A method and apparatus for answering a question based on a proper noun

Abstract

The application provides a question answering method and device based on a proper noun, which can be used in the financial field or other technical fields. The method comprises the following steps: receiving an inquiry request sent by a client, wherein the inquiry request comprises a consultation question; if it is determined that the consultation question comprises a proper noun based on a proper noun list, querying corresponding related information according to the proper noun comprised in the consultation question as reply information of the consultation question; wherein the proper noun list is obtained in advance; and returning the reply information of the consultation question to the client. The device is used for executing the above method. The question answering method and device based on a proper noun provided in the embodiment of the application improve the accuracy of the reply to the consultation question.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and specifically to a question-answering method and apparatus based on proper nouns. Background Technology

[0002] With the development of natural language processing technology, intelligent question answering systems have emerged, which parse questions described in natural language and return answers that match the questions.

[0003] In existing technologies, the question is segmented into words, and then the word vectors of each segmented word are summed to obtain the sentence vector corresponding to the question. Then, based on the sentence vector, the answer corresponding to the most similar preset question in the question-answering database is selected as the answer for question matching. However, this method has a drawback: when proper nouns appear in the question, the user's input may contain incomplete proper nouns, and some specialized business terms are quite similar, resulting in low differentiation between proper nouns. This can easily lead to bias when matching preset questions, and the answer corresponding to the most similar preset question may not be the correct answer, thus reducing the accuracy of the obtained answer. Summary of the Invention

[0004] To address the problems in the prior art, embodiments of the present invention provide a question-answering method and apparatus based on proper nouns, which can at least partially solve the problems existing in the prior art.

[0005] Firstly, this invention proposes a question-answering method based on proper nouns, comprising:

[0006] Receive an inquiry request sent by a client, the inquiry request including a question;

[0007] If it is determined from the list of proper nouns that the consultation question includes proper nouns, then the relevant information corresponding to the proper nouns included in the consultation question is queried as the reply information for the consultation question; wherein, the list of proper nouns is obtained in advance;

[0008] Return the response information for the inquiry to the client.

[0009] Secondly, the present invention provides a question-answering device based on proper nouns, comprising:

[0010] A receiving module is used to receive inquiry requests sent by clients, the inquiry requests including consultation questions;

[0011] The query module is used to, after determining from the list of proper nouns that the consultation question includes proper nouns, query the relevant information corresponding to the proper nouns included in the consultation question as the reply information for the consultation question; wherein, the list of proper nouns is obtained in advance;

[0012] The return module is used to return the response information for the inquiry to the client.

[0013] Thirdly, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the question-answering method based on proper nouns described in any of the above embodiments.

[0014] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the question-answering method based on proper nouns as described in any of the above embodiments.

[0015] Fifthly, the present invention provides a computer program product comprising a computer program that, when executed by a processor, implements the question-answering method based on proper nouns as described in any of the above embodiments.

[0016] The question-answering method and apparatus based on proper nouns provided in this embodiment of the invention can receive inquiry requests sent by clients, including questions. If it is determined from a list of proper nouns that the question includes proper nouns, the method queries the relevant information corresponding to the proper nouns included in the question as the answer information for the question and returns the answer information to the client. Since it can return the relevant information of proper nouns based on the proper nouns included in the question, the accuracy of answering the question is improved. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0018] Figure 1 This is a flowchart illustrating the question-answering method based on proper nouns provided in the first embodiment of the present invention.

[0019] Figure 2 This is a flowchart illustrating the question-answering method based on proper nouns provided in the second embodiment of the present invention.

[0020] Figure 3 This is a flowchart illustrating the question-answering method based on proper nouns provided in the third embodiment of the present invention.

[0021] Figure 4This is a flowchart illustrating the question-answering method based on proper nouns provided in the fourth embodiment of the present invention.

[0022] Figure 5 This is a flowchart illustrating the question-answering method based on proper nouns provided in the fifth embodiment of the present invention.

[0023] Figure 6 This is a flowchart illustrating the question-answering method based on proper nouns provided in the sixth embodiment of the present invention.

[0024] Figure 7 This is a flowchart illustrating the question-answering method based on proper nouns provided in the seventh embodiment of the present invention.

[0025] Figure 8 This is a flowchart illustrating the question-answering method based on proper nouns provided in the eighth embodiment of the present invention.

[0026] Figure 9 This is a schematic diagram of the structure of the question-answering device based on proper nouns provided in the ninth embodiment of the present invention.

[0027] Figure 10 This is a schematic diagram of the structure of the question-answering device based on proper nouns provided in the tenth embodiment of the present invention.

[0028] Figure 11 This is a schematic diagram of the structure of the question-answering device based on proper nouns provided in the eleventh embodiment of the present invention.

[0029] Figure 12 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the twelfth embodiment of the present invention.

[0030] Figure 13 This is a schematic diagram of the structure of the question-answering device based on proper nouns provided in the thirteenth embodiment of the present invention.

[0031] Figure 14 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the fourteenth embodiment of the present invention.

[0032] Figure 15 This is a schematic diagram of the physical structure of the electronic device provided in the fifteenth embodiment of the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments and descriptions of the present invention are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other.

[0034] To facilitate understanding of the technical solution provided in this application, the relevant content of the technical solution in this application will be explained below.

[0035] In general, basic questions often cannot cover all technical terms. For example, basic questions may include "What is the minimum investment amount for short-to-medium term bond open-ended net asset value (NAV) wealth management products?" but not "What is the minimum investment amount for long term bond open-ended NAV wealth management products?". If a customer asks "What is the minimum investment amount for long term bond open-ended NAV wealth management products?", the answer that is most similar to "What is the minimum investment amount for short-to-medium term bond open-ended NAV wealth management products?" may be returned, resulting in an incorrect response.

[0036] This invention addresses the shortcomings of existing technologies in handling questions containing proper nouns, especially those with long proper nouns (more than 15 characters). It proposes a question-answering method based on proper nouns. This method identifies proper nouns from customer inquiries and then queries the relevant information corresponding to the proper nouns as the reply information. This avoids incorrect matching when matching the inquiry with the basic question based on similarity, and improves the accuracy of answering questions containing proper nouns.

[0037] The following describes the specific implementation process of the question-answering method based on proper nouns provided in this embodiment of the invention, using the server as the execution subject.

[0038] Figure 1 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the first embodiment of the present invention, as shown below. Figure 1 As shown, the question-answering method based on proper nouns provided in this embodiment of the invention includes:

[0039] S101. Receive an inquiry request sent by the client, the inquiry request including a question;

[0040] Specifically, when a customer wants to ask a question, they can send an inquiry request to the server through a client. The inquiry request includes the question being asked. The server will receive the inquiry request. The client includes, but is not limited to, devices such as desktop computers, laptops, smartphones, and tablets.

[0041] For example, if customer A wants to inquire about the minimum investment amount for a long-term bond open-ended net asset value (NAV) wealth management product, customer A can enter this question on the intelligent customer service page of the bank's smartphone app. The smartphone will then send an inquiry request carrying the aforementioned question to the mobile banking server, which will then receive the inquiry request.

[0042] S102. If it is determined from the list of proper nouns that the consultation question includes proper nouns, then the relevant information corresponding to the proper nouns included in the consultation question is queried as the reply information for the consultation question; wherein, the list of proper nouns is obtained in advance;

[0043] Specifically, after receiving the query request, the server determines whether the inquiry includes proper nouns based on the proper noun list. That is, it iterates through each proper noun in the list and checks if the inquiry includes any proper noun from that list. If the inquiry includes a word from the proper noun list, then the inquiry includes proper nouns. The server then queries the corresponding relevant information based on the proper nouns included in the inquiry and uses the retrieved relevant information as the response information for the inquiry. If the inquiry does not include any word from the proper noun list, then the inquiry does not include proper nouns. The proper noun list is pre-obtained and includes each proper noun and its alternative names. The relevant information corresponding to each proper noun in the proper noun list is preset and can be set according to actual needs; this embodiment of the invention does not impose limitations on this.

[0044] For example, a proper noun is the full name of a specific financial product, and the relevant information corresponding to the full name of the specific financial product is the full information of the financial product, such as its term, return, and minimum purchase amount.

[0045] S103. Return the response information for the inquiry to the client.

[0046] Specifically, after receiving the response information for the inquiry, the server returns the response information to the client so that the client can view it.

[0047] The question-and-answer method based on proper nouns provided in this embodiment of the invention can receive inquiry requests sent by clients, which include questions. If it is determined from the list of proper nouns that the question includes proper nouns, the method queries the relevant information corresponding to the proper nouns included in the question as the answer information for the question and returns the answer information to the client. Since it can return the relevant information of proper nouns based on the proper nouns included in the question, the accuracy of answering the question is improved.

[0048] Based on the above embodiments, further, if the consultation question is determined to include proper nouns based on the list of proper nouns, the following are included:

[0049] If it is determined that the consultation question includes the subject or derivative of a proper noun in the list of proper nouns, then the consultation question is determined to include a proper noun; wherein, the list of proper nouns includes the subject and derivative of the proper noun.

[0050] Specifically, the server iterates through the proper noun list, examining the full name and alternative names of each proper noun. It scans the consultation question to see if it includes either the full name or an alternative name of the proper noun. If the consultation question includes either the full name or an alternative name of the proper noun, then the consultation question is determined to contain a proper noun. The proper noun list includes both the full name and alternative names of the proper noun. The full name refers to the complete name of the proper noun, while an alternative name refers to a different name or alternative name. Each proper noun has only one full name but can have one or more alternative names. Alternative names can be shortened names or similar descriptions of the full name.

[0051] For example, the subject term of a certain bank wealth management product is: ICBC Wealth Management · Xintianyi Private Banking Premium Medium-Short Term Bond Daily Open Net Asset Value Type Wealth Management Product 19GS2815. The above bank wealth management product can have multiple subject terms as follows:

[0052] Alternate Name 1: ICBC Wealth Management · Xintianyi Private Banking Premium Medium-Short Term Bond Daily Open Net Asset Value Type Wealth Management Product

[0053] Second pseudonym: Xintianyi Private Banking Exclusive Short-Term Bond Daily Open Net Asset Value Wealth Management Product

[0054] Thirdly: Private Bank's Exclusive Short-Term Bond Daily Open Net Asset Value Wealth Management Product

[0055] Doppelganger 4: Short-Term Bond Daily Open Net Asset Value Type Wealth Management Product 19GS2815

[0056] Doppelganger 5: Short- and Medium-Term Bond Daily Open Net Asset Value Type Wealth Management Product

[0057] The subject and derivative terms of proper nouns can be obtained from specialized business terminologies or extracted from actual consultation questions. Specialized business terminologies include, for example, a database of financial product names or a database of standard business names.

[0058] Figure 2 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the second embodiment of the present invention, as shown below. Figure 2 As shown, based on the above embodiments, the question-answering method based on proper nouns provided in this embodiment of the invention further includes:

[0059] S201. If it is determined from the list of proper nouns that the consultation question does not include proper nouns, then the consultation question is segmented to obtain the vocabulary corresponding to the consultation question.

[0060] Specifically, the server determines whether the consultation question includes proper nouns based on a list of proper nouns. If the consultation question does not include any word in the list of proper nouns, then the consultation question does not include proper nouns. The server performs word segmentation on the consultation question to obtain the corresponding vocabulary. Word segmentation can be implemented using word segmentation tools, such as Jieba, PkuSeg, THULAC, etc., selected according to actual needs; this embodiment of the invention does not impose any limitations.

[0061] S202. Perform entity recognition on the words corresponding to the consultation question to obtain the entity corresponding to the consultation question;

[0062] Specifically, the server performs entity recognition on the words corresponding to the consultation question, identifies entities from the words corresponding to the consultation question, and obtains the entity corresponding to the consultation question. In this embodiment of the invention, an entity refers to a proper noun outside the proper noun list, such as a person's name, place name, country name, date, etc., and is set according to actual needs; this embodiment of the invention does not impose any limitations.

[0063] For example, an entity lexicon can be pre-established. When performing entity recognition on the vocabulary corresponding to the consultation question, each word in the vocabulary corresponding to the consultation question is compared with each proper noun in the entity lexicon. If a word in the vocabulary corresponding to the consultation question is the same as a proper noun in the entity lexicon, then the word that is the same as a proper noun in the entity lexicon is considered an entity. If a word in the vocabulary corresponding to the consultation question is not the same as any proper noun in the entity lexicon, then that word is not considered an entity.

[0064] S203. Perform part-of-speech filtering on the remaining words corresponding to the consultation question to obtain the nouns and verbs corresponding to the consultation question; wherein, the remaining words corresponding to the consultation question refer to the words remaining after removing the entity corresponding to the consultation question from the words corresponding to the consultation question;

[0065] Specifically, after identifying the entity corresponding to the consultation question, the server removes the entity corresponding to the consultation question from the vocabulary corresponding to the consultation question to obtain the remaining vocabulary corresponding to the consultation question. Then, the server performs part-of-speech filtering on the remaining vocabulary corresponding to the consultation question, retaining nouns and verbs from the remaining vocabulary corresponding to the consultation question and filtering out other words besides nouns and verbs, thereby obtaining the nouns and verbs corresponding to the consultation question.

[0066] For example, word segmentation tools usually have part-of-speech tagging functionality. You can first use the part-of-speech tagging function of the word segmentation tool to tag each word in the vocabulary corresponding to the consultation question, and then retain the verbs and nouns as the nouns and verbs corresponding to the consultation question.

[0067] S204. Based on the entity corresponding to the consultation question and the nouns and verbs corresponding to the consultation question, obtain the sentence vector corresponding to the consultation question;

[0068] Specifically, the server can obtain the word vectors corresponding to each entity in the entity corresponding to the consultation question, the word vectors corresponding to each noun in the noun corresponding to the consultation question, and the word vectors corresponding to each verb in the verb corresponding to the consultation question. Then, by adding the word vectors corresponding to the entities, the nouns, and the verbs, the sentence vector corresponding to the consultation question can be obtained. The conversion of words into vectors can be achieved using the Word2vec model, a word vector model used to map words to vectors.

[0069] S205. Based on the sentence vector corresponding to the consultation question and the sentence vectors corresponding to each basic question, obtain the first basic question that matches the consultation question; wherein, the sentence vectors corresponding to each basic question are obtained in advance;

[0070] Specifically, the server compares the sentence vector corresponding to the consultation question with the sentence vector corresponding to each basic question, and selects the basic question with the highest similarity to the consultation question as the first basic question matching the consultation question. The sentence vector corresponding to each basic question is obtained in advance, and the basic questions are pre-set and stored in a basic question-and-answer database.

[0071] For example, the cosine similarity between the sentence vector corresponding to the consultation question and the sentence vector corresponding to the basic question can be calculated as the similarity between the consultation question and the basic question.

[0072] S206. Obtain the answer to the first basic question that matches the consultation question from the basic question and use it as the response information for the consultation question.

[0073] Specifically, after obtaining a first basic question matching the consultation question, the server queries the basic question-and-answer database for the answer corresponding to the first basic question matching the consultation question, and uses this answer as the response information for the consultation question. The basic question-and-answer database is pre-set and includes multiple question-and-answer pairs, each including a basic question and its corresponding answer.

[0074] The question-answering method based on proper nouns provided in this invention improves the accuracy of responses by matching entities and retaining verbs and nouns in the question, while retaining keywords that express the intent of the question and eliminating some interfering words.

[0075] Figure 3 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the third embodiment of the present invention, as shown below. Figure 3 As shown, based on the above embodiments, the question-answering method based on proper nouns provided in this embodiment of the invention further includes:

[0076] S301. If it is determined that the number of verbs corresponding to the consultation question is greater than 1, then select one verb from the verbs corresponding to the consultation question as the key verb of the consultation question;

[0077] Specifically, the server counts the number of verbs corresponding to the consultation question. If the number of verbs corresponding to the consultation question is greater than 1 (i.e., greater than or equal to 2), then, due to the existence of multiple verbs, when matching with the basic question later, there may be a situation where the answer corresponding to the basic question with the highest similarity is not the expected answer. The server can then select one verb from the verbs corresponding to the consultation question as the key verb for the consultation question. The specific process of selecting verbs is detailed below and will not be repeated here.

[0078] S302. Based on the nouns corresponding to the consultation question and the key verbs of the consultation question, generate a standard question corresponding to the consultation question according to a unified sentence pattern; wherein, the unified sentence pattern is preset;

[0079] Specifically, the server generates a standard question based on the entity corresponding to the consultation question, the noun corresponding to the consultation question, and the key verbs of the consultation question, following a standardized sentence pattern. This standardized sentence pattern is preset.

[0080] For example, the unified sentence pattern is entity and / or noun + "how" + key verb. When there are multiple entities and nouns corresponding to the consultation question, the order of the entities and nouns corresponding to the consultation question in the unified sentence pattern is the same as the order of the entities and consultation questions in the consultation question.

[0081] S303. Based on the standard question corresponding to the consultation question and the standard questions corresponding to each basic question, obtain the second basic question that matches the consultation question; wherein, the standard questions corresponding to each basic question are obtained in advance;

[0082] Specifically, the server converts the standard question corresponding to the consultation question into a standard question vector, and also converts the standard question corresponding to each basic question into a standard question vector. It calculates the cosine similarity between the standard question vector corresponding to the matching question and the standard question vector corresponding to each basic question, and selects the basic question with the highest similarity to the consultation question as the second basic question for matching the consultation question. The standard questions corresponding to each basic question are obtained in advance, and the specific process for obtaining the standard questions corresponding to each basic question is similar to the process for obtaining the standard questions corresponding to the consultation question.

[0083] S304. Obtain the answer to the second basic question that matches the consultation question from the basic question and answer database, and use it as the response information for the consultation question.

[0084] Specifically, after obtaining a second basic question that matches the consultation question, the server queries the basic question-and-answer database for the answer corresponding to the second basic question that matches the consultation question, and uses it as the response information for the consultation question.

[0085] The question-answering method based on proper nouns provided in this invention selects one verb from multiple verbs, retains the key verb that can express the intent of the question for question matching, further eliminates interfering verbs, and improves the accuracy of the response information to the inquiry.

[0086] Figure 4 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the fourth embodiment of the present invention, as shown below. Figure 4 As shown, based on the above embodiments, the further step of obtaining the standard questions corresponding to each basic question in advance includes:

[0087] S401. Segment the basic questions to obtain the corresponding vocabulary.

[0088] Specifically, the server retrieves a basic question from the basic question-and-answer database, performs word segmentation on the basic question, and obtains the corresponding vocabulary.

[0089] S402. Perform entity recognition on the words corresponding to the basic questions to obtain the entities corresponding to the basic questions;

[0090] Specifically, the server performs entity recognition on the words corresponding to the basic question, identifies entities from the words corresponding to the basic question, and obtains the entities corresponding to the basic question.

[0091] S403. Perform part-of-speech filtering on the remaining vocabulary corresponding to the basic question to obtain the nouns and verbs corresponding to the basic question; whereby the remaining vocabulary corresponding to the basic question refers to the vocabulary remaining after removing the entities corresponding to the basic question from the vocabulary corresponding to the basic question.

[0092] Specifically, after identifying the entity corresponding to the basic question, the server removes the entity corresponding to the basic question from the vocabulary corresponding to the basic question to obtain the remaining vocabulary corresponding to the basic question. Then, it performs part-of-speech filtering on the remaining vocabulary corresponding to the basic question, retaining nouns and verbs in the remaining vocabulary corresponding to the basic question and filtering out other words besides nouns and verbs, thus obtaining the nouns and verbs corresponding to the basic question.

[0093] S404. If it is determined that the number of verbs corresponding to the basic question is greater than 1, then select one verb from the verbs corresponding to the basic question as the key verb of the basic question.

[0094] Specifically, the server counts the number of verbs corresponding to the basic question. If the number of verbs corresponding to the basic question is greater than 1, a verb can be selected from the verbs corresponding to the basic question as the key verb of the basic question. The specific process of selecting verbs is described below and will not be repeated here.

[0095] S405. Based on the entities corresponding to the basic questions, the nouns corresponding to the basic questions, and the key verbs of the basic questions, generate standard questions corresponding to the basic questions according to a unified sentence pattern.

[0096] Specifically, the server generates a standard question corresponding to the basic question based on the entity, noun, and key verb of the basic question, following a standardized sentence pattern. This standardized sentence pattern is preset.

[0097] By iterating through each basic question in the basic question-and-answer database and repeating steps S401, S402, S403, S404, and S405, the standard question corresponding to each basic question in the basic question-and-answer database can be obtained.

[0098] Figure 5 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the fifth embodiment of the present invention, as shown below. Figure 5 As shown, based on the above embodiments, further, selecting a verb from the verbs corresponding to the target question as the key verb of the target question includes:

[0099] S501. Obtain the average TF-IDF score of each verb in the verbs corresponding to the target question; wherein, the average TF-IDF score of each verb is obtained in advance based on the basic question-answering database; the target question is the consultation question or the basic question;

[0100] Specifically, the server can query and obtain the average TF-IDF (term frequency–inverse document frequency) score of each verb in the verbs corresponding to the target question. The average TF-IDF score of each verb is pre-obtained based on a basic question-answering database. The target question is either the consultation question or a basic question, and the process of selecting key verbs from the verbs corresponding to the consultation question is similar to the process of selecting key verbs from the verbs corresponding to the basic question.

[0101] S502. Obtain the verb with the highest average TF-IDF score as the key verb corresponding to the target question.

[0102] Specifically, the server compares the average TF-IDF scores of each verb in the verbs corresponding to the target question, and selects the verb with the highest average TF-IDF score as the key verb corresponding to the target question.

[0103] Figure 6 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the sixth embodiment of the present invention, as shown below. Figure 6 As shown, based on the above embodiments, further, obtaining the average TF-IDF score for each verb based on the basic question-answering database includes:

[0104] S601. Based on each question-answer pair in the basic question-answer database, obtain the word frequency and reverse file frequency of each verb in the verb corresponding to each question-answer;

[0105] Specifically, for each question-answer pair in the basic question-answer database, the server calculates the word frequency and reverse file frequency of each verb in the corresponding verb of the question-answer pair based on the basic question included in the question-answer pair and the answer corresponding to the basic question.

[0106] For example, a question-and-answer pair in the basic question-and-answer database includes the basic question: How do I cancel a bank card? and the corresponding answer: You can cancel your bank card at a local branch, or you can cancel it through online banking or mobile banking; if it is not used for more than 5 years, it will be automatically canceled.

[0107] In the above question-and-answer pairs, the total number of words is 25. The verb "cancel" corresponding to the basic question appears 4 times, and the word frequency (TF) of "cancel" is 4 / 25 = 0.16. If the number of question-and-answer pairs included in the basic question-and-answer database is 100,000, and the word "cancel" appears in 10 question-and-answer pairs, then the reverse file frequency (IDF) of "cancel" is lg(100,000 / 10) = 4.

[0108] S602. Based on the term frequency and reverse file frequency of each verb in each question-answer pair, obtain the TF-IDF score of each verb in each question-answer pair.

[0109] Specifically, for each verb in the question-answer pair, the server can calculate the product of the verb's word frequency and its inverse file frequency as the verb's TF-IDF score.

[0110] For example, if the server calculates that a question-answer pair includes a verb request with a TF of 0.16 and an IDF of 3, then the TF-IDF score of the corresponding request for that question-answer pair is 0.16 × 3 = 0.48.

[0111] S603. Based on the individual TF-IDF scores of each verb, obtain the average TF-IDF score of each verb.

[0112] Specifically, for a given verb, appearing in different question-answer pairs will result in multiple TF-IDF scores. The server calculates the average TF-IDF score of the verb by averaging these individual TF-IDF scores. For each verb appearing in a question-answer pair within the basic question-answer database, the average TF-IDF score for each verb can be calculated.

[0113] For example, if the verb "apply" appears in 10 question-answer pairs in the basic question-answer database, and the 10 TF-IDF scores for the verb "apply" are 0.1, 0.2, 0.3, 0.4, 0.5, 0.5, 0.4, 0.3, 0.2, and 0.1, then the average TF-IDF score for the verb "apply" is (0.1+0.2+0.3+0.4+0.5+0.5+0.4+0.3+0.2+0.1) / 10 = 0.3.

[0114] Figure 7 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the seventh embodiment of the present invention, as shown below. Figure 7 As shown, based on the above embodiments, the further step of obtaining the sentence vectors corresponding to each basic question in advance includes:

[0115] S701. Segment the basic questions to obtain the corresponding vocabulary.

[0116] Specifically, the server retrieves a basic question from the basic question-and-answer database, performs word segmentation on the basic question, and obtains the vocabulary corresponding to the basic question.

[0117] S702. Perform entity recognition on the words corresponding to the basic questions to obtain the entities corresponding to the basic questions;

[0118] Specifically, the server performs entity recognition on the words corresponding to the basic question, identifies entities from the words corresponding to the basic question, and obtains the entities corresponding to the basic question.

[0119] S703. Perform part-of-speech filtering on the remaining vocabulary corresponding to the basic question to obtain the nouns and verbs corresponding to the basic question; whereby the remaining vocabulary corresponding to the basic question refers to the vocabulary remaining after removing the entities corresponding to the basic question from the vocabulary corresponding to the basic question.

[0120] Specifically, after identifying the entity corresponding to the basic question, the server removes the entity corresponding to the basic question from the vocabulary corresponding to the basic question to obtain the remaining vocabulary corresponding to the basic question. Then, it performs part-of-speech filtering on the remaining vocabulary corresponding to the basic question, retaining nouns and verbs in the remaining vocabulary corresponding to the basic question and filtering out other words besides nouns and verbs, thus obtaining the nouns and verbs corresponding to the basic question.

[0121] S704. Based on the entities corresponding to the basic questions and the nouns and verbs corresponding to the basic questions, obtain the sentence vectors corresponding to the basic questions.

[0122] Specifically, the server can obtain the vector corresponding to each entity in the entity corresponding to the basic question, the vector corresponding to each noun in the noun corresponding to the basic question, and the vector corresponding to each verb in the verb corresponding to the basic question. Then, the vectors corresponding to the entities, the vectors corresponding to the nouns, and the vectors corresponding to the verbs are added together to obtain the sentence vector corresponding to the basic question.

[0123] By iterating through each basic question in the basic question-and-answer database and repeating steps S701, S702, S703, and S704, the standard question corresponding to each basic question in the basic question-and-answer database can be obtained.

[0124] Figure 8 This is a flowchart illustrating the question-and-answer method based on proper nouns provided in the eighth embodiment of the present invention, as shown below. Figure 8 As shown, based on the above embodiments, further, obtaining the sentence vector corresponding to the consultation question based on the entity corresponding to the consultation question and the nouns and verbs corresponding to the consultation question includes:

[0125] S801. Obtain the word vector of the entity corresponding to the consultation question, the word vector of the noun corresponding to the consultation question, and the word vector of the verb corresponding to the consultation question;

[0126] Specifically, the server converts each entity in the entity corresponding to the consultation question into a word vector corresponding to each entity in the entity corresponding to the consultation question, converts the word vector corresponding to each noun in the noun corresponding to the consultation question, and converts the word vector corresponding to each verb in the verb corresponding to the consultation question.

[0127] S802. Based on the positions of the entity corresponding to the consultation question, the noun corresponding to the consultation question, and the verb corresponding to the consultation question in the consultation question, add the word vectors of the entity corresponding to the consultation question, the noun corresponding to the consultation question, and the verb corresponding to the consultation question to obtain the sentence vector corresponding to the consultation question.

[0128] Specifically, the server, based on the positions of the entities, nouns, and verbs corresponding to the consultation question in the consultation question, adds up the word vectors corresponding to each entity, noun, and verb in the consultation question to obtain the sentence vector corresponding to the consultation question.

[0129] The following example illustrates the specific implementation process of the question-answering method based on proper nouns provided in this embodiment of the invention.

[0130] A customer wants to inquire about the minimum investment amount for a certain wealth management product offered by a bank. On the intelligent customer service page of the bank's smartphone app, they can enter, "Hello, I would like to know what the minimum investment amount is for a 3-month fixed-income open-ended wealth management product?"

[0131] The smartphone will send an inquiry request to the mobile banking server. The inquiry request includes questions such as: Hello, I would like to know what the minimum purchase amount is for a 3-month fixed-income open-ended wealth management product?

[0132] Upon receiving the aforementioned inquiry, the mobile banking server, based on a list of proper nouns, determines whether the inquiry includes the entity term or a related term of a proper noun. It identifies the inquiry's mention of "

Fixed-income 3-month fixed-term open-ended wealth management product

ICBC Wealth Management · Xinwenli Private Banking Premium Fixed-income 3-month fixed-term open-ended wealth management product 19GS2818

ICBC Wealth Management · Xinwenli Private Banking Premium Fixed-income 3-month fixed-term open-ended wealth management product 19GS2818

[0133] The mobile banking server will send the relevant information corresponding to the entity term "ICBC Wealth Management · Xinwenli Private Banking Premium Fixed Income 3-Month Fixed-Term Open-Ended Wealth Management Product 19GS2818" to the smartphone as the reply to the above inquiry. The smartphone will display the reply to the above inquiry on the intelligent customer service page for the customer to view.

[0134] The API can be used to query information related to the ontology term. The entity term can be passed as an input parameter, and the API returns the relevant information. This method is more accurate than obtaining responses through basic similarity matching questions. It is especially effective for inquiries where the ontology term or its corresponding element exceeds 15 characters.

[0135] The following example illustrates the specific implementation process of the question-answering method based on proper nouns provided in this embodiment of the invention.

[0136] A customer wants to inquire about how the subscription fee for equity funds is calculated. On the intelligent customer service page of the bank's smartphone app, they can enter, "Hello, I would like to know how the subscription fee for equity funds is calculated?"

[0137] The smartphone will send an inquiry request to the mobile banking server. The inquiry request includes questions such as: Hello, I would like to know how the handling fee for subscribing to equity funds is calculated?

[0138] The mobile banking server receives the above inquiry and determines whether it includes the subject or derivative of a proper noun based on a list of proper nouns. If the inquiry does not include any subject or derivative of a proper noun, the server performs word segmentation on the inquiry to obtain the corresponding vocabulary. Then, entity recognition is performed on the corresponding vocabulary to obtain the entity for the inquiry as: "equity fund". "Equity fund" is then removed from the vocabulary, and part-of-speech tagging is performed, retaining both nouns and verbs. The resulting noun for the inquiry is: "handling fee", and the corresponding verbs are: "want", "understand", "subscribe", and "calculate".

[0139] The server obtains the sentence vector corresponding to the above-mentioned inquiry question based on the entities, nouns, and verbs associated with it. Then, based on this sentence vector and the sentence vectors of various basic questions in the basic question-and-answer database, it obtains the matching basic question: "How are the transaction fees for subscribing to equity funds calculated?". The answer to this basic question serves as the response to the above-mentioned inquiry question.

[0140] Figure 9 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the ninth embodiment of the present invention, as shown below. Figure 9 As shown, the question-and-answer device based on proper nouns provided in this embodiment of the invention includes a receiving module 901, a query module 902, and a return module 903, wherein:

[0141] The receiving module 901 is used to receive an inquiry request sent by the client, the inquiry request including a question; the query module 902 is used to, after determining that the question includes proper nouns based on the list of proper nouns, query the relevant information corresponding to the proper nouns included in the question as the reply information for the question; wherein, the list of proper nouns is obtained in advance; the returning module 903 is used to return the reply information for the question to the client.

[0142] Specifically, when a customer wants to ask a question, they can send an inquiry request to the receiving module 901 through the client. The inquiry request includes the question being asked. The receiving module 901 will receive the inquiry request. The client includes, but is not limited to, devices such as desktop computers, laptops, smartphones, and tablets.

[0143] Upon receiving the inquiry request, the query module 902 determines whether the inquiry includes proper nouns based on the proper noun list. Specifically, it iterates through each proper noun in the list, checking if the inquiry includes any proper noun from that list. If the inquiry includes a word from the proper noun list, then the inquiry includes proper nouns. The query module 902 then queries the corresponding relevant information based on the proper nouns included in the inquiry and uses the retrieved relevant information as the response information for the inquiry. If the inquiry does not include any word from the proper noun list, then the inquiry does not include proper nouns. The proper noun list is pre-obtained and includes each proper noun and its alternative names. The relevant information corresponding to each proper noun in the proper noun list is preset and can be set according to actual needs; this embodiment of the invention does not impose limitations on this.

[0144] After obtaining the response information for the inquiry, the return module 903 returns the response information to the client so that the client can view it.

[0145] The question-and-answer device based on proper nouns provided in this embodiment of the invention can receive inquiry requests sent by clients, including questions. If it is determined from the list of proper nouns that the question includes proper nouns, the device queries the relevant information corresponding to the proper nouns in the question as the answer information for the question and returns the answer information to the client. Since it can return the relevant information of proper nouns based on the proper nouns in the question, the accuracy of answering the question is improved.

[0146] Based on the above embodiments, the query module 902 is further specifically used for

[0147] If it is determined that the consultation question includes the subject or derivative of a proper noun in the list of proper nouns, then the consultation question is determined to include a proper noun; wherein, the list of proper nouns includes the subject and derivative of the proper noun.

[0148] Figure 10 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the tenth embodiment of the present invention, as shown below. Figure 10 As shown, based on the above embodiments, the question-answering device based on proper nouns provided in this embodiment further includes a first word segmentation module 904, a first entity recognition module 905, a first part-of-speech filtering module 906, a first acquisition module 907, a second acquisition module 908, and a first acquisition module 909, wherein:

[0149] The first word segmentation module 904 is used to segment the consultation question into words if it is determined from the list of proper nouns that the consultation question does not contain proper nouns, thereby obtaining the vocabulary corresponding to the consultation question; the first entity recognition module 905 is used to perform entity recognition on the vocabulary corresponding to the consultation question, thereby obtaining the entity corresponding to the consultation question; the first part-of-speech filtering module 906 is used to perform part-of-speech filtering on the remaining vocabulary corresponding to the consultation question, thereby obtaining the nouns and verbs corresponding to the consultation question; wherein, the remaining vocabulary corresponding to the consultation question refers to the vocabulary remaining after removing the entity corresponding to the consultation question from the vocabulary corresponding to the consultation question; the first obtaining module 907 is used to obtain the sentence vector corresponding to the consultation question based on the entity corresponding to the consultation question and the nouns and verbs corresponding to the consultation question; the second obtaining module 908 is used to obtain the first basic question matching the consultation question based on the sentence vector corresponding to the consultation question and the sentence vectors corresponding to each basic question; wherein, the sentence vectors corresponding to each basic question are obtained in advance; the first acquiring module 909 obtains the answer corresponding to the first basic question matching the consultation question from the basic question-and-answer database, as the reply information for the consultation question.

[0150] Figure 11 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the eleventh embodiment of the present invention, as shown below. Figure 11 As shown, based on the above embodiments, the question-answering device based on proper nouns provided in this embodiment of the invention further includes a first filtering module 910, a first generation module 911, a third obtaining module 912, and a second acquisition module 913, wherein:

[0151] The first filtering module 910 is used to select a verb from the verbs corresponding to the consultation question as the key verb of the consultation question if it is determined that the number of verbs corresponding to the consultation question is greater than 1; the first generation module 911 is used to generate a standard question corresponding to the consultation question according to a unified sentence pattern based on the nouns corresponding to the consultation question and the key verb of the consultation question; wherein, the unified sentence pattern is preset; the third obtaining module 912 is used to obtain a second basic question matching the consultation question based on the standard question corresponding to the consultation question and the standard questions corresponding to each basic question; wherein, the standard questions corresponding to each basic question are obtained in advance; the second obtaining module 913 is used to obtain the answer corresponding to the second basic question matching the consultation question from the basic question-answer database as the reply information of the consultation question.

[0152] Figure 12 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the twelfth embodiment of the present invention, as shown below. Figure 12As shown, based on the above embodiments, the question-answering device based on proper nouns provided in this embodiment further includes a second word segmentation module 914, a second entity recognition module 915, a second part-of-speech filtering module 916, a second filtering module 917, and a second generation module 918, wherein:

[0153] The second word segmentation module 914 is used to segment the basic question to obtain the vocabulary corresponding to the basic question; the second entity recognition module 915 is used to perform entity recognition on the vocabulary corresponding to the basic question to obtain the entity corresponding to the basic question; the second part-of-speech filtering module 916 is used to perform part-of-speech filtering on the remaining vocabulary corresponding to the basic question to obtain the nouns and verbs corresponding to the basic question; wherein, the remaining vocabulary corresponding to the basic question refers to the vocabulary remaining after removing the entity corresponding to the basic question from the vocabulary corresponding to the basic question; the second filtering module 917 is used to select a verb as the key verb of the basic question if it is determined that the number of verbs corresponding to the basic question is greater than 1; the second generation module 918 is used to generate a standard question corresponding to the basic question according to a unified sentence pattern based on the entity corresponding to the basic question, the noun corresponding to the basic question, and the key verb of the basic question.

[0154] Figure 13 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the thirteenth embodiment of the present invention, as shown below. Figure 13 As shown, based on the above embodiments, the question-answering device based on proper nouns provided in this embodiment further includes a third word segmentation module 919, a third entity recognition module 920, a third part-of-speech filtering module 921, and a fourth acquisition module 922, wherein:

[0155] The third word segmentation module 919 is used to segment the basic question into words to obtain the vocabulary corresponding to the basic question; the third entity recognition module 920 is used to perform entity recognition on the vocabulary corresponding to the basic question to obtain the entities corresponding to the basic question; the third part-of-speech filtering module 921 is used to perform part-of-speech filtering on the remaining vocabulary corresponding to the basic question to obtain the nouns and verbs corresponding to the basic question; wherein, the remaining vocabulary corresponding to the basic question refers to the vocabulary remaining after removing the entities corresponding to the basic question from the vocabulary corresponding to the basic question; the fourth obtaining module 922 is used to obtain the sentence vector corresponding to the basic question based on the entities corresponding to the basic question and the nouns and verbs corresponding to the basic question.

[0156] Figure 14 This is a schematic diagram of the structure of the question-and-answer device based on proper nouns provided in the fourteenth embodiment of the present invention, as shown below. Figure 14 As shown, based on the above embodiments, the first obtaining module 907 further includes a first obtaining unit 9071 and a second obtaining unit 9072, wherein:

[0157] The first obtaining unit 9071 is used to obtain the word vector of the entity corresponding to the consultation question, the word vector of the noun corresponding to the consultation question, and the word vector of the verb corresponding to the consultation question; the second obtaining unit 9072 is used to add the word vector of the entity corresponding to the consultation question, the word vector of the noun corresponding to the consultation question, and the word vector of the verb corresponding to the consultation question according to their positions in the consultation question, to obtain the sentence vector corresponding to the consultation question.

[0158] The embodiments of the device provided in this invention can be used to execute the processing flow of the above-described method embodiments. Its functions will not be repeated here, but can be referred to the detailed description of the above-described method embodiments.

[0159] It should be noted that the question-answering method and apparatus based on proper nouns provided in the embodiments of the present invention can be used in the financial field, or in any technical field other than the financial field. The embodiments of the present invention do not limit the application field of the question-answering method and apparatus based on proper nouns.

[0160] Figure 15 This is a schematic diagram of the physical structure of the electronic device provided in the fifteenth embodiment of the present invention, as shown below. Figure 15 As shown, the electronic device may include a processor 1501, a communications interface 1502, a memory 1503, and a communication bus 1504, wherein the processor 1501, the communications interface 1502, and the memory 1503 communicate with each other via the communication bus 1504. The processor 1501 can invoke logical instructions in the memory 1503 to execute the following methods: receiving an inquiry request sent by a client, the inquiry request including a question; if it is determined based on a list of proper nouns that the question includes proper nouns, then querying the corresponding relevant information based on the proper nouns included in the question as the response information for the question; wherein the list of proper nouns is obtained in advance; and returning the response information for the question to the client.

[0161] Furthermore, the logical instructions in the aforementioned memory 1503 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0162] This embodiment discloses a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments, such as: receiving an inquiry request sent by a client, the inquiry request including a question; if it is determined based on a list of proper nouns that the question includes proper nouns, then querying the relevant information corresponding to the proper nouns included in the question as the reply information for the question; wherein, the list of proper nouns is obtained in advance; and returning the reply information for the question to the client.

[0163] This embodiment provides a computer-readable storage medium storing a computer program that causes a computer to execute the methods provided in the above-described method embodiments. For example, the method includes: receiving an inquiry request sent by a client, the inquiry request including a question; if it is determined based on a list of proper nouns that the question includes proper nouns, then querying relevant information corresponding to the proper nouns included in the question as response information for the question; wherein the list of proper nouns is obtained in advance; and returning the response information for the question to the client.

[0164] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0165] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0166] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0167] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0168] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0169] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A question-and-answer method based on proper nouns, characterized in that, include: Receive an inquiry request sent by a client, the inquiry request including a question; If it is determined from the list of proper nouns that the consultation question includes proper nouns, then the relevant information corresponding to the proper nouns included in the consultation question is queried as the reply information for the consultation question; wherein, the list of proper nouns is obtained in advance; Return the response information for the inquiry to the client; The method further includes: If it is determined from the list of proper nouns that the consultation question does not contain proper nouns, then the consultation question is segmented to obtain the vocabulary corresponding to the consultation question; Entity recognition is performed on the words corresponding to the consultation question to obtain the entity corresponding to the consultation question, wherein the entity is a proper noun outside the proper noun list; The remaining vocabulary corresponding to the consultation question is filtered by part of speech to obtain the nouns and verbs corresponding to the consultation question; wherein, the remaining vocabulary corresponding to the consultation question refers to the vocabulary remaining after removing the entity corresponding to the consultation question from the vocabulary corresponding to the consultation question; Based on the entity corresponding to the consultation question, as well as the nouns and verbs corresponding to the consultation question, obtain the sentence vector corresponding to the consultation question; Based on the sentence vector corresponding to the consultation question and the sentence vectors corresponding to each basic question, a first basic question matching the consultation question is obtained; wherein, the sentence vectors corresponding to each basic question are obtained in advance; The answer to the first basic question that matches the consultation question is obtained from the basic question and answer database and used as the response information for the consultation question.

2. The method according to claim 1, characterized in that, If, based on the list of proper nouns, it is determined that the consultation question includes proper nouns, then: If it is determined that the consultation question includes the subject or derivative of a proper noun in the list of proper nouns, then the consultation question is determined to include a proper noun; wherein, the list of proper nouns includes the subject and derivative of the proper noun.

3. The method according to claim 1, characterized in that, Also includes: If it is determined that the number of verbs corresponding to the consultation question is greater than 1, then one verb is selected from the verbs corresponding to the consultation question as the key verb of the consultation question; Based on the nouns corresponding to the consultation question and the key verbs of the consultation question, a standard question corresponding to the consultation question is generated according to a unified sentence pattern; wherein, the unified sentence pattern is preset; Based on the standard questions corresponding to the consultation question and the standard questions corresponding to each basic question, a second basic question matching the consultation question is obtained; wherein, the standard questions corresponding to each basic question are obtained in advance; The answer to the second basic question that matches the consultation question is obtained from the basic question and answer database and used as the response information for the consultation question.

4. The method according to claim 3, characterized in that, The steps to obtain the standard questions corresponding to each basic question in advance include: The basic questions are segmented to obtain the corresponding vocabulary. Entity recognition is performed on the words corresponding to the basic questions to obtain the entities corresponding to the basic questions; The remaining vocabulary corresponding to the basic question is filtered by part of speech to obtain the nouns and verbs corresponding to the basic question; where the remaining vocabulary corresponding to the basic question refers to the vocabulary remaining after removing the entities corresponding to the basic question from the vocabulary corresponding to the basic question. If it is determined that the number of verbs corresponding to the basic question is greater than 1, then a verb is selected from the verbs corresponding to the basic question as the key verb of the basic question; Based on the entities, nouns, and key verbs of the basic questions, standard questions are generated according to a unified sentence pattern.

5. The method according to claim 1, characterized in that, The steps to obtain the sentence vectors corresponding to each basic question in advance include: The basic questions are segmented to obtain the corresponding vocabulary. Entity recognition is performed on the words corresponding to the basic questions to obtain the entities corresponding to the basic questions; The remaining vocabulary corresponding to the basic question is filtered by part of speech to obtain the nouns and verbs corresponding to the basic question; where the remaining vocabulary corresponding to the basic question refers to the vocabulary remaining after removing the entities corresponding to the basic question from the vocabulary corresponding to the basic question. Based on the entities corresponding to the basic questions, as well as the nouns and verbs corresponding to the basic questions, obtain the sentence vectors corresponding to the basic questions.

6. The method according to claim 1, characterized in that, The step of obtaining the sentence vector corresponding to the consultation question based on the entity corresponding to the consultation question and the nouns and verbs corresponding to the consultation question includes: Obtain the word vectors of the entities corresponding to the consultation question, the word vectors of the nouns corresponding to the consultation question, and the word vectors of the verbs corresponding to the consultation question; Based on the positions of the entity, noun, and verb corresponding to the consultation question in the consultation question, the word vectors of the entity, noun, and verb corresponding to the consultation question are added together to obtain the sentence vector corresponding to the consultation question.

7. A question-and-answer device based on proper nouns, characterized in that, include: A receiving module is used to receive inquiry requests sent by clients, the inquiry requests including consultation questions; The query module is used to, after determining from the list of proper nouns that the consultation question includes proper nouns, query the relevant information corresponding to the proper nouns included in the consultation question as the reply information for the consultation question; wherein, the list of proper nouns is obtained in advance; The return module is used to return the response information for the inquiry to the client. The question-and-answer device based on proper nouns further includes a first word segmentation module, a first entity recognition module, a first part-of-speech filtering module, a first acquisition module, a second acquisition module, and a first obtaining module, wherein: The first word segmentation module is used to segment the consultation question to obtain the vocabulary corresponding to the consultation question if it is determined from the list of proper nouns. The first entity recognition module is used to perform entity recognition on the words corresponding to the consultation question to obtain the entity corresponding to the consultation question, wherein the entity is a proper noun outside the proper noun list; The first part-of-speech filtering module is used to perform part-of-speech filtering on the remaining words corresponding to the consultation question to obtain the nouns and verbs corresponding to the consultation question; wherein, the remaining words corresponding to the consultation question refer to the words remaining after removing the entity corresponding to the consultation question from the words corresponding to the consultation question; The first obtaining module is used to obtain the sentence vector corresponding to the consultation question based on the entity corresponding to the consultation question and the nouns and verbs corresponding to the consultation question; The second obtaining module is used to obtain a first basic question matching the consultation question based on the sentence vector corresponding to the consultation question and the sentence vectors corresponding to each basic question; wherein, the sentence vectors corresponding to each basic question are obtained in advance; The first acquisition module retrieves the answer to the first basic question that matches the consultation question from the basic question and answers database, and uses it as the response information for the consultation question.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images