Extraction method, electronic device and storage medium for name recognition

By combining voice interaction data, historical information and high-frequency statistical analysis, the person name recognition model is used to obtain expert weight scores, which solves the problem of insufficient accuracy and usability of person name recognition in the existing technology, and achieves high accuracy and efficient expert name recognition , simplifies operational steps and improves resource accessibility.

CN114297346BActive Publication Date: 2024-11-29AISPEECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202111629173.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-28
Publication Date
2024-11-29
Estimated Expiration
2041-12-28

AI Technical Summary

Technical Problem

The existing method of identifying person names is insufficient in terms of accuracy and usability of recognition, especially when different homophones or rare characters cannot be recognized, and a large amount of manpower and material resources are required to establish a rule database. The historical limitations of intelligent voice language technology lead to technical thresholds and Higher risk.

Method used

By obtaining interactive voice data in voice interaction, the person name recognition model is used to combine historical medical information, appointment registration scheduling information, high-frequency expert statistical analysis information and high-frequency search term information to obtain expert weight score information, and obtain the current from the expert name library Make an appointment with expert names, using natural language processing and context understanding functions to improve recognition accuracy and efficiency.

Benefits of technology

It improves the accuracy and availability of name recognition, reduces the repetitive operation steps of the agent, improves the accessibility and sense of gain of medical and health resources, and lowers the technical threshold.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114297346B_ABST
    Figure CN114297346B_ABST
Patent Text Reader

Abstract

The present invention discloses an extraction method, electronic device and storage medium for name recognition, wherein the method comprises: obtaining interactive voice data in voice interaction. The interactive voice data includes inquiry voice data. The inquiry voice data is recognized by a name recognition model to obtain name recognition information. The name recognition information is combined with various information to obtain expert weight score information. The name of the currently scheduled expert is obtained from the expert name library according to the expert weight score information. By using a voice recognition extraction model and combining hot word functions, high-frequency expert statistical analysis functions, historical medical information functions, real-time scheduling information for appointment registration, context understanding and other functions, not only can high accuracy and high availability of expert name recognition extraction be achieved, but also steps of seat appointment registration business operations can be saved, freeing the seat personnel from repeated input work. It also improves the accessibility and sense of gain of citizens' medical and health resources.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and in particular relates to an extraction method, electronic equipment and storage medium for name recognition. Background Art

[0002] Current methods for name recognition include rule-based name recognition and extraction, statistics-based name recognition and extraction, and a combination of rule-based and statistics-based methods. The rule-based name recognition and extraction method involves a two-stage process based on learning and application: determining rules and applying them to obtain the information the user requires. The rules for extracting information are based on extensive learning experience and a specific domain. Based on this empirical learning, the language structure and common grammatical structures of the Chinese language are determined, such as "noun" + "verb" and "subject" + "predicate." Once these rules are established, they are stored in a database for matching.

[0003] After the sentence is segmented, the structure of the sentence is matched with the rules in the information database to determine the sentence structure. After syntactic analysis, dialogue analysis, and knowledge acquisition, the useful information in the sentence is extracted, and the extraction task is completed. Therefore, for rule-based information extraction, the correctness of the rules themselves is the key to successful extraction.

[0004] Statistical name recognition and extraction methods are currently a commonly used Chinese information extraction technology, and are also relatively accurate and flexible. Common statistical models include the Markov model, the Hidden Markov model, and the Maximum Entropy model. The Hidden Markov model has developed a relatively robust algorithm with a strong theoretical foundation, making it well-suited for natural language processing.

[0005] A name recognition and extraction method that combines rules and statistics is an ideal information extraction technology. While both rule-based and statistical methods have their own advantages, they also inevitably have disadvantages. Therefore, experts have proposed a method that combines the two. This hybrid approach can reduce the complexity of the recognition problem through heuristic rules, while also increasing the system's adaptability through statistical models and mitigating the negative impact of database incompleteness.

[0006] The aforementioned rule-based name recognition and extraction method requires manual creation of a rule base beforehand, and performs recognition based on the character characteristics of names. Its drawbacks are that the patterns of expert names are complex, making rule development difficult and requiring significant manpower and resources. Furthermore, the rules developed are often too rigid, with limited coverage and the possibility of rule conflicts.

[0007] The statistical name recognition and extraction method first establishes a statistical model for Chinese expert name recognition, and uses the existing training corpus to estimate the parameters in the statistical model. Then, the parameter values ​​are assigned to the statistical model to perform Chinese expert name recognition. Its disadvantage is that one of the errors that are easy to occur in the statistical expert name recognition model is the expert name delimitation error, especially the inability to determine whether the last character is part of the real expert name or a separate word.

[0008] In actual application, it is necessary not only to identify the extracted expert's name accurately but also to make it usable. For example, when making an appointment, I want to make an appointment with a certain expert in the dermatology department of a certain hospital. If the same hospital has multiple experts with the same pronunciation but different characters, it is possible that an expert has already left. At this time, if the name extracted is the name of an expert who has left, it is an invalid expert name.

[0009] The inventors discovered that names with the same pronunciation but different characters, names of people and places with the same pronunciation but different characters, or uncommon characters cannot be identified and extracted. The accuracy and usability of the identified and extracted names are low. Existing technology centers often strengthen vertical acoustic and language models and develop more rules when recognizing names. Therefore, when an expert's name changes, the language model must be retrained or rules must be added or deleted. Due to the historical limitations of intelligent speech and language technology, the direct use of raw speech data still has high technical barriers and risks. A large amount of raw data is stored in text form. Summary of the Invention

[0010] The embodiments of the present invention aim to solve at least one of the above technical problems.

[0011] In a first aspect, an embodiment of the present invention provides a method for extracting name recognition, comprising: obtaining interactive voice data from a voice interaction. The interactive voice data includes inquiry voice data. Recognizing the inquiry voice data using a name recognition model to obtain name recognition information. Combining the name recognition information with one or more of historical medical information, appointment registration and scheduling information, high-frequency expert statistical analysis information, and high-frequency search term information to obtain expert weight score information. Retrieving the name of the currently scheduled expert from an expert name database based on the expert weight score information.

[0012] In a second aspect, an embodiment of the present invention provides an electronic device comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any one of the above-mentioned extraction methods for name recognition of the present invention.

[0013] In a third aspect, an embodiment of the present invention provides a storage medium, in which one or more programs including execution instructions are stored. The execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute any of the above-mentioned extraction methods for name recognition of the present invention.

[0014] In a fourth aspect, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes any one of the above-mentioned extraction methods for name recognition.

[0015] In this embodiment of the present invention, a name recognition model is used to identify inquiry voice data to obtain name recognition information. This name recognition information is then combined with one or more of historical medical records, appointment registration and scheduling information, statistical analysis of high-frequency experts, and high-frequency search terms to obtain expert weight score information. Based on this expert weight score information, the name of the currently scheduled expert is retrieved from the expert name database. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A flowchart of an embodiment of a method for extracting name recognition according to the present invention;

[0018] Figure 2 A flowchart of another embodiment of the method for extracting name recognition of the present invention;

[0019] Figure 3 A flow chart of a method for extracting name recognition provided by one embodiment of the present invention;

[0020] Figure 4 FIG. 1 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0022] It should be noted that, unless there is any conflict, the embodiments and features in the embodiments of this application can be combined with each other.

[0023] The present invention may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0024] In the present invention, "module", "device", "system" and the like refer to related entities applied to a computer, such as hardware, a combination of hardware and software, software or software in execution, etc. Specifically, for example, an element can be, but is not limited to, a process running on a processor, a processor, an object, an executable element, an execution thread, a program and / or a computer. In addition, an application or script program running on a server, or a server can all be an element. One or more elements can be in an execution process and / or thread, and an element can be localized on a computer and / or distributed between two or more computers, and can be run by various computer-readable media. An element can also communicate through local and / or remote processes based on a signal having one or more data packets, for example, a signal from a data packet interacting with another element in a local system, a distributed system, and / or a signal from a network on the Internet that interacts with other systems via signals.

[0025] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include" and "comprise" include not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or device. In the absence of further limitations, the elements defined by the phrase "include..." do not exclude the presence of other identical elements in the process, method, article or device that includes the elements.

[0026] The embodiment of the present invention provides a method for extracting a name for recognition, which can be applied to an electronic device. The electronic device can be a computer, a server, or other electronic products, and the present invention is not limited thereto.

[0027] Please refer to Figure 1, which shows a method for extracting name recognition provided by an embodiment of the present invention. Figure 1 As shown, a method for extracting name recognition includes:

[0028] Step S101: Acquire interactive voice data.

[0029] In this step, the interactive voice data from the voice interaction is obtained. The interactive voice data includes inquiry voice data. For example, if a patient calls to make an appointment, the patient's phone voice is obtained, including the patient's inquiry voice content, such as: Is the specialist available?

[0030] Step S102: Acquire name recognition information.

[0031] In this step, the name recognition model recognizes the query voice data to obtain name recognition information. For example, the name recognition model recognizes the name information of multiple experts from step 101.

[0032] Step S103: Obtain expert weight score information.

[0033] In this step, the name recognition information is combined with one or more of the following information: historical medical records, appointment scheduling information, high-frequency expert statistical analysis information, and high-frequency search term information to obtain expert weight score information. For example, based on the name information of multiple experts, one or more of the following information: the names of experts in the patient's previous medical records, the scheduling times of each expert, the names of high-frequency experts (with many registered patients), and the names of experts frequently searched through the hospital registration platform are obtained, and the name information of the multiple experts obtained in step S102 is obtained.

[0034] Step S104, obtaining the name of the currently scheduled expert.

[0035] In this step, the name of the currently scheduled expert is obtained from the expert name database based on the expert weight score information.

[0036] In some optional embodiments, the name recognition model in step S102 is a recognition model based on natural language processing, thereby ensuring recognition accuracy and efficiency.

[0037] In some optional embodiments, step S103 also includes: combining name recognition information with historical medical information, appointment registration and scheduling information, high-frequency expert statistical analysis information, high-frequency search term information, and context understanding information in interactive voice data to obtain expert weight score information.

[0038] In some optional embodiments, the name recognition information includes a speech recognition confidence weight value w1.

[0039] In some optional embodiments, historical medical information includes: historical medical information expert name weight w2; appointment registration and scheduling information includes: appointment registration and scheduling information weight w4; high-frequency expert statistical analysis information includes: high-frequency expert name weight w3; high-frequency search term information includes: hot word expert name weight w5; context understanding information includes: context understanding function weight w6.

[0040] In some optional embodiments, the context understanding function weight w6 is obtained based on the context recognition model of natural language processing and the interactive voice data.

[0041] In some optional embodiments, step S103 includes: obtaining expert weight score information according to formula 1.

[0042] Expert weight score = w4*[w1*score(ASR)+w2*score(H)+w3*score(F)+w5*score(P)+w6*score(C)] Formula 1

[0043] in:

[0044] score(H) = the proportion of a particular expert appearing in a patient's history;

[0045] For example: When the number has been opened 10 times in total, the number of times the number is frequently hung up.

[0046] score(F) = the percentage of a specialist appearing in all historical records; the proportion of specialists providing treatments on the appointment registration platform.

[0047] score(P) = the popularity of a hospital expert's name in the hot word database; the proportion of searches.

[0048] score(C) = the score of an expert name in the context, for example, the patient says Doctor Zhang, Gong Chang Zhang, not Li Zao Zhang; the score in the communication.

[0049] The percentage of a particular specialist's appearances in a patient's history refers to the percentage of each specialist's appearances across all of the patient's appointments. For example, if a patient has 20 appointments with specialists, and the current specialist is Zhang**, who has been seen four times, then Zhang**'s appearance percentage is 20%. And so on.

[0050] The proportion of a specialist appearing in all historical records refers to the proportion of each specialist appearing in the registration records of all patients.

[0051] The popularity of a hospital expert's name in the hot word library refers to the popularity of whether patients have seen or not seen or consulted this expert, that is, the number of times the expert's name has been searched and inquired.

[0052] An expert's score in context - this mainly refers to the current patient's telephone voice, the current patient's context department name or hospital name during the communication process.

[0053] Reference Figure 2 In some optional embodiments, after step S104, the following steps are further included:

[0054] Step S105, making an appointment. In this step, an appointment is made in the registration system according to the current appointment expert name information.

[0055] Another embodiment of the present invention provides a method for extracting name recognition, which aims to solve two core problems: 1) the accuracy of name recognition extraction; 2) the effectiveness of name recognition extraction.

[0056] Specifically, the following six methods are used to ensure that name recognition extraction is not only accurate but also effective. Take the extraction of medical expert names from telephone appointment registration conversation data as an example.

[0057] 1. Based on the general acoustic and language recognition models, vertical field-specific language model training is carried out based on the data provided by the National Health Commission's appointment registration platform to ensure the accuracy of recognition of conversations between citizens and agents.

[0058] 2. Based on the acoustic and language models, we've added a separate hotword feature for expert names. This means that when the system recognizes a pinyin combination that matches an expert's name, it will forcibly convert it into the expert's name. If there are expert names with the same pronunciation but different characters, both will be recognized and displayed, ensuring the accuracy of expert name recognition and extraction.

[0059] 3. The system conducts self-learning and intelligent labeling of citizens' historical appointment registration information. When a citizen makes another appointment with a specialist, the system predicts and determines the name of the expert in this appointment registration information based on data analysis of the citizen's historical registration and medical information. It then combines the language recognition model and the expert name hot word function to ensure that the expert name for each appointment registration is not only accurately identified and extracted, but also available to the agent.

[0060] 4. Conduct statistical analysis on the names of experts that appear during the appointment registration or consultation process for citizens across the city. During the appointment registration process, refer to the frequency of appearance of the expert names to assist in the identification and extraction of expert names. That is, the high-frequency expert names that appear during the appointment registration and consultation process will be given priority in identification and extraction in a single appointment registration.

[0061] 5. During the appointment registration process, expert names are identified and extracted based on the context of the conversation between the citizen and the agent. For example, when hospital or department information appears in the appointment registration conversation, the expert name is matched with the hospital name and department name in the context when extracting the expert name. This ensures that the expert name identified and extracted during the appointment registration process is the expert of the hospital or department. This further ensures the high accuracy and availability of the identified and extracted expert names.

[0062] 6. In the actual appointment registration business, citizens' appointment registration resources are based on the appointment registration scheduling information. Therefore, in the actual appointment registration business, on the basis of methods 1-5, we further add the identification and extraction of expert name information based on the real-time appointment registration scheduling information to ensure that the accuracy of our identification of the extracted expert names is high, and that the extracted expert names are 100% available in this registration business.

[0063] To ensure that the accuracy of the expert name extraction for citizen appointment registration is close to 100% and that the extracted expert names are 100% valid and available, please refer to Figure 3 , which shows a method for extracting name recognition provided by another embodiment of the present invention. The above-mentioned accurate effect can be achieved.

[0064] When citizens call to make an appointment, in addition to using the voice recognition model to identify and extract expert names to support the appointment service, the expert names can also be identified and extracted more accurately and effectively based on whether it is the citizen's first call to make an appointment, the content of the conversation between the citizen and the agent, and the real-time registration schedule information for this week.

[0065] In any case, if only expert names are identified and extracted, the conversation between the citizen and the agent can be extracted through [2-Speech Recognition Model]. In this case, the accuracy and effectiveness of the expert names extracted and recognized are low and cannot meet the application requirements. When a citizen makes an appointment for the first time, if the hospital and department are not mentioned above, expert names are identified and extracted.

[0066] When a citizen makes an appointment for the first time, if the hospital name or department name is not mentioned in the text, the conversation between the citizen and the agent can be recognized through [2-Speech Recognition Model] and combined with [7-Hot Word Function] and [5-High-Frequency Expert Statistical Analysis Function] to identify and extract the expert name to ensure the accuracy of expert name recognition, but it cannot guarantee that the identified and extracted expert name is 100% usable.

[0067] If the same citizen makes multiple appointments and the hospital or department name is not mentioned in the text, the conversation between the citizen and the agent can be identified and extracted using [2-Speech Recognition Model] and combined with [7-Hot Word Function], [5-High-Frequency Expert Statistical Analysis Function], and [4-Historical Medical Information Function] to identify and extract the expert name. This can ensure a higher accuracy rate for expert name recognition, but it cannot guarantee 100% accuracy of the identified and extracted expert names.

[0068] Expert name recognition and extraction occurs when a citizen makes multiple appointments and the hospital or department name appears in the text. When a citizen makes an appointment and the hospital or department name appears in the text, the conversation between the citizen and the agent is recognized using the [2-Speech Recognition Model] and combined with the [7-Hot Word Function], [5-High-Frequency Expert Statistical Analysis Function], [4-Historical Medical Information Function], [6-Appointment Scheduling Information], and [8-Context Understanding Function] to identify and extract the expert name. This approach essentially guarantees 100% accuracy in expert name recognition and extraction, and 100% usability of the identified and extracted expert names.

[0069] The weights in the calculation method for extracting name recognition in the present invention are shown in Table 1.

[0070] Table 1

[0071]

[0072] Expert total score

[0073] =w4*[w1*score(ASR)+w2*score(H)+w3*score(F)+w5*score(P)+w6*score(C)]

[0074] By labeling the training data and performing maximum likelihood estimation, the values ​​of the above weights are calculated;

[0075] score(H) = the proportion of a particular expert appearing in a patient's history;

[0076] score(F) = the percentage of an expert appearing in all historical records;

[0077] score(P) = popularity of a hospital expert's name in the hot word database;

[0078] score(C) = the score of the expert name in the context, e.g., the patient said "Dr. Zhang," not "Li Zao Zhang";

[0079] The beneficial effects of the present invention are: by combining a speech recognition extraction model with hot word functions, high-frequency expert statistical analysis functions, historical medical information functions, real-time appointment scheduling information, and context understanding functions, it not only achieves high accuracy and high usability in expert name recognition and extraction, but also saves agents from repetitive data entry work (an average of 2.4 steps are reduced) during the appointment registration process. This also improves the accessibility and sense of achievement of medical and health resources for citizens.

[0080] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of combined actions, but those skilled in the art should be aware that the present invention is not limited by the order of the actions described, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0081] In some embodiments, an embodiment of the present invention provides a non-volatile computer-readable storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute any of the above-mentioned extraction methods for name recognition of the present invention.

[0082] In some embodiments, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes any of the above-mentioned extraction methods for name recognition.

[0083] In some embodiments, an embodiment of the present invention also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute a method for extracting name recognition.

[0084] Figure 4 This is a hardware structure diagram of an electronic device for performing a name recognition extraction method provided by another embodiment of the present application, such as Figure 4 As shown, the device includes:

[0085] One or more processors 410 and memory 420, Figure 4A processor 410 is taken as an example.

[0086] The device for performing the extraction method for name recognition may further include: an input device 430 and an output device 440 .

[0087] The processor 410, the memory 420, the input device 430 and the output device 440 may be connected via a bus or other means. Figure 4 The bus connection is taken as an example.

[0088] Memory 420, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the name recognition extraction method in the embodiments of the present application. Processor 410 executes the non-volatile software programs, instructions, and modules stored in memory 420 to execute various server functional applications and data processing, thereby implementing the name recognition extraction method in the aforementioned method embodiment.

[0089] The memory 420 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the extraction device for name recognition, etc. In addition, the memory 420 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 420 may optionally include a memory remotely located relative to the processor 410, and these remote memories may be connected to the extraction device for name recognition via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0090] The input device 430 can receive input digital or character information and generate signals related to user settings and function control of the extraction device for name recognition. The output device 440 can include a display device such as a display screen.

[0091] The one or more modules are stored in the memory 420 , and when executed by the one or more processors 410 , perform the extraction method for name recognition in any of the above method embodiments.

[0092] The above-mentioned product can execute the extraction method provided in the embodiment of this application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the extraction method provided in the embodiment of this application.

[0093] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:

[0094] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and their primary purpose is to provide voice and data communications. These terminals include smartphones, multimedia phones, feature phones, and low-end phones.

[0095] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers and have computing and processing capabilities, and generally also have mobile Internet access. These terminals include PDAs, MIDs, and UMPCs.

[0096] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players, handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0097] (4) Other onboard electronic devices with data interaction functions, such as onboard computer devices installed in vehicles.

[0098] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0099] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, or of course by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the relevant technology, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the extraction method described in each embodiment or certain parts of the embodiment.

[0100] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for extracting names for expert appointment registration, comprising: Acquiring interactive voice data in the voice interaction, wherein the interactive voice data includes inquiry voice data; Recognize the inquiry voice data using a name recognition model to obtain name recognition information; The name recognition information is combined with one or more of the historical medical information, appointment registration and scheduling information, high-frequency expert statistical analysis information, and high-frequency search word information to obtain expert weight score information. =w4*[w1*score(ASR)+w2*score(H)+w3*score(F)+w5*score(P)+w6*score(C)], w4 is 1 when the expert is scheduled during the appointment time period; otherwise it is 0. score(ASR) is the confidence of speech recognition, w1 is the confidence weight of speech recognition, score(H) is the percentage of expert names appearing in all the registration records of the current patient, w2 is the weight of expert names in historical medical information, score(F) is the percentage of expert names appearing in all patients’ registration records, w3 is the weight of high-frequency expert names, score(P) = the search ratio of the expert name in the hot word database; w5 is the weight of the hot word expert name, score(C) = the score of the expert name in context understanding; w6 is the weight of the context understanding function; The name of the currently scheduled expert is obtained from the expert name database according to the expert weight score information.

2. The extraction method according to claim 1, wherein: In the step of obtaining name recognition information by recognizing the inquiry voice data through a name recognition model, the name recognition model is a recognition model based on natural language processing.

3. The extraction method according to claim 1, wherein: The context understanding function weight w6 is obtained based on the context recognition model of natural language processing and the interactive voice data.

4. The extraction method according to claim 1, wherein: After the step of obtaining the current appointment expert name from the expert name database according to the expert weight score information, the method also includes: making an appointment in the registration system according to the current appointment expert name information.

5. An electronic device comprising: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method according to any one of claims 1 to 4.

6. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Citation Information

Patent Citations

  • Method, device and system for realizing voice registration

    CN105913846A

  • Hot word acquisition method, voice recognition method and related equipment

    CN112735428A