the evaluation.

[0100] Next, returning to Figure 4, the mismatch determination unit determines whether a mismatch has occurred between the above-mentioned care evaluation score and the above-mentioned calculated emotional score (S400). Here, a mismatch means that the tendencies between the care evaluation score and the emotional score are inconsistent.

[0101] For example, as illustrated in FIG. 9, the mismatch determination unit (3200) can obtain a care evaluation score (1201) and an emotional score (1111) obtained by emotionally analyzing the care evaluation content. Next, the mismatch determination unit can calculate the absolute value (3201) of the difference between the care evaluation score (1201) and the emotional score (1111). Next, the mismatch determination unit can determine whether the absolute value (3201) of the difference between the care evaluation score and the emotional score exceeds a threshold value (3202). Next, the mismatch determination unit can determine whether the absolute value (3201) of the difference between the care evaluation score and the emotional score for each item exceeds the threshold value (3202) (3203).

[0102] In one embodiment, if the difference between the care evaluation score and the sentiment score calculated through sentiment analysis of the care evaluation content (1100) exceeds a pre-stored threshold, a mismatch may be determined to have occurred. The occurrence of the mismatch may be determined using the following mathematical formula.

[0103] (|Nursing Care Evaluation Score (1200) - Emotional Score| > Threshold)

[0104] In one embodiment, if the absolute value of the difference between the care evaluation score and the care evaluation content (the emotional score calculated by sentiment analysis) divided by the care evaluation score exceeds a pre-stored standard, it may be determined that a mismatch has occurred. The determination of whether the mismatch has occurred may be determined using the following mathematical formula.

[0105] (|Nursing care evaluation score (1200) - emotional score| / |Nursing care evaluation score (1200)| > threshold)

[0106] In one embodiment, if the correlation coefficient between the care evaluation score and the sentiment score calculated by sentiment analysis of the care evaluation content is close to 0, it may be determined that a mismatch has occurred.

[0107] The determination of whether the above mismatch has occurred is not limited to these embodiments, and may include all types of methods for determining whether the tendency between the above nursing evaluation score and the emotional score calculated by emotional analysis of the nursing evaluation content is consistent.

[0108] Next, if the occurrence of the above mismatch is determined, the score filtering unit adjusts the nursing evaluation score (S500). Here, adjusting the nursing evaluation score means adjusting the nursing evaluation score so that its tendency matches the sentiment score calculated through sentiment analysis of the nursing evaluation content.

[0109] In one embodiment, adjusting the care assessment score may increase or decrease the care assessment score for a particular item where a mismatch occurred such that the difference between the care assessment score and the emotional score for that item falls below a pre-stored threshold.

[0110] For example, as illustrated in FIG. 10, the score filtering unit (3300) can obtain content regarding whether a mismatch has occurred between the care evaluation score (1201) and the care evaluation content, the emotional score (1111) calculated by emotional analysis, and the occurrence of a mismatch between the care evaluation score and the emotional score (3203). Next, the score filtering unit (3300) can adjust the care evaluation score only for the 'cleanliness' and 'skill' items in which a mismatch has occurred. Next, the score filtering unit (3300) can reduce the 5 points of the 'cleanliness' item to 4 points and reduce the 5 points of the 'skill' item to 4.7 points so that the absolute value of the difference between the care evaluation score (1201) and the emotional score (1111) does not exceed the threshold value (3203). Next, the score filtering unit can calculate the adjusted care evaluation score (1211).

[0111] Returning to Figure 4, the nursing evaluation score adjusted by the score filtering unit can be stored in the database (S600).

[0112] Meanwhile, if the occurrence of the above mismatch is not determined, the care evaluation score is terminated without being adjusted.

[0113] This method of filtering care evaluation scores can increase the reliability of care evaluation scores and improve their accuracy through sentiment analysis. Furthermore, it can minimize users' temporary emotions and biases, providing a more fair evaluation. Furthermore, by maintaining consistency between care evaluation scores and care evaluation content, other users can gain useful information about caregivers.

[0114] So far, embodiments of a method for filtering evaluation scores by utilizing mismatches between the emotional scores calculated by sentiment analysis of nursing evaluation content and the nursing evaluation scores have been described. According to the described embodiments, since the nursing evaluation scores are adjusted by sentiment analysis of the nursing evaluation content, factors that may affect the nursing evaluation scores themselves cannot be considered. However, in some embodiments, the nursing evaluation scores can be first adjusted by considering factors that may affect the nursing evaluation scores themselves, and then the nursing evaluation scores can be secondarily adjusted by utilizing mismatches between the emotional scores and the nursing evaluation scores. Hereinafter, related embodiments will be described with reference to FIGS. 11 to 13, and for the convenience of understanding, the description will focus on the parts that have been changed compared to FIG. 4.

[0115] Figure 11 illustrates a method for primarily adjusting nursing assessment scores using scoring tendency data. Below, a method for filtering assessment scores using mismatches according to Figure 11 is described.

[0116] As illustrated in Fig. 11, the server receives care evaluation content, care evaluation score, and data not entered by the user, and obtains scoring tendency data corresponding to the content not entered by the user from the database (S210).

[0117] In one embodiment, the scoring tendency of the scoring tendency data may include a scoring tendency according to the relationship between the user and the patient. The relationship between the user and the patient may include mother and child, mother and daughter, father and son, father and daughter, husband and wife, brothers and sisters, siblings, relatives, friends, acquaintances, colleagues, etc. Specifically, when the relationship between the user and the patient is mother and child, there is a possibility that a tendency to give a more extreme nursing evaluation score may be shown than when the relationship is mother and child. This tendency may be interpreted as a scoring tendency according to the relationship between the user and the patient. In this case, the score analysis unit may obtain data related to the relationship between the user and the patient from the user terminal, define the relationship between the user and the patient, and obtain a scoring tendency according to the defined relationship.

[0118] In one embodiment, the scoring tendency of the scoring tendency data may include a scoring tendency according to the patient's health status. The patient's health status may include healthy, average, critical, etc. Specifically, the more critical the patient's health, the more likely it is that a lower nursing evaluation score will be given. This pattern can be interpreted as a scoring tendency according to the patient's health status. In this case, the score analysis unit may obtain data related to the patient's health from the user terminal, define the patient's health using the data related to the patient's health, and obtain a scoring tendency according to the defined patient's health.

[0119] In one embodiment, as illustrated in FIG. 12, the scoring tendency of the scoring tendency data may include a scoring tendency according to the difficulty of providing care to a patient. The difficulty of providing care to a patient may be defined as very difficult, difficult, average, easy, very easy, etc. based on data not entered by the user (such as the patient's personal information, physical information, disease, symptoms, the caregiver's personal information, and the care location). Specifically, as the difficulty of providing care increases, the caregiver's satisfaction with the care may decrease. Therefore, a pattern in which the caregiver inputs a lower care evaluation score as the difficulty of providing care increases may be observed. This pattern may be interpreted as a scoring tendency according to the difficulty of providing care to a patient. In this case, the score analysis unit may obtain data related to the difficulty of providing care to a patient from a user terminal, define the difficulty of providing care to a patient using content related to the difficulty of providing care, and obtain scoring tendency data according to the defined difficulty of providing care.

[0120] For example, as illustrated in FIG. 12, the server (3000) can obtain nursing assessment content (1003-1 to 1003-3) containing content not input by the user from multiple user terminals. Next, the server can obtain content related to nursing difficulty from the nursing assessment content containing content not input by the user. Next, the server can define nursing difficulty levels (1303-1 to 1303-3) using the content related to nursing difficulty levels. Next, the score analysis unit (3500) of the server can obtain nursing assessment scores (1203-1 to 1203-3) and nursing difficulty content (1303-1 to 1303-3) related to the nursing assessment scores. Next, the score analysis unit (3500) can extract scoring tendency data (4203) according to care difficulty using the care evaluation scores (1203-1 to 1203-3) and care difficulty content (1303-1 to 1303-3). Next, the scoring tendency data (4203) according to care difficulty can be stored in a database.

[0121] Returning to Figure 11, next, the score filtering unit primarily adjusts the nursing assessment score based on the acquired scoring tendency data (S310). Here, the primary adjustment of the nursing assessment score can be performed using a method including Min-Max normalization, Z-score normalization, robust normalization, maximum absolute value normalization, etc. based on the scoring tendency of the acquired scoring tendency data.

[0122] In one embodiment, the primary adjustment of the care assessment score may include an adjustment that takes into account scoring tendencies based on the user-patient relationship. Specifically, the closer the user-patient relationship, the more difficult it is for the user to objectively evaluate the caregiver. Therefore, the normalization of the care assessment score may be an adjustment that takes this tendency into account.

[0123] In one embodiment, the normalization of the care assessment score may include adjustments that account for scoring trends based on the patient's health status. Specifically, the more critical the patient's health status, the more difficult it is for the user to objectively evaluate the caregiver. Therefore, the normalization of the care assessment score may be a normalization that compensates for this pattern.

[0124] In one embodiment, normalizing the care assessment score may include adjusting to account for scoring tendencies based on the difficulty of caregiving for the patient.

[0125] For example, as illustrated in FIG. 13, the server (3000) may define the patient's nursing difficulty (1303) using nursing evaluation content received from the user terminal (1000) or data not input by the user. Next, the server may obtain scoring tendency data (4203) according to the nursing difficulty corresponding to the patient's nursing difficulty previously stored in the database (4000). Next, the score filtering unit (3300) of the server may primarily adjust the nursing evaluation score (1204) obtained from the user terminal in consideration of the scoring tendency data (4203) according to the nursing difficulty. Next, the server may obtain the nursing evaluation score (1223) that has been primarily adjusted. Specifically, the more difficult the nursing difficulty, the more likely it is that the user will be dissatisfied with the caregiver, and thus, the user may tend to input a low nursing evaluation score. For example, as the difficulty of caregiving increases, there may be a tendency to input caregiving assessment scores within a range of 1 to 3. Therefore, the score filtering unit may interpret 1 as the lowest score, 2 as the midpoint, and 3 as the highest. Therefore, the score filtering unit may adjust 3 to the maximum score of 5, and 2 to the midpoint of 3.

[0126] The above scoring tendency and the method of primarily adjusting the nursing evaluation score based on the above scoring tendency are not limited to these embodiments.

[0127] This method of filtering evaluation scores using scoring tendency data can minimize distortion in user evaluations. Some users tend to input excessively high or low scores based on emotional input. By accounting for this tendency, score variance can be reduced. Furthermore, because certain users may be overly strict or lenient in their evaluation criteria, analyzing and adjusting their tendencies can increase the objectivity of evaluations. Consequently, caregivers can identify areas for improvement based on accurate feedback, and this also has the positive effect of increasing customer trust in caregiver evaluations.

[0128] So far, we have described embodiments of a method for filtering assessment scores that primarily adjusts care evaluation scores using scoring tendency data. Furthermore, in some embodiments, care evaluation scores can be adjusted by considering factors that may influence the sentiment analysis of care evaluation content. In Figure 14 below, we will describe the adjustment of a sentiment dictionary that can influence the sentiment analysis of care evaluation content.

[0129] Figure 14 illustrates a method for adjusting a sentiment dictionary used in sentiment analysis. Below, a method for adjusting a sentiment dictionary according to Figure 14 is described.

[0130] The sentiment analysis of the present disclosure may include sentiment analysis using a sentiment dictionary. General sentiment dictionaries may not reflect the sentiments associated with the caregiving domain. Therefore, sentiment analysis of caregiving evaluation content may require a sentiment dictionary specialized for the caregiving domain.

[0131] As illustrated in FIG. 14, the emotional dictionary adjustment unit (3400) of the server (3000) can obtain general emotional dictionary contents (4104) from the database (4000). In addition, the emotional dictionary adjustment unit can obtain a plurality of nursing evaluation contents (1104-1 and 1104-2) and a plurality of nursing evaluation scores (1204-1 and 1204-2) from the database. Next, the emotional dictionary adjustment unit can generate an emotional score of a word using the plurality of nursing evaluation contents (1104-1 and 1104-2) and the plurality of nursing evaluation scores (1204-1 and 1204-2).

[0132] In one embodiment, the emotional pre-conditioning unit may calculate a correlation coefficient between words included in multiple nursing evaluation contents and multiple nursing evaluation scores, and if the correlation coefficient is less than a preset threshold, assign an emotional score of 0.

[0133] In one embodiment, the sentiment pre-conditioning unit calculates a correlation coefficient between words included in multiple nursing assessment content and multiple nursing assessment scores. If the correlation coefficient exceeds a preset threshold, a score is assigned proportionally to the correlation coefficient. The sentiment pre-conditioning unit assigns a higher score to words with higher correlation coefficients, enabling more precise sentiment analysis.

[0134] In one embodiment, the emotional dictionary adjustment unit may adjust pre-stored emotional dictionary data using correlation coefficients between words included in a plurality of nursing evaluation contents and a plurality of nursing evaluation scores. The adjustment may be calculated through a mathematical formula (Y=X+s*r*a), where Y is the adjusted emotional score, X is the pre-stored emotional score, s is the sign of the pre-stored emotional score, r is the correlation coefficient calculated for each word, and a is a weight. Here, s may be +1 or -1, and a may correspond to a pre-set weight that is 10 or greater than 10.

[0135] For example, as illustrated in FIG. 14, the emotional dictionary data (4104) previously stored in the database may not be specialized for the caregiving domain. Specifically, the word "gratitude" may be interpreted as having a positive meaning even in the caregiving domain. Meanwhile, the word "worry" may have either a positive or negative meaning in the caregiving domain. Furthermore, the word "proudly" may have a negative meaning in the caregiving domain, unlike its typical meaning.

[0136] For example, as shown in Figure 14, among the multiple care evaluation contents, “Thank you so much. Thanks to you, I didn’t have to worry.” (1104-1) and the related care evaluation score (1204-1) have the maximum score of 5 for all items. Therefore, the words (such as “so much,” “thank you,” and “worried”) in the corresponding care evaluation contents (1104-1) are likely to have positive meanings related to the care domain. On the other hand, among the multiple care evaluation contents, “I was worried because you didn’t come often. Nevertheless, you confidently asked for a 3-day supply.” (1104-2) and the related care evaluation score (1204-2) have the minimum score of 1 for all items. Therefore, the words (such as “often,” “worried,” and “confidently”) in the corresponding care evaluation contents (1104-2) are likely to have negative meanings related to the care domain. The emotional pre-regulation unit (3400) considers the nursing evaluation contents (1104-1 to 1104-2) and the nursing evaluation scores (1204-1 to 1204-2), and thus can assign a more positive emotional score to the word "thank you" because there are more high-scoring nursing evaluation scores. Furthermore, the word "worry" can be assigned a neutral emotional score because there are both high-scoring and low-scoring nursing evaluation scores. Meanwhile, the word "proudly" can be assigned a more negative emotional score because there are more low-scoring nursing evaluation scores.

[0137] The above emotional dictionary and the method for generating the emotional dictionary are not limited to these embodiments. The emotional dictionary may include any type of database capable of indicating emotional scores for words. Furthermore, the method for generating the emotional dictionary may include any type of generation method capable of increasing the efficiency of emotional analysis for a specific domain in relation to a specific domain. Specifically, the method for generating the emotional dictionary may include both a method for generating the dictionary using only care evaluation content and care evaluation scores, and a method for adjusting previously stored emotional dictionary data with care evaluation content and care evaluation scores.

[0138] This method of filtering evaluation scores using a tuned sentiment dictionary can accurately perform sentiment analysis tailored to specific fields and contexts. Specifically, this method can effectively handle specialized terminology and specific expressions commonly used in specific industries. Therefore, this method of filtering evaluation scores using a tuned sentiment dictionary can increase the consistency between evaluation scores and evaluation content, thereby enhancing the reliability of evaluation scores.

[0139] So far, we have described embodiments of methods for filtering evaluation scores using mismatches between emotional scores and caregiving evaluation scores. However, in some embodiments, the caregiving evaluation score can be adjusted or deleted using mismatches between the guardian's access location to the caregiver matching application and the caregiving location. Relevant embodiments are described below with reference to FIGS. 15 through 18.

[0140] FIG. 15 is a flowchart illustrating a method for deleting a care evaluation score by utilizing a mismatch between a caregiver matching application access location and a care location according to some embodiments of the present disclosure.

[0141] In step S700, the server can obtain care evaluation content, care evaluation score, and caregiver matching application access location information from the user terminal. Here, the caregiver matching application access location information can be obtained from data not entered by the user. For example, when a guardian accesses the caregiver matching application, the user terminal can store GPS location information at the time of access and transmit the GPS location information upon request from the server.

[0142] In step S800, the server may obtain location information regarding the provision of care services from the user terminal. This location information may refer to GPS information regarding the actual location where the patient's care is provided. For example, the server may obtain GPS information corresponding to the location of the care provided, entered by the guardian when uploading the caregiver recruitment notice, as GPS information regarding the actual location where the patient's care is provided.

[0143] In step S900, the mismatch determination unit may determine whether there is a mismatch between the guardian's caregiver matching application access location and the care service provision location. Here, a mismatch may mean that the distance between the application access location and the care service provision location is greater than a pre-stored threshold.

[0144] For example, if the caregiver and the actual care location are more than 100km apart, objective assessment of the caregiving process can be difficult because the caregiver cannot observe the caregiving process. The distance between the caregiver and the actual care location can be determined by measuring the distance between the caregiver's access point in the caregiver matching application and the care service location. Therefore, if the distance between the caregiver's access point in the caregiver matching application and the care service location is more than 100km, a mismatch can be considered to have occurred.

[0145] In step S1000, the score filtering unit can delete the care evaluation score and care evaluation content if a mismatch occurs. The distance between the caregiver and the care service provider makes it difficult to objectively evaluate the caregiver's care services. Therefore, by deleting assessments that are difficult to objectively assess, the reliability of the care evaluation score and care evaluation content can be secured.

[0146] For example, if the distance between the guardian's access location to the caregiver matching application and the location where the care service is provided is more than 100 km, it is difficult to objectively evaluate the guardian's care service, so the score filtering unit may delete the care evaluation score and care evaluation content, which are evaluations of the guardian's care service.

[0147] Meanwhile, if the mismatch determination unit determines that there is no mismatch between the caregiver application access location and the care service provision location, the mismatch determination unit determines whether the caregiver cannot access the care service provision location (S1100). Even if the caregiver and the care location are close together, if the caregiver cannot access the care location, it may be difficult for the caregiver to objectively assess the care service.

[0148] For example, suppose the distance between the caregiver's access point to the caregiver matching application and the care location is within 100km, but the patient's care location is an isolation ward or the patient's illness is a contagious disease requiring isolation. Under this assumption, the caregiver cannot access the care location. Therefore, it may be difficult for the caregiver to objectively evaluate the care service. Therefore, even if there is no mismatch between the caregiver and the care location, if the caregiver cannot access the care location, the score filter unit may delete the care evaluation score and care evaluation content.

[0149] The above 100 km is an example for the convenience of understanding, and the threshold for determining the above mismatch is not limited to 100 km.

[0150] Meanwhile, if the mismatch judgment unit determines that the caregiver can access the care location, there is no need to delete the care evaluation score and care evaluation content because the caregiver is likely to make an objective judgment about the care service.

[0151] Figure 16 illustrates an example of a method for filtering care evaluation scores using mismatches between the caregiver matching application access location and the caregiving location, according to some embodiments of the present disclosure. Figure 16 will be described below.

[0152] As illustrated in Figure 16, the timing of a caregiver's access to the caregiver matching application can influence the caregiver's evaluation of the caregiver's care service. For example, during the care service provision period (p100), the caregiver may be located close to the patient, preventing a mismatch between the caregiver's access location to the caregiver matching application and the care service provision location. Therefore, if the caregiver enters care evaluation content and care evaluation scores at a point in time (t1) during the care service provision period, the mismatch will not occur, and thus the care evaluation content and care evaluation scores will not be deleted.

[0153] However, although the caregiver may be close to the patient during the care service provision period, he or she may be far away from the patient during the non-care service provision period (p200). Since the caregiver observed the care service during the care service provision period, it can be considered that he or she can objectively evaluate the care service. However, if the caregiver inputs the care evaluation content and care evaluation score at one point in time (t2) during the non-care service provision period (p200), a mismatch may occur between the caregiver matching application access location and the care service provision location. If the score filtering unit deletes the care evaluation content and care evaluation score of the caregiver due to a mismatch even though the caregiver conducted an objective evaluation, it may be difficult to consider that the actual evaluation was reflected.

[0154] Therefore, to overcome the above problem, the user terminal and server can store the access location of the guardian / caregiver matching application during the nursing service provision period (p100). Next, if the guardian inputs the nursing evaluation content and nursing evaluation score at one point in time (t2) during the nursing service non-provision period, the mismatch determination unit can determine whether a mismatch has occurred based on the previously stored access location of the guardian / caregiver matching application during the nursing service provision period.

[0155] Figure 17 illustrates an example of a method for adjusting a care evaluation score by utilizing a mismatch between a guardian's caregiver matching application access location and the caregiving location, according to some embodiments of the present disclosure. Below, Figure 16 will be described, focusing on the changed portions compared to Figure 15 for ease of understanding.

[0156] In step S1200, if the mismatch determination unit determines that a mismatch has occurred, the score filtering unit may adjust the care evaluation score. For example, the score filtering unit may adjust the care evaluation score to be closer to the average care evaluation score for the caregiver of the caregiver in proportion to the distance between the caregiver's access location to the caregiver matching application and the care service provision location.

[0157] For example, as illustrated in FIG. 18, the server can obtain information regarding the distance (1315) between the guardian's access location to the caregiver matching application and the location where care services are provided, using data (1205) not entered by the user. The score filtering unit can adjust the care evaluation score (1205) by considering the distance (1315) between the guardian's access location to the caregiver matching application and the location where care services are provided. The adjusted care evaluation score (1215) can be stored in a database.

[0158] This evaluation score filtering method filters evaluation-related data and analyzes only the important portions, thereby reducing unnecessary data processing and efficiently utilizing server computing resources. Furthermore, compared to inefficient simple processing methods, this evaluation score filtering method can reduce redundant data processing. While simple numerical evaluations process all users' evaluations equally, the evaluation score filtering method of the present disclosure learns data patterns and trends in advance, thereby preventing redundant or unnecessary computations.

[0159] It should be noted that the effects of the present disclosure are not limited to the above effects.

[0160] The technical ideas of the present disclosure, which have been described with reference to FIGS. 1 to 18, may be implemented as computer-readable codes on a computer-readable medium. The computer-readable recording medium may be, for example, a removable recording medium (such as a USB storage device or a removable hard disk). The computer program recorded on the computer-readable recording medium may be transmitted to another computing device via a network such as the Internet and installed on the other computing device, thereby allowing the computer program to be used on the other computing device.

[0161] Hereinafter, the hardware configuration of an exemplary computing device according to some embodiments of the present disclosure will be described with reference to FIG. 19. The computing device may be, for example, a user terminal (1000) or a server (3000).

[0162] FIG. 19 is an exemplary hardware configuration diagram that may implement a computing device in various embodiments of the present disclosure. A computing device (5000) according to the present embodiment may include one or more processors (5100), a system bus (5600), a communication interface (5200), a memory (5400) that loads a computer program (5500) executed by the processor (5100), and a storage (5300) that stores the computer program (5500). Only components related to the embodiments of the present disclosure are illustrated in FIG. 19. Therefore, those skilled in the art will appreciate that other general components may be included in addition to the components illustrated in FIG. 19.

[0163] The processor (5100) controls the overall operation of each component of the computing device (5000). The processor (5100) may include at least one of a Central Processing Unit (CPU), a Micro Processor Unit (MPU), a Micro Controller Unit (MCU), a Graphics Processing Unit (GPU), or any other type of processor well known in the art of the present disclosure. In addition, the processor (5100) may perform operations for at least one application or program for executing methods / operations according to various embodiments of the present disclosure. The computing device (5000) may include two or more processors.

[0164] The memory (5400) stores various contents, commands, and / or information. The memory (5400) can load one or more programs (190) from the storage (5300) to execute methods / operations according to various embodiments of the present disclosure. An example of the memory (5400) may be, but is not limited to, RAM. The system bus (5600) provides communication functions between components of the computing device (5000).

[0165] The system bus (5600) may be implemented as various types of buses, such as an address bus, a data bus, and a control bus. The communication interface (5200) supports wired and wireless Internet communication of the computing device (5000). The communication interface (5200) may also support various communication methods other than Internet communication. To this end, the communication interface (5200) may be configured to include a communication module well known in the technical field of the present disclosure. The storage (5300) may non-temporarily store one or more computer programs (5500). The storage (5300) may be configured to include a non-volatile memory such as a flash memory, a hard disk, a removable disk, or any type of computer-readable recording medium well known in the technical field of the present disclosure.

[0166] The computer program (5500) may include one or more instructions implementing methods / operations according to various embodiments of the present disclosure. When the computer program (5500) is loaded into the memory (5400), the processor (5100) may execute the one or more instructions to perform the methods / operations according to various embodiments of the present disclosure.

[0167] Although the embodiments of the present disclosure have been described with reference to the attached drawings, those skilled in the art will appreciate that the embodiments of the present disclosure may be implemented in other specific forms without altering the technical concepts or essential features thereof. Therefore, it should be understood that the embodiments described above are exemplary in all respects and not restrictive. The scope of protection of the present disclosure should be interpreted by the claims below, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of the technical ideas defined by the present disclosure.

Claims

1. In a method performed by a computing system, A step of obtaining first evaluation content and a first evaluation score related to the first evaluation content; A step of obtaining a second evaluation score by performing sentiment analysis on the first evaluation content; A step of determining whether a mismatch occurs between the first evaluation score and the second evaluation score; If the occurrence of the above mismatch is determined, a step of adjusting the first evaluation score; and comprising a step of storing the adjusted first evaluation score; How to filter evaluation scores.

2. In paragraph 1, The above first evaluation content is, Evaluation content for matched caregivers entered through the caregiver matching platform. How to filter evaluation scores.

3. In paragraph 2, The step of obtaining the second evaluation score by analyzing the above sentiment is: A step of sentiment analysis comprising at least one of the personal information of the patient of the caregiver, the personal information of the user of the caregiver, the consultation content of the user, and the user's access record to the caregiver matching platform, together with the first evaluation content. How to filter evaluation scores.

4. In paragraph 2, The step of obtaining the above first evaluation score is: A step of adjusting the first evaluation score by considering at least one of an evaluation scoring tendency according to a relationship between a user of the caregiver matching platform and a patient, an evaluation scoring tendency according to the difficulty of caregiving for the patient, and an evaluation scoring tendency according to the health status of the patient is further included. How to filter evaluation scores.

5. In paragraph 2, The above sentiment analysis is, A sentiment analysis using a sentiment lexicon generated by obtaining a correlation coefficient between a first word and a first emotion using a plurality of nursing evaluation contents and a plurality of nursing evaluation scores related to the plurality of nursing evaluation contents, calculating a sentiment score of the first word proportional to the correlation coefficient if the correlation coefficient of the first emotion is greater than or equal to a threshold, and calculating a sentiment score of the first word as 0 if the correlation coefficient of the first emotion is less than the threshold. How to filter evaluation scores.

6. In a method performed by a computing system, A step of obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform and a first evaluation score related to the first evaluation content; A step of obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; A step of determining whether there is a mismatch between the geographic location information of the user terminal and the service provision location; and If the occurrence of the above mismatch is determined, a step of deleting the first evaluation content and the first evaluation score is included. How to filter evaluation scores.

7. In paragraph 6, The step of determining whether there is a mismatch above is: A step of determining the geographic location information of the user terminal by using the access location for the service match platform during the service provision period of the first matching of the user terminal, How to filter evaluation scores.

8. In a method performed by a computing system, A step of obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform and a first evaluation score related to the first evaluation content; A step of obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; A step of determining whether a mismatch occurs between the geographic location information of the user terminal and the service provision location; If it is determined that the above mismatch has not occurred, a step of determining whether the user of the user terminal cannot access the service provision location; and If it is determined that the above service provision location is inaccessible, a step of deleting the first evaluation content and the first evaluation score is included. How to filter evaluation scores.

9. In paragraph 8, The step of determining whether there is a mismatch above is: A step of determining the geographical location information of the user terminal by using the access location to the service match platform during the service provision period of the first matching of the user terminal. How to filter evaluation scores.

10. In a method performed by a computing system, A step of obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform and a first evaluation score related to the first evaluation content; A step of obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; A step of determining whether there is a mismatch between the geographic location information of the user terminal and the service provision location; When the occurrence of the above mismatch is determined, a step of adjusting the first evaluation score; and comprising a step of storing the adjusted first evaluation score; How to filter evaluation scores.

11. In paragraph 10, The step of determining whether there is a mismatch above is: A step of determining the geographic location information of the user terminal by using the access location for the service match platform during the service provision period of the first matching of the user terminal, How to filter evaluation scores.

12. Processor; network interface; memory; and A computer program loaded into the above memory and executed by the above processor, The above processor, An instruction for obtaining first evaluation data and a first evaluation score related to the first evaluation data from a user terminal; An instruction for obtaining a second evaluation score by performing sentiment analysis on the first evaluation data; An instruction for determining whether a mismatch occurs between the first evaluation score and the second evaluation score; and Including an instruction for adjusting the first evaluation score when a mismatch between the first evaluation score and the second evaluation score is determined to occur; Evaluation score filtering device.

13. Processor; network interface; memory; and A computer program loaded into the above memory and executed by the above processor, The above processor, An instruction for obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform and a first evaluation score related to the first evaluation content; An instruction for obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; An instruction for determining whether there is a mismatch between the geographic location information of the user terminal and the service provision location; and In case the occurrence of the above mismatch is determined, an instruction for deleting the first evaluation content and the first evaluation score is performed; Evaluation score filtering device.

14. Processor; network interface; memory; and A computer program loaded into the above memory and executed by the above processor, The above processor, An instruction for obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform and a first evaluation score related to the first evaluation content; An instruction for obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; An instruction for determining whether a mismatch occurs between the geographic location information of the user terminal and the service provision location; If it is determined that the above mismatch has not occurred, an instruction for determining whether the user of the user terminal cannot access the service provision location; and If it is determined that the above service provision location is inaccessible, instructions for deleting the first evaluation content and the first evaluation score; including performing Evaluation score filtering device.

15. Processor; network interface; memory; and A computer program loaded into the above memory and executed by the above processor, The above processor, An instruction for obtaining, from a user terminal, first evaluation content for a service provider connected through a first matching via a service matching platform and a first evaluation score related to the first evaluation content; An instruction for obtaining information on a service provision location of the service provider from information on the first matching, wherein the service provision location is geographic location information; An instruction for determining whether there is a mismatch between the geographic location information of the user terminal and the service provision location; An instruction for adjusting the first evaluation score when the occurrence of the above mismatch is determined; and Including an instruction to store the adjusted first evaluation score. Evaluation content filtering device.

Images