Online inquiry evaluation processing method and device

Abstract

The invention discloses an online inquiry evaluation processing method and device, and relates to the technical field of internet medical treatment. A specific embodiment of the method comprises the steps of classifying evaluation texts; and displaying the evaluation text corresponding to the specific classification in response to a triggering operation of the user on the specific classification screening key through the man-machine interaction control. According to the embodiment, the evaluation texts are automatically classified, the manual participation workload is reduced, and the classification screening keys are set according to classification, so that the user can trigger the specific classification screening keys in the interface as required to quickly screen and display the corresponding evaluation texts, and the user experience is improved. The problem that a user needs to consume a lot of time to manually browse and discriminate massive evaluations at present is effectively solved, and powerful support is provided for continuous improvement of service quality and user experience.

Description

Technical Field

[0001] This invention relates to the field of internet medical technology, and in particular to an online consultation evaluation processing method and apparatus. Background Technology

[0002] With the popularization of internet-based healthcare services, patient reviews have become an important reference for doctors to improve their services and for patients to make decisions. Furthermore, regulations require that no reviews be deleted. Current technology relies on operators to manually review and mark invalid reviews, lacking an effective classification mechanism for detailed differentiation. At the same time, doctors lack a screening mechanism, making it impossible to quickly locate the required reviews based on actual needs. This forces doctors to spend a significant amount of time manually browsing and verifying a massive number of reviews. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide an online consultation evaluation processing method and apparatus, which can at least solve the problem of the lack of evaluation classification and screening mechanisms in the prior art.

[0004] To achieve the above objectives, according to one aspect of the present invention, an online consultation evaluation processing method is provided, comprising: Classify the evaluation texts; In response to a user's triggering operation of a specific category filter key through a human-computer interaction control, the evaluation text corresponding to the specific category is displayed.

[0005] To achieve the above objectives, according to another aspect of the present invention, an online consultation evaluation processing device is provided, comprising: The classification module is used to classify the evaluation text; The filtering module is used to display the evaluation text corresponding to the specific category in response to the user's trigger operation of the filter key for a specific category through the human-computer interaction control.

[0006] To achieve the above objectives, according to another aspect of the present invention, an electronic device for online consultation evaluation processing is provided.

[0007] The electronic device of this invention includes: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any of the online consultation evaluation processing methods described above.

[0008] To achieve the above objectives, according to another aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements any of the online consultation evaluation processing methods described above.

[0009] To achieve the above objectives, according to another aspect of the present invention, a computing program product is provided. One such computing program product includes a computer program, which, when executed by a processor, implements the online consultation and evaluation processing method provided in the present invention.

[0010] According to the solution provided by the present invention, one embodiment of the invention has the following advantages or beneficial effects: by automatically classifying the evaluation text, the classification efficiency is improved, the workload of manual intervention is reduced, and the classification filter key is set according to the classification, so that users can trigger a specific classification filter key on the interface as needed to quickly filter and display the corresponding evaluation text. This effectively solves the problem that users need to spend a lot of time manually browsing and screening a large number of evaluations, and provides strong support for the continuous improvement of service quality and user experience.

[0011] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0012] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein: Figure 1 This is a schematic diagram of the main process of an online consultation evaluation processing method according to an embodiment of the present invention; Figure 2(a) is a schematic diagram of the existing evaluation sorted by submission time; Figure 2(b) is a schematic diagram showing only the high-value assessment; Figure 2(c) is a schematic diagram of marking read evaluation text and adjusting the display mode of read evaluation text; Figure 2(d) is a schematic diagram showing the highlighting of valuable information in the evaluation text; Figure 3 This is a flowchart illustrating an optional online consultation evaluation processing method according to an embodiment of the present invention. Figure 4 This is a flowchart illustrating another optional online consultation evaluation processing method according to an embodiment of the present invention; Figure 5 This is a flowchart illustrating another optional online consultation evaluation processing method according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the main modules of an online consultation evaluation and processing device according to an embodiment of the present invention; Figure 7 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied; Figure 8 This is a schematic diagram of the structure of a computer system suitable for implementing the embodiments of the present invention, such as a mobile device or server. Detailed Implementation

[0013] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0014] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0015] Where there is no conflict, the embodiments and features in the embodiments of this invention can be combined with each other. The acquisition, transmission, storage, use, and processing of data in the technical solutions of this invention comply with the relevant provisions of national laws and regulations, are used for legal and reasonable purposes, and are not shared, disclosed, or sold outside of these legal uses, and are subject to supervision and management by regulatory authorities.

[0016] Regarding user information, necessary measures should be taken to prevent unauthorized access to such personal information data, ensure that personnel authorized to access such data comply with relevant laws and regulations, and safeguard the security of user personal information. Once this user personal information data is no longer needed, risks should be minimized by restricting or even prohibiting data collection and / or deleting the data. Where applicable, including in certain relevant applications, user privacy should be protected through data de-identification, such as by removing specific identifiers (e.g., date of birth), controlling the amount or specificity of stored data (e.g., collecting location data at the city level rather than the specific address level), controlling how data is stored, and / or other de-identification methods.

[0017] With the widespread adoption of internet-based healthcare services, patient reviews of doctors have become a crucial reference for improving doctor services and influencing patients' medical decisions. According to e-commerce laws, patient reviews (including negative ones) cannot be deleted. Currently, operations staff manually review reviews, marking those containing malicious comments (such as "The doctor is incompetent, a fraud") or irrelevant information (such as "The weather is bad today") as "invalid" in the backend. However, these marked invalid reviews are still displayed in their entirety to doctors, who lack a filtering function, thus failing to effectively hide invalid reviews.

[0018] Against this backdrop, the existing evaluation processing model faces two challenges: on the one hand, it needs to avoid the interference caused to doctors by negative expressions (or aggressive words) in invalid evaluations; on the other hand, it needs to help doctors efficiently screen out constructive and valuable evaluations (such as "patient during consultation but not detailed enough in explaining medication") so as to quickly obtain content that is valuable for improving their own services.

[0019] Furthermore, existing technologies have the following drawbacks: doctors cannot accurately distinguish between "malicious attacks" and "constructive criticism." For example, in a review stating "The doctor's response is slow; I hope for faster response," some doctors might mistakenly interpret the first part as a negative review and selectively ignore it, leading to the omission of valuable feedback. Therefore, it is urgent to find a balance between "compliance in not deleting reviews" and "optimizing the doctor's experience," implementing intelligent classification of reviews and providing filtering functions so that valuable reviews can be prioritized and displayed according to the doctor's needs.

[0020] See Figure 1 The diagram shows the main flowchart of an online consultation evaluation processing method provided by an embodiment of the present invention, which includes the following steps: S101: Classify the evaluation texts; S102: In response to the user's triggering operation of the filter key for a specific category through the human-computer interaction control, the evaluation text corresponding to the specific category is displayed.

[0021] This solution boasts excellent versatility and, beyond the healthcare sector, can be widely applied to target business scenarios requiring intelligent classification and evaluation. Examples include intelligent classification of product and store reviews in e-commerce, course and teacher reviews in online education, bank service and wealth management product reviews in financial services, restaurant and movie reviews in catering and entertainment, government service satisfaction evaluations in government services, and travel service evaluations in transportation. This solution effectively identifies malicious or low-quality reviews and extracts constructive information from valuable reviews, demonstrating significant cross-industry application value. A detailed explanation will primarily focus on the healthcare scenario as a target business scenario.

[0022] For step S101, the evaluation text to be classified is determined. This evaluation text can be entered by the patient in real time or it can be historical evaluation text before the implementation of this solution. That is, this solution supports both offline batch processing and online real-time processing modes. After determining the evaluation text, special characters such as \n are removed using regular expressions to eliminate meaningless characters, thus obtaining standardized evaluation text. Then, a fine-tuned AI (Artificial Intelligence) model is used to constructively evaluate the standardized evaluation text to determine its specific classification. This solution preferably sets three classifications: value evaluation, general value evaluation, and invalid evaluation.

[0023] In one alternative implementation, the AI ​​model is based on BERT (Bidirectional Encoder Representations from Transformers) or a variant thereof. BERT is a pre-trained language model based on a bidirectional Transformer encoder, capable of effectively capturing semantic relationships between words. Its core feature is achieving deep semantic understanding through a "pre-training + fine-tuning" paradigm. Its bidirectional attention mechanism can simultaneously capture the "left→right" and "right→left" contextual relationships in the text, overcoming the limitations of traditional keyword matching. In medical evaluation processing scenarios, the model, fine-tuned with the corpus, can accurately identify implicit attack semantics (such as irony), assess the business relevance, constructiveness, and authenticity of evaluations, providing semantic support for malicious evaluation judgment and value scoring.

[0024] For step S102, to facilitate doctors in viewing the evaluations they need based on their actual requirements, a filtering entry can be set up on the doctor's front-end interface. For example, after logging into the internet healthcare system, a doctor enters the personal evaluation management page, where three filtering buttons are displayed at the top: "High-Value Evaluation," "Ordinary-Value Evaluation," and "Invalid Evaluation." Doctors can click the corresponding filtering item through the human-computer interaction control according to their actual needs. The front-end responds to this operation, filtering the evaluations corresponding to that category and displaying them in order, eliminating the need for doctors to search through each evaluation individually, thereby improving the efficiency of doctors browsing evaluations.

[0025] For example, when a doctor clicks the "High-Value Evaluation" button via the human-computer interaction control, the system immediately filters and displays constructive evaluation texts that have been intelligently categorized and marked. See Figure 2(a) for an example of existing evaluations sorted by submission time. With this solution, after the doctor clicks "High-Value Evaluation," only constructive evaluation texts are displayed (see Figure 2(b)), such as "The treatment process was patient and meticulous, which made me feel very reassured, but I hope for more night shift hours," "I hope for more detailed medication instructions," and "The doctor's response was slow; I hope for faster response times." Non-high-value evaluation texts are not displayed.

[0026] There are no restrictions on the display method; for example, the evaluation texts can be displayed in reverse chronological order of submission time. In one alternative implementation, the evaluations are categorized based on their value scores, and therefore can be displayed in descending order of value scores. Furthermore, the system can display the value score to the left of each evaluation to provide doctors with a reference.

[0027] In this way, when doctors need to view evaluations under a specific category, they only need to click the corresponding category filter to quickly filter and display relevant evaluation content, eliminating the need to search through a massive number of evaluations one by one, thus significantly improving work efficiency. Furthermore, unless the doctor actively clicks the invalid evaluation filter, invalid evaluations will not be displayed to the doctor in the default interface, achieving an automatic filtering function for invalid evaluations.

[0028] In one optional implementation, this solution provides a one-click positioning control, such as adding a "Quickly locate high-value" button. After the doctor clicks the "High-value evaluation" filter, the focus will be positioned on the first high-value evaluation in the displayed evaluation text list, helping the doctor to quickly focus on useful information.

[0029] In addition, considering the continuous accumulation of evaluations, a "read" function can be set up. After a doctor views a certain evaluation text, he can mark it as "read" through a tagging interactive control. The front end responds to this operation by hiding the evaluation text or moving the evaluation text to the end of the list, as shown in Figure 2(c), and adjusting the display order of the remaining evaluation texts in the list to avoid the problem of users having to flip through pages frequently because the first high-value evaluation remains unchanged.

[0030] In another optional implementation, this solution also includes a viewing interaction control. When a user clicks on any review through this control to enter the details page, the "valuable information points" in the review are automatically highlighted (e.g., underlined), such as "faster response," as shown in Figure 2(d), to reduce user reading time and improve information acquisition efficiency.

[0031] It is understood that the specific values ​​listed in the above hypothetical description of online consultation evaluation processing are merely illustrative and not restrictive. In practice, other values ​​can be set according to needs. Similarly, the specific values ​​listed in the subsequent embodiments are also just examples and are not the only limitation.

[0032] The method provided in the above embodiments automatically classifies the evaluation text and sets category filter keys according to the classification, so that users can trigger specific category filter keys on the interface as needed to quickly filter and display the corresponding evaluation text. This effectively solves the problem that existing users need to spend a lot of time manually browsing and screening massive amounts of evaluations, and provides strong support for the continuous improvement of service quality and user experience.

[0033] See Figure 3 The diagram illustrates an optional online consultation evaluation processing method according to an embodiment of the present invention, including the following steps: S301: Identify the information in the evaluation text that corresponds to a specific field related to the target business scenario, and determine the true evaluation value based on the identification result of the specific field information; S302: Identify semantic information related to the target business scenario in the evaluation text, and determine the business relevance based on the semantic information identification results; S303: Identify improvement suggestion information related to the target business scenario in the evaluation text, and determine a constructive score based on the identification result of the improvement suggestion information; S304: Determine the value score of the evaluation text based on at least one of the evaluation true value, the business relevance, and the constructive score, so as to determine the classification of the evaluation text according to the value score.

[0034] Taking the medical scenario as an example, based on the specific field information, semantic information, and improvement suggestion information related to the medical scenario in the evaluation text, the evaluation true value, business relevance, and constructive score of the evaluation text are determined respectively.

[0035] The evaluation of real values ​​reflects the ability to semantically capture details of the scene, assessing whether it includes information such as specific time, event, and result (e.g., "I had a consultation on October 5, 2024, waited 20 minutes, and the consultation ended in 5 minutes" scores 10 points, while "poor experience" scores 1 point). The reason for identifying the scene here is that some evaluations are unrelated to medical scenarios. For example, "I went to the park for two hours today" also includes information such as time, event, and result, but it does not fall within the scope of medical evaluation.

[0036] Identifying business relevance, such as treatment relevance, involves recognizing semantic information in the evaluation text that is directly related to medical behavior (such as consultation content, medication recommendations, examination procedures, etc.) and outputting a score from 0 to 10 (e.g., "It is recommended to adjust the dosage of antihypertensive medication" scores 8 points). Typically, the evaluation text is processed based on pre-defined prompts to obtain the treatment relevance score. For example, evaluation text containing medically relevant prompts such as "consultation," "medication," "examination," and "diagnosis," such as "The doctor inquired about the symptoms in detail and arranged the corresponding examinations," will obtain a high treatment relevance score.

[0037] The constructive score is determined by identifying whether the evaluation text contains specific suggestions for improvement (e.g., "I hope the doctor can explain the lab report in detail" scores 9 points, while "poor service" scores 2 points). Based on the identification results and the identified suggestions, the constructive score is determined. For example, some evaluation texts may include suggestions for improvement, but their descriptions have a low relevance to the doctor's service. For instance, "The waiting time is too long; I suggest optimizing the process" or "The drug prices are too high" will receive a relatively lower constructive score than "I hope the doctor can explain the meaning of each indicator on the lab report in detail." This can be addressed by injecting "suggestion-style sentence patterns" (e.g., "hope + improvement point," "suggestion + optimization plan") into the model during fine-tuning, which can effectively improve the accuracy of identifying different levels of constructiveness.

[0038] This solution calculates the value score of the evaluation text based on at least one of the following: evaluation authenticity, business relevance, and constructiveness score. When considering multiple indicators, a weighted summation method can be used, for example: Business Relevance × 0.4 + Constructiveness Score × 0.3 + Evaluation Authenticity × 0.3, where the weights can be adjusted based on practical experience. Finally, the evaluation text is classified based on the value score, for example, ≥6 points is "high-value evaluation", 3-5 points is "ordinary-value evaluation", and <3 points is "invalid evaluation" (not malicious but lacking substantial information, such as "good"). Invalid evaluations can be automatically marked "invalid" and a message can be displayed stating "This evaluation has been identified as valueless content".

[0039] The method provided in the above embodiments classifies evaluation texts by automatically performing multi-dimensional value scoring, significantly improving the efficiency of evaluation classification, reducing the workload of manual intervention, and facilitating the rapid selection of corresponding evaluation texts based on the classification results. For complex evaluation texts that include both criticisms and improvement suggestions, value scoring is used for quantitative processing, avoiding the confusion caused by users' subjective judgments and improving the objectivity and consistency of evaluation classification.

[0040] See Figure 4 The diagram illustrates another optional online consultation evaluation processing method according to an embodiment of the present invention, including the following steps: S401: Identify negative expression words in the evaluation text to calculate the percentage of negative expression characters; S402: In response to the fact that the proportion of negative expression characters is greater than or equal to a preset threshold, the evaluation text is classified as an invalid evaluation; S403: In response to the fact that the proportion of negative expression characters is less than the preset threshold, identify the information in the evaluation text that corresponds to a specific field related to the target business scenario, and determine the true value of the evaluation based on the identification result of the specific field information; S404: Identify semantic information related to the target business scenario in the evaluation text, and determine the business relevance based on the semantic information identification results; S405: Identify improvement suggestion information related to the target business scenario in the evaluation text, and determine a constructive score based on the identification result of the improvement suggestion information; S406: Determine the value score of the evaluation text based on at least one of the evaluation true value, the business relevance, and the constructive score, so as to determine the classification of the evaluation text according to the value score.

[0041] This implementation method, after standardizing the evaluation text, identifies negative expression words in the standardized evaluation text and calculates the negative expression character ratio using the formula: negative expression character ratio = total number of negative expression words / total number of characters in the evaluation text.

[0042] In one optional implementation, firstly, a preset sensitive word library corresponding to the medical scenario (such as "treating human life like dirt" or "cheating money") is used to identify suspected sensitive words (including homophonic alternatives such as "yongyi") in the evaluation text; then, a preset negative expression word library corresponding to the medical scenario (such as "qunyi" or "scammer") is used to identify negative expression words from the suspected sensitive words. The identification of suspected sensitive words can employ semantic vector retrieval, such as performing word segmentation on the text and calculating the semantic vector similarity between the segmented words and sensitive words in the preset sensitive word library. If the semantic vector similarity is higher than a preset semantic vector threshold, it is determined to be a suspected sensitive word; otherwise, it is considered a non-sensitive word.

[0043] The identification of suspected sensitive words, the identification of negative expressions, and the calculation of the proportion of negative expression characters mentioned above can all be achieved through a large-scale model. This model can be combined with the aforementioned... Figure 1 The model used can be integrated or a separate model; there is no restriction here. In the case of integration, the model can construct a two-branch semantic understanding framework of "negative expression recognition - constructive evaluation".

[0044] It should be noted that both the preset sensitive word database and the preset negative expression database are pre-organized and established by the operations staff, and the specific organization method is not limited here. Furthermore, these two databases, especially the preset negative expression database, support dynamic expansion. For example, by calculating the word vector similarity between candidate words and each negative expression word in the preset negative expression database, candidate words with higher word vector similarity are used as expanded words of the corresponding negative expression words (such as associating "unethical doctor" with "quack doctor"), thereby realizing intelligent expansion of the database.

[0045] This solution establishes a rule for determining malicious reviews (a type of invalid review). When the proportion of negative descriptive characters is ≥30% (i.e., a preset threshold, only an example, adjustable in practice), the review text is directly classified as a malicious review. In one optimized implementation, this solution pre-assigns a negative level (or aggression level) to each negative descriptive word in a preset negative descriptive word library. Extended words of negative descriptive words inherit the negative level of the negative descriptive word. For example, five levels are set, with level 5 being the highest. The malicious review determination rule is that any of the following conditions must be met to be considered a "malicious review": the proportion of negative descriptive characters is ≥30%, or it contains ≥4 levels of negative descriptive words.

[0046] Therefore, even if the negative expression characters in the evaluation text account for less than 30%, as long as any of the identified negative expression words has a negative level greater than or equal to 4, the evaluation text is classified as a malicious evaluation, and no further action is required. For evaluation texts with a negative expression character ratio of <30% (and not containing negative expression words of level 4 or higher), a constructive evaluation needs to be performed using a fine-tuned model; see [link to details]. Figure 1 As shown in the description.

[0047] The method provided in the above embodiments filters malicious reviews by automatically identifying the proportion of negative descriptive words in the review text and classifies valuable reviews by performing multi-dimensional value scoring. This significantly improves the efficiency of review classification, reduces the workload of manual intervention, and facilitates the rapid selection of corresponding review texts based on the classification results.

[0048] See Figure 5 The diagram illustrates another optional online consultation evaluation processing method according to an embodiment of the present invention, including the following steps: S501: Obtain the submission time of the evaluation text, and determine a time factor based on the difference between the submission time and the current time; S502: Identify business terms related to the target business scenario in the evaluation text, and determine semantic weights based on the business terms; wherein, the semantic weights are determined based on the total number of characters of the business terms and the total number of characters of the evaluation text; S503: Calculate a comprehensive score for the evaluation text based on at least one of the time factor and semantic weight, and the value score, so as to rank the evaluation texts in the same category according to the comprehensive score.

[0049] Existing review management lacks an effective sorting mechanism. The system defaults to displaying reviews in chronological order of submission or filtering them based on patient tags (such as age, disease type, etc.). Doctors have to read each sentence carefully to obtain useful information, making it difficult to quickly locate reviews with constructive value. Furthermore, the display of negative expressions in malicious reviews in chronological order can exacerbate the negative experience for doctors and reduce their willingness to view reviews.

[0050] To address this issue, this solution optimizes the sorting of evaluation texts after classification. Specifically, invalid evaluations are sorted in reverse chronological order of submission time (facilitating the tracking of the latest invalid feedback). In one optional implementation, high-value evaluations can be sorted in descending order of constructive scores, while ordinary value evaluations can be sorted in descending order of overall score. Alternatively, both high-value and ordinary value evaluations can be sorted in descending order of overall score.

[0051] The calculation process for the overall score includes: 1. Time Factor: A quantitative indicator for measuring the timeliness of evaluations. Different scores are assigned based on the time elapsed since the evaluation was submitted, used to balance "usefulness" and "timeliness" in the ranking process. Example settings: Evaluation submission time ≤ 7 days, time factor = 10; 7-30 days, time factor = 7; > 30 days, time factor = 4. This time factor ensures that new evaluations have appropriate weight while preventing older, high-value evaluations from being completely buried, thus improving the rationality of the ranking.

[0052] 2. Semantic Weight: This is determined by the "information density" score output by the model, with an example range of 0-5 points. It addresses the misjudgment problem of "long texts being low-value" or "short texts being high-value." For example, it can identify business terms in the evaluation text that are related to the target business scenario, calculate the percentage of the total number of characters of these business terms divided by the total number of characters in the evaluation text, and use this as the semantic weight.

[0053] 3. Calculate the comprehensive score of the evaluation text based on at least one of the time factor and semantic weight, as well as the value score. Considering all three factors, the comprehensive score = value score × 0.7 + time factor × 0.3 + semantic weight × 0.1. The weights here can also be adjusted based on practical experience.

[0054] In one optional implementation, a hierarchical sorting mechanism can be set up. For example, high-value evaluations (≥6 points) account for the top 30%, with "strongly constructive" evaluations (constructive ≥8 points) identified by the model prioritized at the top; ordinary value evaluations (3-5 points) account for the middle 50%, arranged in descending order of comprehensive score; and invalid evaluations account for the bottom 20%, arranged in descending order of evaluation submission time. Taking a medical scenario as an example, the evaluation display interface on the doctor's end is divided into three independent areas, displaying high-value evaluations, ordinary value evaluations, and invalid evaluations respectively. The evaluations in each area can be scrolled in response to the doctor's operation, realizing categorized display and convenient browsing, while shielding the doctor from interference from invalid evaluations.

[0055] The method provided in the above embodiments further introduces time factors and semantic weights on the basis of value scoring to comprehensively score the evaluation text. It can effectively balance the display priority of new and old evaluations, avoid high-value but outdated evaluations from occupying high positions for a long time, and solve the problem of value misjudgment caused by differences in text length through semantic weights. This makes the evaluation ranking more reasonable and intelligent, and improves the efficiency of users obtaining valuable evaluations.

[0056] In summary, this solution identifies negative words in review texts and then uses an AI model for multi-dimensional analysis to determine whether the review text is malicious. If not malicious, it assesses the constructive value of the review text for the user (e.g., a doctor's subsequent treatment). Based on this value score, the review text is categorized to prevent users from ignoring review texts that contain both criticism and suggestions for improvement. By introducing time factors and semantic weights on top of the value score, the review text is comprehensively evaluated and displayed according to a structure of high-value reviews -> moderate-value reviews -> invalid reviews. This effectively improves the efficiency of users obtaining valuable reviews while significantly reducing the negative interference of invalid reviews, helping users efficiently extract improvement suggestions to optimize service quality.

[0057] See Figure 6 The diagram shows the main modules of an online consultation evaluation processing device 600 provided in an embodiment of the present invention, including: Classification module 601 is used to classify evaluation texts; The filtering module 602 is used to display the evaluation text corresponding to the specific category in response to the user's trigger operation on the filter key of a specific category through the human-computer interaction control.

[0058] In the apparatus of this invention, the classification module 601 is used for: Identify information in the evaluation text that corresponds to specific fields related to the target business scenario, and determine the true evaluation value based on the identification results of the specific field information; Identify semantic information in the evaluation text that is related to the target business scenario, and determine the business relevance based on the semantic information identification results; Identify improvement suggestions related to the target business scenario in the evaluation text, and determine a constructive score based on the identification results of the improvement suggestions; The value score of the evaluation text is determined based on at least one of the evaluation true value, the business relevance, and the constructiveness score, so as to determine the classification of the evaluation text according to the value score.

[0059] In the apparatus of this invention, the classification module 601 is used for: Identify negative descriptive words in the evaluation text to calculate the proportion of negative descriptive characters; If the proportion of negative expression characters is greater than or equal to a preset threshold, the evaluation text is classified as an invalid evaluation. In response to the fact that the proportion of negative expression characters is less than the preset threshold, information in the evaluation text that corresponds to a specific field related to the target business scenario is identified.

[0060] In the apparatus of this invention, the classification module 601 is used for: By using a preset sensitive word library corresponding to the target business scenario, suspected sensitive words in the evaluation text can be identified; By using a pre-defined negative expression word library corresponding to the target business scenario, negative expression words in the suspected sensitive words are identified; Based on the total number of characters in the negative expression words and the total number of characters in the evaluation text, the proportion of negative expression characters in the evaluation text is determined.

[0061] In the apparatus of this invention, the classification module 601 is further used for: Receive the negative level configured for each negative expression word in the preset negative expression lexicon; If the evaluation text contains negative words with a negative level greater than or equal to a preset level, the evaluation text is classified as an invalid evaluation.

[0062] In the apparatus of this invention, the classification module 601 is used for: The value score of the evaluation text is obtained by weighted summation based on at least one of the evaluation true value, the business relevance, and the constructive score; Determine the interval in which the value score is hit, and use the category corresponding to the interval as the category of the evaluation text.

[0063] The apparatus for implementing this invention further includes a sorting module, used for: Obtain the submission time of the evaluation text, and determine a time factor based on the difference between the submission time and the current time; Identify business terms related to the target business scenario in the evaluation text, and determine semantic weights based on the business terms; wherein, the semantic weights are determined based on the total number of characters of the business terms and the total number of characters of the evaluation text; Based on at least one of the time factor and semantic weight of the evaluation text, and the value score, a comprehensive score of the evaluation text is calculated, so as to rank the evaluation texts in the same category according to the comprehensive score.

[0064] In the apparatus of this invention, the classification includes one of high-value evaluation, ordinary-value evaluation, and invalid evaluation. The display priority of high-value evaluation is higher than that of ordinary-value evaluation, and the display priority of ordinary-value evaluation is higher than that of invalid evaluation. The sorting module is further configured to: For evaluation texts under high-value evaluation, they are sorted in descending order of constructive score; For evaluation texts under ordinary value assessment, sort them in descending order of comprehensive score; For invalid evaluations, the evaluation texts are sorted in reverse chronological order of submission time.

[0065] The apparatus of the present invention further includes a read module, which is used to: hide the target evaluation text or move the target evaluation text to the end of the evaluation text list in response to a user marking the target evaluation text in the displayed evaluation text list as read by the user through the marking interaction control, and adjust the display order of the remaining evaluation texts in the evaluation text list.

[0066] The device for implementing the present invention also includes a highlighting module, which is used to: in response to a user's viewing operation of a target evaluation text in the displayed evaluation text list through a viewing interaction control, highlight the information in the target evaluation text that is associated with the target business scenario.

[0067] Furthermore, the specific implementation details of the device described in the embodiments of the present invention have been described in detail in the above-described method, so the details will not be repeated here.

[0068] Figure 7 An exemplary system architecture 700 to which embodiments of the present invention can be applied is shown, including terminal devices 701, 702, 703, network 704, and server 705 (only an example).

[0069] Terminal devices 701, 702, and 703 can be various electronic devices with displays and support for web browsing, and have various communication client applications installed. Users can use terminal devices 701, 702, and 703 to interact with server 705 through network 704 to receive or send messages, etc.

[0070] Network 704 is a medium used to provide a communication link between terminal devices 701, 702, 703 and server 705. Network 704 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0071] Server 705 can be a server providing various services, such as a backend management server supporting shopping websites browsed by users using terminal devices 701, 702, and 703 (this is just an example). The backend management server can analyze and process received data such as product information query requests, and feed back the processing results (e.g., target push information, product information—this is just an example) to the terminal devices. It should be noted that the method provided in this embodiment of the invention is generally executed by server 705, and correspondingly, the apparatus is generally set in server 705.

[0072] It should be understood that Figure 7 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0073] The following is for reference. Figure 8It shows a schematic diagram of the structure of a computer system 800 suitable for implementing a terminal device of the present invention. Figure 8 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0074] like Figure 8 As shown, the computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 802 or programs loaded from storage section 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the system 800. The CPU 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0075] The following components are connected to I / O interface 805: an input section 806 including a keyboard, mouse, etc.; an output section 807 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to I / O interface 805 as needed. A removable medium 811, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 810 as needed so that computer programs read from it can be installed into storage section 808 as needed.

[0076] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 809, and / or installed from removable medium 811. When the computer program is executed by central processing unit (CPU) 801, it performs the functions defined above in the system of this invention.

[0077] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0078] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0079] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be housed in a processor; for example, a processor may be described as including a classification module and a filtering module. The names of these modules do not necessarily limit the module itself; for example, a filtering module may also be described as a "filter display module".

[0080] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to perform any of the online consultation evaluation processing methods described above.

[0081] The computer program product of the present invention includes a computer program that, when executed by a processor, implements the online consultation evaluation processing method in the embodiments of the present invention.

[0082] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An online consultation evaluation processing method, characterized in that, The method comprises: classifying the evaluation text; in response to a user triggering a specific classification filter key through a human-computer interaction control, displaying the evaluation text corresponding to the specific classification.

2. The method of claim 1, wherein, The classification of the evaluation text comprises: identifying information corresponding to a specific field related to a target business scenario in the evaluation text, determining an evaluation truth value based on the specific field information recognition result; identifying semantic information related to the target business scenario in the evaluation text, determining a business relevance based on the semantic information recognition result; identifying improvement suggestion information related to the target business scenario in the evaluation text, determining a constructive score based on the improvement suggestion information recognition result; determining a value score of the evaluation text based on at least one of the evaluation truth value, the business relevance, and the constructive score, to determine the classification of the evaluation text according to the value score.

3. The method of claim 2, wherein, The identification of information corresponding to a specific field related to a target business scenario in the evaluation text comprises: identifying negative expression words in the evaluation text to calculate a negative expression character proportion; in response to the negative expression character proportion being greater than or equal to a preset threshold, classifying the evaluation text as invalid evaluation; in response to the negative expression character proportion being less than the preset threshold, identifying information corresponding to a specific field related to a target business scenario in the evaluation text.

4. The method of claim 3, wherein, The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises:

5. The method according to claim 3 or 4, characterized in that, receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises:

6. The method of claim 2, wherein, receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises:

7. The method according to claim 1 or 2, characterized in that, receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The method further comprises: receiving a negative level configured for each negative expression word in the preset negative expression word library; in response to the presence of a negative expression word with a negative level greater than or equal to a preset level in the evaluation text, classifying the evaluation text as invalid evaluation. The comprehensive score of the evaluation text is calculated based on at least one of a time factor and a semantic weight of the evaluation text and the value score, so as to sort the evaluation texts under the same category according to the comprehensive score.

8. The method of claim 7, wherein, The category includes one of a high-value evaluation, a common-value evaluation and an invalid evaluation, the display priority of the high-value evaluation is greater than that of the common-value evaluation, and the display priority of the common-value evaluation is greater than that of the invalid evaluation, after the evaluation texts are classified, the method further includes: The evaluation texts under the high-value evaluation are sorted in descending order of the constructive score; The evaluation texts under the common-value evaluation are sorted in descending order of the comprehensive score; The evaluation texts under the invalid evaluation are sorted in descending order of the submission time of the evaluation.

9. The method according to claim 1 or 8, characterized in that, The method further includes: In response to a user marking a target evaluation text in the displayed list of evaluation texts through a marking interactive control, the target evaluation text is hidden or moved to the end of the list of evaluation texts, and the display order of the remaining evaluation texts in the list of evaluation texts is adjusted.

10. The method of claim 1 or 8, wherein, The method further includes: In response to a user viewing a target evaluation text in the displayed list of evaluation texts through a viewing interactive control, information associated with the target business scenario in the target evaluation text is highlighted.

11. An online consultation evaluation processing apparatus characterized by comprising: The method includes: a classification module configured to classify the evaluation texts; a screening module configured to display the evaluation texts corresponding to a specific category in response to a user triggering a specific category screening key through a human-computer interaction control.

12. An electronic device, comprising: The method includes: one or more processors; a storage device configured to store one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any one of claims 1-10.

13. A computer readable medium having stored thereon a computer program, characterized in that The program is executed by the processor to implement the method of any one of claims 1-10.

14. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1-10.

Images