Intelligent response method and device, electronic equipment and storage medium

Abstract

The present disclosure provides a smart answering method and device, electronic equipment and storage medium. The method comprises: obtaining a user question from a client; classifying the user question by a natural language understanding model to obtain a target intent category of the user question; obtaining a target composite business answer corresponding to the target intent category, the target composite business answer including general answers to different sub-class questions in the target intent category; matching a response paragraph corresponding to the user question from the target composite business answer; distinguishing the response paragraph in the target composite business answer; returning the target composite business answer to the client to display the target composite business answer including the distinguished response paragraph on the client. The method can reduce the reading difficulty of the target composite business answer, assist users in quickly locating the answer, and improve the user experience.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to an intelligent response method, apparatus, electronic device, and storage medium. Background Technology

[0002] During customer service responses, a complex business answer can usually be obtained by matching the user's question with the relevant information. However, these complex business answers are often very lengthy, making it difficult for users to read and quickly find the paragraph or sentence that corresponds to their question, resulting in a poor user experience.

[0003] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The purpose of this disclosure is to provide an intelligent response method, device, electronic device, and storage medium that can reduce the reading difficulty of the target complex business answer, assist users in quickly locating the answer, and improve the user experience.

[0005] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0006] This disclosure provides an intelligent response method, comprising: obtaining a user question from a client; classifying the user question using a natural language understanding model to obtain a target intent category of the user question; obtaining a target composite business answer corresponding to the target intent category, wherein the target composite business answer includes a general answer for different subcategories of questions within the target intent category; matching a response paragraph corresponding to the user question from the target composite business answer; displaying the response paragraphs in the target composite business answer differently; and returning the target composite business answer to the client to display the target composite business answer, including the differentiated response paragraphs, on the client.

[0007] In some exemplary embodiments of this disclosure, matching the response paragraph corresponding to the user question from the target composite business answer includes: segmenting the user question into words to obtain the segmentation result of the user question; segmenting the target composite business answer according to the punctuation in the target composite business answer to obtain the unit paragraph of the target composite business answer; determining the business key sentence according to the segmentation result and the unit paragraph; and using the business key sentence as the starting point, matching the response paragraph corresponding to the user question from the target composite business answer using regular expressions.

[0008] In some exemplary embodiments of this disclosure, determining business key sentences based on word segmentation results and unit paragraphs includes: matching target business words in the word segmentation results of the user question; if the word segmentation results of the user question include target business words, then determining the unit paragraph containing the most different target business words as the business key sentence; if the word segmentation results of the user question do not include target business words, then determining the similarity between each unit paragraph of the user question and the target composite business answer; and determining the unit paragraph with a similarity greater than a similarity threshold as the business key sentence.

[0009] In some exemplary embodiments of this disclosure, distinguishing the response paragraph in the target composite business answer includes: if the length of the response paragraph is less than a preset ratio of the length of the target composite business answer, then distinguishing the response paragraph in the target composite business answer using a first method; and distinguishing the business key sentences in the response paragraph using a second method.

[0010] In some exemplary embodiments of this disclosure, different sub-type questions include a first sub-type question and a second sub-type question; wherein, the above method further includes: counting the frequency of the response paragraphs displaying different responses to user questions to obtain the response statistics of the response paragraphs; determining a first response paragraph to be split and a second response paragraph to be split from the target composite business answer based on the response statistics, wherein the first response paragraph to be split corresponds to the first sub-type question and the second response paragraph to be split corresponds to the second sub-type question; determining the weighted difference between the first response paragraph to be split and the second response paragraph to be split; if the weighted difference exceeds a preset difference, then splitting the first response paragraph to be split and the second response paragraph to be split from the target composite business answer.

[0011] In some exemplary embodiments of this disclosure, the method further includes: if no response paragraph corresponding to the user question is matched from the target composite business answer, then counting the frequency of the response paragraph corresponding to the user question not matched from the target composite business answer to obtain the trigger count of the business key sentence; obtaining business key sentences whose trigger count exceeds a first count; clustering the user questions corresponding to the business key sentences whose trigger count exceeds the first count to obtain a first clustering result; and adding a new intent category and its corresponding business answer based on the first clustering result.

[0012] In some exemplary embodiments of this disclosure, the method further includes: if the word segmentation result of the user question does not contain the target business term, then adding the user question to the uncovered question list; counting the acquisition frequency of user questions in the uncovered question list to obtain the question count of the user questions; acquiring user questions whose question count exceeds a second count; clustering the user questions whose question count exceeds the second count to obtain a second clustering result; and adding new intent categories and their corresponding business answers based on the second clustering result.

[0013] This disclosure provides an intelligent response method, comprising: receiving a user question, obtaining a corresponding target composite service answer based on the user question, matching a response paragraph corresponding to the user question from the target composite service answer, and displaying the response paragraphs in the target composite service answer differently; and displaying the target composite service answer including the displayed response paragraphs.

[0014] This disclosure provides an intelligent response device, comprising: a user question acquisition module for acquiring user questions from a client; an intent category acquisition module for classifying user questions using a natural language understanding model to obtain a target intent category for the user questions; a business answer acquisition module for obtaining a target composite business answer corresponding to the target intent category, wherein the target composite business answer includes a general answer for different subcategories of questions within the target intent category; a response paragraph matching module for matching response paragraphs corresponding to the user questions from the target composite business answer; a differentiation display module for differentiating and displaying the response paragraphs in the target composite business answer; and a business answer return module for returning the target composite business answer to the client to display the target composite business answer, including the differentiated response paragraphs, on the client.

[0015] This disclosure provides an intelligent response device, comprising: a user question receiving module, configured to receive user questions, obtain corresponding target composite business answers based on user questions, match response paragraphs corresponding to user questions from target composite business answers, and distinguish and display the response paragraphs in target composite business answers; and a business answer display module, configured to display the target composite business answer including the distinguished response paragraphs.

[0016] This disclosure provides an electronic device, including: at least one processor; and a storage device for storing at least one program, which, when executed by the at least one processor, causes the at least one processor to implement any of the above-described smart response methods.

[0017] This disclosure provides a computer-readable storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements any of the intelligent response methods described above.

[0018] The intelligent response method provided in some embodiments of this disclosure can obtain the corresponding target composite business answer based on the user's question. The response paragraph corresponding to the user's question can be accurately matched from the target composite business answer. The response paragraphs in the target composite business answer can be displayed differently, which can reduce the reading difficulty of the target composite business answer, help users quickly locate the answer, and improve the user experience.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0021] Figure 1 A schematic diagram of an exemplary system architecture to which the smart response method of embodiments of the present disclosure can be applied is shown.

[0022] Figure 2 This is a flowchart illustrating an intelligent response method according to an exemplary embodiment.

[0023] Figure 3 This is a schematic diagram illustrating a type of display of a composite business answer in related technologies.

[0024] Figure 4 This is a schematic diagram of an interface for a target composite business answer that includes distinctly displayed response paragraphs, according to an exemplary embodiment.

[0025] Figure 5 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0026] Figure 6 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0027] Figure 7 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0028] Figure 8 This is a schematic diagram of an interface for a target composite business answer that includes distinctly displayed response paragraphs, according to an exemplary embodiment.

[0029] Figure 9This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0030] Figure 10 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0031] Figure 11 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0032] Figure 12 This is a flowchart illustrating an intelligent response method according to an exemplary embodiment.

[0033] Figure 13 This is a block diagram illustrating a smart response device according to an exemplary embodiment.

[0034] Figure 14 This is a block diagram illustrating a smart response device according to an exemplary embodiment.

[0035] Figure 15 This is a schematic diagram of the structure of an electronic device according to an exemplary embodiment. Detailed Implementation

[0036] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0037] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0038] Figure 1 A schematic diagram of an exemplary system architecture to which the smart response method of embodiments of the present disclosure can be applied is shown.

[0039] like Figure 1As shown, the system architecture may include a server 101, a network 102, and a client 103. The network 102 serves as the medium for providing a communication link between the client 103 and the server 101. The network 102 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0040] Server 101 can be a server that provides various services, such as a backend management server that supports the device operated by the user using client 103. The backend management server can analyze and process received requests and other data, and feed the processing results back to client 103.

[0041] Client 103 can be a mobile terminal such as a mobile phone, game console, tablet computer, e-book reader, smart glasses, smart home device, AR (Augmented Reality) device, VR (Virtual Reality) device, etc. Alternatively, client 103 can also be a personal computer, such as a laptop computer and a desktop computer, etc.

[0042] For example, a user can input a user question through client 103. Client 103 can process the user question, obtain the corresponding target composite business answer based on the user question, match the response paragraph corresponding to the user question from the target composite business answer, distinguish and display the response paragraph in the target composite business answer, and display the target composite business answer including the distinguished response paragraph.

[0043] Client 103 can also send user questions to server 101. Server 101 can obtain user questions from client 103. Server 101 can classify user questions using a natural language understanding model to obtain the target intent category of the user question. Server 101 can obtain the target composite business answer corresponding to the target intent category. The target composite business answer includes the general answer for different subcategories of questions in the target intent category. Server 101 can match the response paragraph corresponding to the user question from the target composite business answer. Server 101 can differentiate and display the response paragraph in the target composite business answer. Server 101 can return the target composite business answer to client 103 so that client 103 can display the target composite business answer including the differentiated response paragraph.

[0044] It should be understood that Figure 1 The number of clients, networks, and servers in this example is merely illustrative. Server 101 can be a single physical server, a server cluster consisting of multiple servers, or a cloud server. Depending on actual needs, it can have any number of terminal devices, networks, and servers.

[0045] The steps of the intelligent response method in the exemplary embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings and examples.

[0046] Figure 2 This is a flowchart illustrating an intelligent response method according to an exemplary embodiment. The method provided in this disclosure can be derived from, for example... Figure 1 This disclosure may be executed on the server or client shown, but is not limited thereto.

[0047] like Figure 2 As shown, the intelligent response method provided in this embodiment may include the following steps.

[0048] In step S202, the user question is obtained from the client.

[0049] In this embodiment of the disclosure, the user can input a user question through the client, the client can obtain and process the user question, or the client can send the user question to the server, and the server can process the user question after obtaining it from the client.

[0050] For example, user questions could include "View pickup code", "Why can't I combine payments", "I want to modify order information", etc.

[0051] In step S204, the user's question is classified using a natural language understanding model to obtain the target intent category of the user's question.

[0052] Among them, the natural language understanding model can be an NLU (Natural Language Understanding) model.

[0053] In this embodiment of the disclosure, the NLU model can be trained based on the user question training set and the target intent category training set, so that the NLU model can accurately classify user questions.

[0054] The intent category can be set as needed. For example, in an online shopping scenario, the intent category can include self-pickup policy related matters, order merging inquiries, order modification inquiries, etc.

[0055] For example, by classifying the user question "View pickup code" using an NLU model, we can obtain the target intent category of "self-pickup policy related"; by classifying the user question "Why can't I combine payments" using an NLU model, we can obtain the target intent category of "order merging consultation"; and by classifying the user question "I want to modify order information" using an NLU model, we can obtain the target intent category of "order modification consultation".

[0056] In step S206, the target composite business answer corresponding to the target intent category is obtained. The target composite business answer includes the general answer to different sub-category questions in the target intent category.

[0057] Among them, different sub-type questions can be sub-type questions with different frequencies, such as high-frequency sub-type questions and low-frequency sub-type questions; different sub-type questions can also be sub-type questions with different content. For example, when the target intent category is "modify order inquiry", it can include two sub-type questions: "modify order address inquiry" and "modify order time inquiry".

[0058] In this embodiment of the disclosure, corresponding composite service answers can be pre-set for intent categories. Based on the pre-set correspondence between intent categories and composite service answers, the target composite service answer corresponding to the target intent category can be obtained.

[0059] For example, user questions, the target intent categories corresponding to user questions, and the target composite business answers corresponding to target intent categories can be shown in Table 1.

[0060] Table 1

[0061]

[0062]

[0063] In step S208, the response paragraph corresponding to the user's question is matched from the target composite business answer.

[0064] The response paragraph can be the statement in the target composite business answer that corresponds to the user's question.

[0065] In this embodiment of the disclosure, the target composite business answer can be divided into multiple unit paragraphs according to the punctuation in the target composite business answer; according to the matching degree between the user question and the unit paragraphs, the unit paragraph with the highest matching degree with the user question can be taken as the response paragraph corresponding to the user question; alternatively, the user question can be segmented into words, and the business key sentence can be determined according to the word segmentation result of the user question and the unit paragraphs, and the response paragraph corresponding to the user question can be obtained by matching with the business key sentence as the starting point.

[0066] For example, if the answer to the target complex business scenario is, "You can view the pickup code in the logistics information on the order details page. If it has arrived at the self-pickup point / locker, the pickup code will be displayed on the order details page, and you will also receive a pickup code SMS. If you do not receive the pickup code SMS, it may be that the order has not yet arrived at the self-pickup point / locker, please wait patiently. If the pickup code is not displayed on the order page, you can directly contact the delivery person for consultation. We apologize for any inconvenience caused.", the response paragraph could be: "If you have not received the pickup code SMS, it may be that the order has not yet arrived at the self-pickup point / locker, please wait patiently."

[0067] In step S210, the response paragraphs in the target composite business answer are displayed separately.

[0068] The method of distinguishing the response paragraphs in the target composite business answer may include at least one of the following: highlighting the response paragraphs in the target symbol business answer, color-coding the response paragraphs in the target composite business answer, bolding the response paragraphs in the target composite business answer, and underlining the response paragraphs in the target composite business answer.

[0069] In step S212, the target composite business answer is returned to the client to display the target composite business answer, including the differentiated response paragraph, on the client.

[0070] In this embodiment of the disclosure, a target composite service answer including a distinct response paragraph can be sent to a client, and the client can receive and display the target composite service answer including the distinct response paragraph.

[0071] Figure 3 This is a schematic diagram illustrating a type of display of a composite business answer in related technologies.

[0072] Figure 4 This is a schematic diagram of an interface for a target composite business answer that includes distinctly displayed response paragraphs, according to an exemplary embodiment.

[0073] refer to Figure 3 and Figure 4 In this embodiment of the disclosure, the response paragraphs in the target composite business answer are displayed separately, which can help users quickly locate the answer and reduce the difficulty of reading for users.

[0074] The intelligent response method provided in this disclosure can obtain the corresponding target composite business answer based on the user's question. It can accurately match the response paragraph corresponding to the user's question from the target composite business answer and display the response paragraph in the target composite business answer differently. This can reduce the reading difficulty of the target composite business answer, help users quickly locate the answer, and improve the user experience.

[0075] Figure 5 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0076] In this embodiment of the disclosure, with Figure 2 The intelligent response method shown is different from that of the other methods. Figure 5 The intelligent response method shown further provides how to match the response paragraph corresponding to the user's question from the target composite business answer, that is, it provides an embodiment of the above step S208.

[0077] Figure 5 The intelligent response method shown may include the following steps.

[0078] In step S502, the user question is segmented into words to obtain the segmentation results of the user question.

[0079] In this embodiment of the disclosure, the user's question can be segmented using Jieba segmentation to obtain the segmentation result of the user's question.

[0080] For example, by segmenting the user question "view pickup code", the word segmentation result of the user question can be "view, pickup code"; by segmenting the user question "why did I not receive the pickup code", the word segmentation result of the user question can be "no, received, pickup code".

[0081] In step S504, the target composite business answer is segmented according to the punctuation in the target composite business answer to obtain the unit paragraphs of the target composite business answer.

[0082] Punctuation marks can include, but are not limited to, periods, commas, semicolons, ellipses, etc.

[0083] In this embodiment of the disclosure, the target composite business answer can be segmented using punctuation marks in the target composite business answer as boundaries to obtain unit paragraphs of the target composite business answer.

[0084] For example, using the commas and periods as boundaries in the target composite business answer, "You can view the pickup code in the logistics information on the order details page. If it has arrived at the self-pickup point / locker, the pickup code will be displayed on the order details page, and you will also receive a pickup code SMS. If you do not receive the pickup code SMS, it may be because the order has not yet arrived at the self-pickup point / locker. Please wait patiently. If the pickup code is not displayed on the order page, you can directly contact the delivery person for consultation. We apologize for any inconvenience caused.", the target composite business answer can be divided into the following unit paragraphs:

[0085] "You can view the pickup code in the logistics information on the order details page,"

[0086] "If you have already arrived at the self-pickup point / locker,"

[0087] "The order details page will display the pickup code,"

[0088] "You will also receive a pickup code via SMS."

[0089] "If you have not received the pickup code via SMS,"

[0090] "It's possible that the ordered items haven't arrived at the pickup point / locker yet."

[0091] Please wait patiently.

[0092] "If the order page does not display the pickup code,"

[0093] "You can contact the delivery person directly for inquiries,"

[0094] We apologize for any inconvenience this may cause.

[0095] In step S506, key business sentences are determined based on word segmentation results and unit paragraphs.

[0096] Among them, the key business sentence can be one of the paragraphs in the target compound business answer.

[0097] In this embodiment of the disclosure, key business sentences can be determined from unit paragraphs based on the word segmentation results and the degree of matching between unit paragraphs.

[0098] Taking the above paragraph as an example, "You can view the pickup code in the logistics information on the order details page," can be identified as the key business sentence.

[0099] In step S508, starting from the key business sentence, regular expressions are used to match and obtain the response paragraph corresponding to the user's question from the target composite business answer.

[0100] In this embodiment of the disclosure, a question-response sentence template can be pre-set. Based on the question-response sentence template module, regular expressions are used to select sentences from the target composite business answers that start with the business key sentence and end with the first period after the business key sentence as the response paragraph corresponding to the user question.

[0101] The question-response template can be shown in Table 2.

[0102] Table 2

[0103]

[0104] For example, the composite business answer, key business sentences, and response paragraphs corresponding to user questions can be shown in Table 3.

[0105] Table 3

[0106]

[0107]

[0108] Figure 6 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0109] In this embodiment of the disclosure, with Figure 5 The intelligent response method shown is different from that of the other methods. Figure 6 The intelligent response method shown further provides how to determine key business sentences based on word segmentation results and unit paragraphs, that is, it provides an embodiment of the above step S506.

[0110] Figure 6 The intelligent response method shown may include the following steps.

[0111] In step S602, the target business words in the word segmentation results of the user's question are matched.

[0112] In this embodiment of the disclosure, a business term library can be pre-set and updated regularly. The word segmentation result of the user's question can be determined based on the business term library to determine whether the target business term is included. If the word segmentation result of the user's question includes the target business term, step S604 can be executed; if the word segmentation result of the user's question does not include the target business term, step S606 can be executed.

[0113] In this embodiment of the disclosure, online user corpus can be acquired, the user corpus can be segmented into words to obtain the segmentation results of the user corpus, the business weight of each segmented word in the segmentation results of the user corpus can be determined according to the business word formula, and the segmented words with preset threshold business weights can be added to the business word lexicon.

[0114] The business weights can be determined according to the following formula:

[0115] Weight(w)=WordTotal(w) / Intent(w)*lg(Intent(w))

[0116] Where w represents a segmented word in the segmentation results of the user corpus, Weight(w) represents the business weight of segmented word w, WordTotal(w) represents the total number of occurrences of segmented word w, and Intent(w) represents the number of sentences containing segmented word w and having intent.

[0117] For example, the business term database may include: "search, view, check, pickup code, send, obtain, express delivery, pick up, collect, information, door-to-door, delivery, no, received, not sent, logistics, arrived, order, cannot be found, buy, purchase, payment, merge, pay, less, goods, replenish, ...".

[0118] In step S604, if the word segmentation result of the user's question includes target business words, then the unit paragraph containing the most different target business words is determined as the business key sentence.

[0119] For example, if the word segmentation result of a user's question is "view, pickup code", and the above business word lexicon includes the two business words "view" and "pickup code", then the unit paragraph containing the most of the two target business words "view" and "pickup code" can be identified as the business key sentence. Taking the unit paragraph of the above response paragraph as an example, the business key sentence can be "You can view the pickup code in the logistics information on the order details page".

[0120] In step S606, if the word segmentation result of the user question does not contain the target business term, then the similarity of each unit paragraph of the user question and the target composite business answer is determined.

[0121] For example, if a user's question is "Why haven't I received the pickup information?", the word segmentation result of the user's question is "received, pickup information". The above business word lexicon does not contain "received" and "pickup information", that is, the word segmentation result of the user's question does not contain the target business word. Therefore, the similarity between the various unit paragraphs in the user's question and the target composite business answer can be determined.

[0122] The similarity model can be invoked, or the DSSM (Deep Structured Semantic Model) can be used to calculate the similarity between each unit paragraph in the user question and the target composite business answer. Other methods can also be used to calculate the similarity, and this disclosure does not restrict this.

[0123] In step S608, the unit paragraphs with a similarity greater than the similarity threshold are identified as key business sentences.

[0124] The similarity threshold can be set according to the actual situation.

[0125] For example, if a user's question is "Why haven't I received a pickup notification?", the corresponding paragraphs in the target composite business answer would be:

[0126] "You can view the pickup code in the logistics information on the order details page,"

[0127] "If you have already arrived at the self-pickup point / locker,"

[0128] "The order details page will display the pickup code,"

[0129] "You will also receive a pickup code via SMS."

[0130] "If you have not received the pickup code via SMS,"

[0131] "It's possible that the ordered items haven't arrived at the pickup point / locker yet."

[0132] Please wait patiently.

[0133] "If the order page does not display the pickup code,"

[0134] "You can contact the delivery person directly for inquiries,"

[0135] We apologize for any inconvenience this may cause.

[0136] For example, the similarity between the user question "Why haven't I received the pickup information?" and the above-mentioned paragraph units can be calculated. The paragraph units with a similarity greater than the similarity threshold can be used as business key sentences. The business key sentence could be "If you haven't received the pickup code SMS,..."

[0137] In this embodiment of the disclosure, the key business sentences can be accurately determined based on the word segmentation results of the user's question and the unit paragraphs of the compound business answer, which facilitates accurate matching to obtain the response answer in the future.

[0138] Figure 7 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0139] In this embodiment of the disclosure, with Figure 2 The intelligent response method shown is different from that of the other methods. Figure 6 The intelligent response method shown further provides how to differentiate and display response paragraphs in the target composite business answer, that is, it provides an embodiment of the above step S210.

[0140] Figure 7 The intelligent response method shown may include the following steps.

[0141] In step S702, the length of the response paragraph is calculated as a proportion of the length of the target composite business answer.

[0142] In this embodiment of the disclosure, the text length of the response paragraph and the text length of the target composite business answer can be calculated, and the text length of the response paragraph can be divided by the text length of the target composite business answer to obtain the ratio of the length of the response paragraph to the length of the target composite business answer.

[0143] In step S704, if the length of the response paragraph is greater than or equal to the length of the target composite business answer, the response paragraph in the target composite business answer will not be displayed separately.

[0144] The preset ratio can be set according to actual needs; for example, it can be set to 60%.

[0145] For example, if the length of the response paragraph accounts for 60% or more of the length of the target composite business answer, then the response paragraph in the target composite business answer will not be displayed separately. Therefore, excessively long separate displays, such as excessively long highlighted displays, may actually negatively impact the user's reading experience.

[0146] In step S706, if the length of the response paragraph is less than the length of the target composite business answer, the response paragraph in the target composite business answer is displayed differently using the first method.

[0147] The first method can be one of the following: highlighting, color annotation, bolding, or underlining. For example, the first method is bolding.

[0148] For example, if the length of the response paragraph accounts for less than 60% of the length of the target composite business answer, then the response paragraph in the target composite business answer should be bolded.

[0149] Figure 8 This is a schematic diagram of an interface for a target composite business answer that includes distinctly displayed response paragraphs, according to an exemplary embodiment.

[0150] refer to Figure 8 You can bold the response paragraph in the target compound business answer: "If you have not received the pickup code SMS, it may be because the order has not yet arrived at the self-pickup point / locker. Please wait patiently."

[0151] In step S708, the key business sentences in the response paragraph are displayed differently using the second method.

[0152] The second method can be one of the following that differs from the first method: highlighting, color annotation, bolding, or underlining. For example, the second method is to add an underline.

[0153] refer to Figure 8 You can respond to the key business sentence in the paragraph, "If you have not received the pickup code SMS, it may be because the ordered goods have not yet arrived at the self-pickup point / self-pickup locker," by adding an underline.

[0154] In this embodiment of the disclosure, a first method is used to distinguish and display the response paragraphs in the target composite business answer, and a second method is used to distinguish and display the key business sentences in the response paragraphs, which allows users to see the key business sentences and response paragraphs more quickly and accurately.

[0155] Figure 9 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0156] In this embodiment of the disclosure, Figure 2 Based on the intelligent response method shown, Figure 9 The intelligent response method shown may include the following steps.

[0157] In step S902, the frequency of the response paragraphs displaying different user questions is counted to obtain the number of responses in the response paragraphs.

[0158] In this embodiment of the disclosure, a timed statistical task can be set to count the frequency of different display of user questions in each response paragraph of the target composite business answer, and use this frequency as the statistical count of the corresponding response paragraph. The timed trigger can be daily.

[0159] In this embodiment of the disclosure, after matching and obtaining the response paragraph, the unique identifier of the response paragraph can be obtained. The unique identifier of the response paragraph and the text of the response paragraph can be stored, or the unique identifier of the response paragraph and the text of the response paragraph can be calculated using MD5 (Message-Digest Algorithm) and stored. The response statistics corresponding to the response paragraph are incremented by one, and the user question corresponding to the response paragraph is inserted into the question record list corresponding to the response paragraph, so that this information can be sent to the operations personnel for analysis later.

[0160] In step S904, based on the number of responses, the first response segment to be split and the second response segment to be split are determined from the target composite business answers. The first response segment to be split corresponds to the first sub-category question, and the second response segment to be split corresponds to the second sub-category question.

[0161] In an exemplary embodiment, the different subclasses of problems include a first subclass problem and a second subclass problem.

[0162] In this embodiment of the disclosure, the response segment with the highest number of response statistics and the response segment with the second highest number of response statistics can be regarded as the first response segment to be split and the second response segment to be split, respectively.

[0163] For example, the first sub-category question is a high-frequency sub-category question, and the second sub-category question is a low-frequency sub-category question; or, when the target intent category is "modify order inquiry", the first sub-category question can be "modify order address inquiry", and the second sub-category question can be "modify order time inquiry".

[0164] Taking "order modification inquiry" as an example, the corresponding target composite business answer for this target intent category could be: "You can click the 'Modify Order' button on the order details page to modify the order information. If you want to modify the order address, you can click the 'Modify Order Address' button to enter the address modification interface. If you want to modify the order delivery time, you can click the 'Estimated Delivery Time' button to modify it."

[0165] The response paragraphs for this objective-based composite business answer could be: "You can click the 'Modify Order' button on the order details page to modify your order information.", "If you want to modify the order address, you can click the 'Modify Order Address' button to enter the address modification interface.", "If you want to modify the order delivery time, you can click the 'Estimated Delivery Time' button to make the modification."

[0166] For example, if the response paragraph "If you want to modify your order address, you can click the 'Modify Order Address' button to enter the address modification interface." has the highest number of responses, then "If you want to modify your order address, you can click the 'Modify Order Address' button to enter the address modification interface." will be designated as the first response paragraph to be split. If the response paragraph "If you want to modify your order delivery time, you can click the 'Estimated Delivery Time' button to modify it." has the second highest number of responses, then "If you want to modify your order delivery time, you can click the 'Estimated Delivery Time' button to modify it." will be designated as the second response paragraph to be split.

[0167] In step S906, the weighted difference between the first response segment to be split and the second response segment to be split is determined.

[0168] In this embodiment of the disclosure, the weighted difference between the first response segment to be split and the second response segment to be split can be determined by the following formula:

[0169] Weighted difference = lg(topTrigger(2)) × ((topTrigger(1) - topTrigger(2)) / topTrigger(1))

[0170] Where topTrigger(1) represents the number of responses to the first response segment to be split, and topTrigger(2) represents the number of responses to the second response segment to be split.

[0171] For example, by substituting the response counts of the first response paragraph to be split, "If you want to modify the order address, you can click the 'Modify Order Address' button to enter the address modification interface.", and the response counts of the second response paragraph to be split, "If you want to modify the order delivery time, you can click the 'Estimated Delivery Time' button to modify it.", into the above formula, we can obtain the weighted difference between the first and second response paragraphs to be split.

[0172] In step S908, if the weighted difference exceeds the preset difference, the first and second response paragraphs to be split are separated from the target composite business answer.

[0173] The preset difference can be set as needed.

[0174] For example, if the weighted difference between the first paragraph to be split, "If you want to modify the order address, you can click the 'Modify Order Address' button to enter the address modification interface," and the second paragraph to be split, "If you want to modify the order delivery time, you can click the 'Estimated Delivery Time' button to modify it," exceeds a preset difference, then the first paragraph to be split, "If you want to modify the order address, you can click the 'Modify Order Address' button to enter the address modification interface," and the second paragraph to be split, "If you want to modify the order delivery time, you can click the 'Estimated Delivery Time' button to modify it," can be separated from the target composite business answer.

[0175] The first segment of the response to be split, "If you want to change your order address, you can click the 'Change Order Address' button to enter the address modification interface," corresponds to the intent category of "Order address modification inquiry." The second segment, "If you want to change your order delivery time, you can click the 'Estimated Delivery Time' button to modify it," corresponds to the intent category of "Order time modification inquiry." Therefore, when subsequently obtaining user questions, if the target intent category of the user question is "Order address modification inquiry," the target business answer can be matched as "If you want to change your order address, you can click the 'Change Order Address' button to enter the address modification interface."; if the target intent category of the user question is "Order time modification inquiry," the target business answer can be matched as "If you want to change your order delivery time, you can click the 'Estimated Delivery Time' button to modify it."

[0176] In this embodiment of the disclosure, if the weighted difference exceeds the preset difference, the first and second response segments to be split and their corresponding user questions can be sent to the operations personnel. The operations personnel can then decide whether to split the first and second response segments to be split.

[0177] In this embodiment of the disclosure, based on the number of responses in each response segment of the composite business answer, it can be determined whether to separate independent intent categories and business answers from the composite business answer. This can improve the rationality of intelligent response classification, obtain intent categories and business answers that are more accurately matched to user questions, and provide more accurate business answers in user interactions, thereby improving user experience.

[0178] Figure 10 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0179] In this embodiment of the disclosure, Figure 5 In the intelligent response method shown, after determining the key business sentences based on word segmentation results and unit paragraphs, Figure 10 The intelligent response method shown may include the following steps.

[0180] In step S1002, if no response paragraph corresponding to the user question is obtained from the target composite business answer, the frequency of the response paragraph corresponding to the user question not being obtained from the target composite business answer is counted to obtain the number of trigger statistics of business key sentences.

[0181] In this embodiment of the disclosure, if a business key sentence corresponding to a user question is determined, but no corresponding response paragraph is found in the target composite business answer, this situation can be classified as feedback not covered. The unique identifier of the business key sentence can be obtained, the unique identifier of the business key sentence and the MD5 of the business key sentence text can be calculated, the trigger count of the business key sentence can be incremented by one, and the user question corresponding to the response paragraph can be inserted into the question record list corresponding to the business key sentence, so that this information can be sent to the operations personnel for analysis later.

[0182] In this embodiment of the disclosure, a timed statistical task can be set to count the frequency of the response paragraph corresponding to the user question that is not matched from the target composite business answer, and the frequency is used as the number of trigger statistics for the business key sentence corresponding to the user question.

[0183] It should be noted that if the business key sentence corresponding to the user's question is determined, and the corresponding response paragraph is obtained by matching it from the target composite business answer, then it does not fall under the case of feedback not being covered, and it is not necessary to add one to the trigger count of that business key sentence.

[0184] In step S1004, the business key sentence that triggers the statistical quantity to exceed the first statistical quantity is obtained.

[0185] The first statistical quantity can be set according to the actual situation.

[0186] If the number of trigger statistics for a business key sentence exceeds the first statistical number, it can be considered that there are many user questions corresponding to that business key sentence and no corresponding response paragraph.

[0187] In step S1006, user questions corresponding to business key sentences whose number of trigger statistics exceeds the first number of statistics are clustered to obtain the first clustering result.

[0188] In this embodiment of the disclosure, the CURE (Clustering Using Representatives) text clustering algorithm can be used to cluster the user questions corresponding to business key sentences that trigger statistical counts exceeding a first statistical count, and obtain the first clustering result. Other clustering algorithms can also be used for clustering, and this disclosure does not limit this.

[0189] In step S1008, based on the first clustering result, new intent categories and their corresponding business answers are added.

[0190] In this embodiment of the disclosure, the first clustering result and the corresponding business keywords can be sent to the operations personnel to assist them in analysis and to add new intent categories and their corresponding business answers in a targeted manner; alternatively, the analysis can be performed automatically based on the first clustering result to add new intent categories and their corresponding business answers.

[0191] Figure 11 This is a flowchart illustrating another smart response method according to an exemplary embodiment.

[0192] In this embodiment of the disclosure, Figure 6 Based on the intelligent response method shown, Figure 11 The intelligent response method shown may include the following steps.

[0193] In step S1102, if the word segmentation result of the user question does not contain the target business term, the user question is added to the list of uncovered questions.

[0194] In this embodiment of the disclosure, if the word segmentation result of a user question does not contain the target business term, this situation can be classified as an uncovered question. The user question can be added to the uncovered question list to facilitate the subsequent addition of intent categories and business answers to the user question.

[0195] In step S1104, the frequency of user questions in the uncovered question list is counted to obtain the total number of user questions.

[0196] In this embodiment of the disclosure, a timed statistical task can be set to count the frequency of access to each user's question in the uncovered question list, and this frequency can be used as the statistical count of each user's question. The access frequency of each user's question can be the number of times the user's question is accessed; that is, the number of times each user's question is accessed can be entered for different users.

[0197] In this embodiment of the disclosure, if the word segmentation result of the user question does not contain the target business word, the number of questions corresponding to the user question can be incremented by one.

[0198] In step S1106, user issues whose number of issues exceeds the second statistical number are obtained.

[0199] The second statistical quantity can be set according to the actual situation.

[0200] If the number of user questions collected exceeds the second statistical count, it can be assumed that the user has many questions about that user question and there are no corresponding key business sentences and response paragraphs.

[0201] In step S1108, user issues whose statistical number exceeds the second statistical number are clustered to obtain the second clustering result.

[0202] In this embodiment of the disclosure, the CURE text clustering algorithm can be used to cluster user questions whose question count exceeds the second statistical count to obtain the second clustering result. Other clustering algorithms can also be used for clustering, and this disclosure does not limit this.

[0203] In step S1110, based on the second clustering result, new intent categories and their corresponding business answers are added.

[0204] In this embodiment of the disclosure, the second clustering result can be sent to the operations personnel to assist them in analysis and to add new intent categories and their corresponding business answers in a targeted manner; alternatively, the analysis can be performed automatically based on the second clustering result to add new intent categories and their corresponding business answers.

[0205] In this embodiment of the disclosure, uncovered user issues can be quickly identified based on clustering results, eliminating the need for manual analysis and mining of issues from millions of user issues daily. Furthermore, new intent categories and corresponding business answers can be added to uncovered user issues, thereby improving the coverage of user issue responses, enhancing the efficiency of business answer optimization, and improving user experience.

[0206] Figure 12 This is a flowchart illustrating an intelligent response method according to an exemplary embodiment. The method provided in this disclosure can be derived from, for example... Figure 1 This can be executed in the client shown, but this disclosure is not limited thereto.

[0207] like Figure 12 As shown, the intelligent response method provided in this embodiment may include the following steps.

[0208] In step S1202, a user question is received to obtain the corresponding target composite service answer based on the user question. The response paragraph corresponding to the user question is matched from the target composite service answer, and the response paragraph in the target composite service answer is displayed separately.

[0209] In this embodiment of the disclosure, the user can input a user question through the client, the client can receive the user question, the client can obtain the corresponding target composite business answer based on the user question, match the response paragraph corresponding to the user question from the target composite business answer, and distinguish and display the response paragraph in the target composite business answer. The specific steps can be referred to the above embodiment; the client can also send the user question to the server, the server processes the user question, and obtain the target composite business answer including the distinguished response paragraph.

[0210] In step S1204, the target composite business answer, including the differentiated response paragraph, is displayed.

[0211] In this embodiment of the disclosure, the client can display the target composite service answer after receiving the response paragraph with a distinct display; the client can also receive the target composite service answer with a distinct display sent by the server and display it.

[0212] The interface diagram for the target composite business answer, which includes differentiated display of response paragraphs, can be shown as follows: Figure 4 or Figure 8 As shown.

[0213] It should be noted that the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0214] The following are embodiments of the apparatus disclosed herein, which can be used to execute embodiments of the method disclosed herein. For details not disclosed in the apparatus embodiments of this disclosure, please refer to the embodiments of the method disclosed herein.

[0215] Figure 13 This is a block diagram illustrating a smart response device according to an exemplary embodiment.

[0216] like Figure 13 As shown, the intelligent response device 1300 may include: a user question acquisition module 1302, an intent category acquisition module 1304, a business answer acquisition module 1306, a response paragraph matching module 1308, a differentiation display module 1310, and a business answer return module 1320.

[0217] Specifically, the user question acquisition module 1302 can be used to acquire user questions from the client; the intent category acquisition module 1304 can be used to classify user questions through a natural language understanding model to obtain the target intent category of the user questions; the business answer acquisition module 1306 can be used to obtain the target composite business answer corresponding to the target intent category, which includes the general answer for different sub-categories of questions in the target intent category; the response paragraph matching module 1308 can be used to match the response paragraphs corresponding to the user questions from the target composite business answers; the differentiation display module 1310 can be used to differentiate the response paragraphs in the target composite business answers; and the business answer return module 1312 can be used to return the target composite business answer to the client so that the target composite business answer including the differentiated response paragraphs can be displayed on the client.

[0218] In an exemplary embodiment, the response paragraph matching module 1308 may include: a word segmentation result obtaining unit, which can be used to segment the user question into words to obtain the word segmentation result of the user question; a business answer segmentation unit, which can be used to segment the target composite business answer according to the punctuation in the target composite business answer to obtain the unit paragraph of the target composite business answer; a business key sentence determination unit, which can be used to determine the business key sentence according to the word segmentation result and the unit paragraph; and a response paragraph matching unit, which can be used to use regular expressions to match the response paragraph corresponding to the user question from the target composite business answer, starting from the business key sentence.

[0219] In an exemplary embodiment, the business key sentence determination unit may include: matching target business words in the word segmentation results of the user question; a first business key sentence determination subunit, which can be used to determine the unit paragraph containing the most different target business words as the business key sentence if the word segmentation results of the user question include target business words; a similarity determination unit, which can be used to determine the similarity of each unit paragraph of the user question and the target composite business answer if the word segmentation results of the user question do not contain target business words; and a second business key sentence determination subunit, which can be used to determine the unit paragraph with a similarity greater than a similarity threshold as the business key sentence.

[0220] In an exemplary embodiment, the differentiation display module 1310 may include: a first differentiation display unit, which can be used to differentiate the response paragraph in the target composite business answer using a first method if the length of the response paragraph is less than a preset ratio of the length of the target composite business answer; and a second differentiation display unit, which can be used to differentiate the business key sentences in the response paragraph using a second method.

[0221] In an exemplary embodiment, different sub-type questions include a first sub-type question and a second sub-type question; wherein, the apparatus 1300 may further include: a response statistics quantity acquisition module, which can be used to count the frequency of the response paragraphs responding to user questions and displaying them differently, and to obtain the response statistics quantity of the response paragraphs; a response paragraph to be split determination module, which can be used to determine a first response paragraph to be split and a second response paragraph to be split from the target composite business answer based on the response statistics quantity, wherein the first response paragraph to be split corresponds to the first sub-type question, and the second response paragraph to be split corresponds to the second sub-type question; a weighted difference determination module, which can be used to determine the weighted difference between the first response paragraph to be split and the second response paragraph to be split; and a response paragraph splitting module, which can be used to split the first response paragraph to be split and the second response paragraph to be split from the target composite business answer if the weighted difference exceeds a preset difference.

[0222] In an exemplary embodiment, the apparatus 1300 may further include: a trigger statistics acquisition module, configured to, if no matching response paragraph corresponding to a user question is obtained from the target composite business answer, count the frequency of the number of times a trigger statistics of a business key sentence is obtained; a business key sentence acquisition module, configured to acquire business key sentences whose trigger statistics exceed a first statistics; a first clustering result acquisition module, configured to cluster the user questions corresponding to business key sentences whose trigger statistics exceed the first statistics, and obtain a first clustering result; and a first addition module, configured to, based on the first clustering result, add new intent categories and their corresponding business answers.

[0223] In an exemplary embodiment, the apparatus 1300 may further include: a user question adding module, configured to add the user question to an uncovered question list if the word segmentation result of the user question does not contain the target business term; a question count obtaining module, configured to count the acquisition frequency of user questions in the uncovered question list to obtain the question count of the user questions; a question acquisition module, configured to acquire user questions whose question count exceeds a second count; a second clustering result obtaining module, configured to cluster the user questions whose question count exceeds the second count to obtain a second clustering result; and a second adding module, configured to add new intent categories and their corresponding business answers based on the second clustering result.

[0224] Figure 14 This is a block diagram illustrating a smart response device according to an exemplary embodiment.

[0225] like Figure 14 As shown, the intelligent response device 1400 may include: a user question receiving module 1402 and a business answer display module 1404.

[0226] The user question receiving module 1402 can be used to receive user questions, obtain corresponding target composite business answers based on user questions, match the response paragraphs corresponding to user questions from the target composite business answers, and display the response paragraphs in the target composite business answers differently; the business answer display module 1404 can be used to display the target composite business answer including the distinguished response paragraphs.

[0227] It should be noted that the block diagrams shown in the above figures are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0228] Figure 15 This is a schematic diagram illustrating the structure of an electronic device according to an exemplary embodiment. It should be noted that... Figure 15 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0229] like Figure 15 As shown, the electronic device 1500 includes a central processing unit (CPU) 1501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1502 or a program loaded from a storage section 1508 into a random access memory (RAM) 1503. The RAM 1503 also stores various programs and data required for the operation of the system 1500. The CPU 1501, ROM 1502, and RAM 1503 are interconnected via a bus 1504. An input / output (I / O) interface 1505 is also connected to the bus 1504.

[0230] The following components are connected to I / O interface 1505: an input section 1506 including a keyboard, mouse, etc.; an output section 1507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1508 including a hard disk, etc.; and a communication section 1509 including a network interface card such as a LAN card, modem, etc. The communication section 1509 performs communication processing via a network such as the Internet. A drive 1510 is also connected to I / O interface 1505 as needed. Removable media 1511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1510 as needed so that computer programs read from them can be installed into storage section 1508 as needed.

[0231] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present 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 1509, and / or installed from removable medium 1511. When the computer program is executed by central processing unit (CPU) 1501, it performs the functions defined above in the system of the present invention.

[0232] 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.

[0233] 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.

[0234] The units described in the embodiments of the present invention can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor can be described as including a sending unit, an acquisition unit, a determining unit, and a first processing unit. The names of these units do not necessarily limit the specific unit; for example, the sending unit can also be described as "a unit that sends an image acquisition request to a connected server."

[0235] 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 that, when executed by the device, cause the device to: obtain a user question from a client; classify the user question using a natural language understanding model to obtain a target intent category of the user question; obtain a target composite business answer corresponding to the target intent category, the target composite business answer including general answers for different subcategories of questions within the target intent category; match a response paragraph corresponding to the user question from the target composite business answer; differentiate the display of the response paragraph in the target composite business answer; and return the target composite business answer to the client to display the target composite business answer, including the differentiated response paragraph, on the client.

[0236] Exemplary embodiments of the present invention have been specifically shown and described above. It should be understood that the present invention is not limited to the detailed structures, arrangements, or implementation methods described herein; rather, the present invention is intended to cover various modifications and equivalent arrangements contained within the spirit and scope of the appended claims.

Claims

1. A smart response method, characterized in that, include: Get user questions from the client; The user's question is classified using a natural language understanding model to obtain the target intent category of the user's question; Obtain the target composite business answer corresponding to the target intent category, wherein the target composite business answer includes general answers to different sub-categories of questions within the target intent category; Match the response paragraphs corresponding to the user's question from the target composite business answer; The response paragraphs in the target composite business answer are displayed differently; The target composite business answer is returned to the client to display the target composite business answer, including the differentiated response paragraph, on the client. The process of matching the response paragraph corresponding to the user question from the target composite business answer includes: The user question is segmented into words to obtain the segmentation results of the user question. Based on the punctuation in the target composite business answer, the target composite business answer is segmented to obtain the unit paragraphs of the target composite business answer; Based on the word segmentation results and the unit paragraphs, determine the key business sentences; Starting with the key business sentence, regular expressions are used to match and obtain the response paragraph corresponding to the user question from the target composite business answer. The different subclasses of the problem include a first subclass problem and a second subclass problem; wherein, the method further includes: The frequency with which the response paragraphs are displayed differently in response to the user's question is counted to obtain the statistical count of the responses for each response paragraph. Based on the number of responses, a first response segment to be split and a second response segment to be split are determined from the target composite business answers. The first response segment to be split corresponds to the first sub-category question, and the second response segment to be split corresponds to the second sub-category question. Determine the weighted difference between the first response segment to be split and the second response segment to be split; If the weighted difference exceeds the preset difference, then the first response paragraph to be split and the second response paragraph to be split are separated from the target composite business answer.

2. The method according to claim 1, characterized in that, Based on the word segmentation results and the unit paragraphs, key business sentences are determined, including: Match the target business words in the word segmentation results of the user's question; If the word segmentation result of the user question includes the target business word, then the unit paragraph containing the most different target business words is determined as the business key sentence; If the word segmentation result of the user question does not contain the target business term, then the similarity of each unit paragraph of the user question and the target composite business answer is determined. The paragraphs with a similarity greater than the similarity threshold are identified as the key business sentences.

3. The method according to claim 1, characterized in that, Differentiated display of response paragraphs in the target composite service answer, including: If the length of the response paragraph is less than the length of the target composite business answer, then the response paragraph in the target composite business answer is displayed differently using the first method. The key business sentences in the response paragraph are displayed differently using a second method.

4. The method according to claim 1, characterized in that, Also includes: If no matching response paragraph for the user question is obtained from the target composite business answer, the frequency of the number of times no matching response paragraph for the user question is obtained from the target composite business answer is counted, and the number of trigger statistics for the business key sentence is obtained. Retrieve key business sentences whose number of triggered statistics exceeds the first statistical count; Cluster the user questions corresponding to the business key sentences whose number of trigger statistics exceeds the first statistical number to obtain the first clustering result; Based on the first clustering result, add new intent categories and their corresponding business answers.

5. The method according to claim 2, characterized in that, Also includes: If the word segmentation result of the user question does not contain the target business term, then the user question will be added to the list of uncovered questions. Count the frequency of user questions retrieved from the uncovered question list to obtain the total number of user questions. Retrieve user issues whose total number of issues exceeds the second set of statistical counts; The user questions whose statistical number exceeds the second statistical number are clustered to obtain the second clustering result; Based on the second clustering result, add new intent categories and their corresponding business answers.

6. A smart response method, characterized in that, include: The system receives user questions, obtains corresponding target composite business answers based on the user questions, matches response paragraphs corresponding to the user questions from the target composite business answers, and displays the response paragraphs in the target composite business answers differently. Matching response paragraphs corresponding to the user questions from the target composite business answers includes: segmenting the user questions into words to obtain segmentation results; segmenting the target composite business answers based on punctuation to obtain unit paragraphs of the target composite business answers; determining key business sentences based on the segmentation results and the unit paragraphs; and using the key business sentences as starting points, matching response paragraphs corresponding to the user questions from the target composite business answers using regular expressions. Display the target complex business answer, including differentiated response paragraphs; The different subclasses of the problem include a first subclass problem and a second subclass problem; the method further includes: The frequency with which the response paragraphs are displayed differently in response to the user's question is counted to obtain the statistical count of the responses for each response paragraph. Based on the number of responses, a first response segment to be split and a second response segment to be split are determined from the target composite business answers. The first response segment to be split corresponds to the first sub-category question, and the second response segment to be split corresponds to the second sub-category question. Determine the weighted difference between the first response segment to be split and the second response segment to be split; If the weighted difference exceeds the preset difference, then the first response paragraph to be split and the second response paragraph to be split are separated from the target composite business answer.

7. A smart response device, characterized in that, include: The user issue retrieval module is used to retrieve user issues from the client. The intent category acquisition module is used to classify the user question using a natural language understanding model to obtain the target intent category of the user question; The business answer acquisition module is used to obtain the target composite business answer corresponding to the target intent category, wherein the target composite business answer includes the general answer to different sub-category questions in the target intent category; The response paragraph matching module is used to match the response paragraph corresponding to the user question from the target composite business answer; wherein, matching the response paragraph corresponding to the user question from the target composite business answer includes: segmenting the user question into words to obtain the word segmentation result of the user question; segmenting the target composite business answer according to the punctuation in the target composite business answer to obtain the unit paragraph of the target composite business answer; determining the business key sentence according to the word segmentation result and the unit paragraph; and using the business key sentence as the starting point, matching the response paragraph corresponding to the user question from the target composite business answer using regular expressions; The differentiation display module is used to differentiate and display the response paragraphs in the target composite business answer. A business answer return module is used to return the target composite business answer to the client, so as to display the target composite business answer, including the differentiated response paragraph, on the client; The different sub-type questions include a first sub-type question and a second sub-type question; the intelligent response device further includes: a response segment splitting module, used to count the frequency of the response segment responding to the user question and displaying the difference, and to obtain the response statistics of the response segment; based on the response statistics, to determine a first response segment to be split and a second response segment to be split from the target composite business answer, wherein the first response segment to be split corresponds to the first sub-type question and the second response segment to be split corresponds to the second sub-type question; to determine the weighted difference between the first response segment to be split and the second response segment to be split; if the weighted difference exceeds a preset difference, then the first response segment to be split and the second response segment to be split are split from the target composite business answer.

8. A smart response device, characterized in that, include: A user question receiving module is used to receive user questions, obtain corresponding target composite business answers based on the user questions, match response paragraphs corresponding to the user questions from the target composite business answers, and distinguish and display the response paragraphs in the target composite business answers. The matching of response paragraphs corresponding to the user questions from the target composite business answers includes: segmenting the user questions into words to obtain the word segmentation results; segmenting the target composite business answers according to punctuation marks to obtain unit paragraphs of the target composite business answers; determining key business sentences based on the word segmentation results and the unit paragraphs; and using the key business sentences as starting points, matching response paragraphs from the target composite business answers using regular expressions. The system obtains the response paragraphs corresponding to the user's question; wherein, different sub-type questions include a first sub-type question and a second sub-type question; the system also counts the frequency of the response paragraphs' differentiating displays in response to the user's question, and obtains the response statistics of the response paragraphs; based on the response statistics, a first response paragraph to be split and a second response paragraph to be split are determined from the target composite business answer, wherein the first response paragraph to be split corresponds to the first sub-type question, and the second response paragraph to be split corresponds to the second sub-type question; the system determines the weighted difference between the first response paragraph to be split and the second response paragraph to be split; if the weighted difference exceeds a preset difference, the first response paragraph to be split and the second response paragraph to be split are split from the target composite business answer; The business answer display module is used to display target composite business answers, including differentiated response paragraphs.

9. An electronic device, characterized in that, include: At least one processor; A storage device for storing at least one program, which, when executed by the at least one processor, causes the at least one processor to implement the smart response method as described in any one of claims 1 to 5 or the smart response method as described in claim 6.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that, When the executable instructions are executed by the processor, they implement the smart response method as described in any one of claims 1 to 5 or the smart response method as described in claim 6.

Images