Information completion method, device and equipment suitable for inquiry and medium

Abstract

The application relates to the technical field of artificial intelligence and intelligent medical treatment, and discloses an information completion method and device suitable for diagnosis, equipment and a medium, wherein the method comprises the following steps: acquiring an information completion signal based on a preconfigured information completion mode; searching for diagnosis dialogues from a diagnosis dialogue library corresponding to the information completion mode according to to-be-completed doctor information corresponding to the information completion signal, so as to obtain a first diagnosis dialogue set; performing a bad recall result deletion process on the first diagnosis dialogue set according to the to-be-completed doctor information, so as to obtain a second diagnosis dialogue set; performing sorting on the second diagnosis dialogue set, so as to obtain a third diagnosis dialogue set; and generating an information completion result corresponding to the to-be-completed doctor information according to the diagnosis dialogue with the highest ranking in the third diagnosis dialogue set. Therefore, the personalized demand of patients is met, and the diagnosis efficiency is effectively improved.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and smart healthcare, and in particular to a method, apparatus, equipment and medium for information completion in medical consultation. Background Technology

[0002] In recent years, online consultations have proliferated, and public acceptance has gradually increased, leading to a surge in online consultation traffic and posing significant challenges to doctors. Currently, most doctors configure their responses with everyday language and medical terminology for quick replies, triggered by keywords or sent with a single click. However, doctors in a particular department often handle a wide variety of diseases, frequently mixing up keywords, and browsing through them individually is time-consuming, making it difficult to effectively improve consultation efficiency. Furthermore, patients' growing demand for personalized services means that doctors consistently providing the same messages can lead patients to believe that online consultations are automated, reducing their trust in the service. Summary of the Invention

[0003] Based on this, it is necessary to address the technical problem that existing online consultation technologies, which rely on keyword triggering or one-click sending of pre-configured quick information, are unable to effectively improve consultation efficiency and meet patients' personalized needs. Therefore, a method, device, equipment, and medium suitable for information completion in online consultations are proposed.

[0004] Firstly, a method for information completion in medical consultations is provided, the method comprising:

[0005] Based on the pre-configured information completion mode, obtain the information completion signal;

[0006] Based on the doctor information to be completed corresponding to the information completion signal, search for consultation scripts from the consultation script library corresponding to the information completion mode to obtain the first consultation script set;

[0007] Based on the doctor information to be supplemented, the first consultation script set is processed by deleting bad recall results to obtain the second consultation script set.

[0008] The second set of consultation scripts is sorted to obtain the third set of consultation scripts.

[0009] Based on the highest-ranked consultation script in the third consultation script set, generate the information completion result corresponding to the doctor information to be completed.

[0010] Secondly, an information completion device suitable for medical consultation is provided, the device comprising:

[0011] The signal acquisition module is used to acquire information completion signals based on a pre-configured information completion mode;

[0012] The search module is used to search for consultation scripts from the consultation script library corresponding to the information completion mode based on the doctor information to be completed corresponding to the information completion signal, and obtain the first consultation script set.

[0013] The deletion processing module is used to perform bad recall result deletion processing on the first consultation script set according to the doctor information to be supplemented, so as to obtain the second consultation script set.

[0014] The sorting module is used to sort the second set of consultation scripts to obtain the third set of consultation scripts;

[0015] The information completion result determination module is used to generate the information completion result corresponding to the doctor information to be completed based on the highest-ranked consultation script in the third consultation script set.

[0016] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described information completion method applicable to medical consultation.

[0017] Fourthly, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described information completion method applicable to medical consultation.

[0018] The information completion method applicable to medical consultation in this application obtains an information completion signal based on a pre-configured information completion mode; according to the doctor information to be completed corresponding to the information completion signal, it searches for medical consultation scripts from the medical consultation script library corresponding to the information completion mode to obtain a first set of medical consultation scripts; according to the doctor information to be completed, it performs bad recall result deletion processing on the first set of medical consultation scripts to obtain a second set of medical consultation scripts; it sorts the second set of medical consultation scripts to obtain a third set of medical consultation scripts; and according to the medical consultation script with the highest ranking in the third set of medical consultation scripts, it generates an information completion result corresponding to the doctor information to be completed. Compared to sending pre-configured quick messages via keyword triggering or a single mouse click, this application utilizes a consultation script library corresponding to the information completion mode, catering to doctors' communication habits and meeting patients' personalized needs. By deleting bad recall results from the first consultation script set, it avoids bad recall results from entering the sorting stage, improving the accuracy of the determined information completion results. By responding to information completion signals to generate information completion results in real time, doctors only need to input partial information for automatic information completion, reducing the amount of text input by doctors and effectively improving consultation efficiency. Attached Figure Description

[0019] 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.

[0020] in:

[0021] Figure 1 This is an application environment diagram for an information completion method applicable to medical consultation in one embodiment;

[0022] Figure 2 This is a flowchart of an information completion method applicable to a medical consultation in one embodiment;

[0023] Figure 3 This is a structural block diagram of an information completion device suitable for medical consultation in one embodiment;

[0024] Figure 4 This is a structural block diagram of a computer device in one embodiment. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] The information completion method for medical history taking provided in this invention can be applied to, for example... Figure 1 In this application environment, client 110 communicates with server 120 via a network. Server 120 can obtain information completion signals from client 110 based on a pre-configured information completion mode. Server 120 searches for consultation scripts in the consultation script library corresponding to the information completion mode based on the doctor information to be completed corresponding to the information completion signal, obtaining a first consultation script set. Based on the doctor information to be completed, server 120 performs bad recall result deletion processing on the first consultation script set, obtaining a second consultation script set. Server 120 sorts the second consultation script set, obtaining a third consultation script set. Based on the consultation script with the highest ranking in the third consultation script set, server 120 generates the information completion result corresponding to the doctor information to be completed. This satisfies patients' personalized needs and effectively improves consultation efficiency.

[0027] The client 110 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server 120 can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention will now be described in detail through specific embodiments.

[0028] Please see Figure 2 As shown, Figure 2 A flowchart illustrating an information completion method for medical consultation provided in an embodiment of the present invention includes the following steps:

[0029] S1: Obtain information completion signals based on pre-configured information completion modes;

[0030] Pre-configured information completion mode, with a pre-configured consultation script library.

[0031] The consultation script library contains at least consultation scripts, which are the texts in which doctors respond to patients.

[0032] The information completion signal is a signal used to complete the information entered by the doctor.

[0033] Specifically, in the pre-configured information completion mode, when a doctor replies to a patient, the information completion signal is automatically triggered when the number of characters in the text corresponding to the information entered by the doctor in the input box exceeds the preset trigger character limit, or the information completion signal is triggered when the doctor enters information in the input box and actively clicks the trigger button.

[0034] S2: Based on the doctor information to be completed corresponding to the information completion signal, search for consultation scripts from the consultation script library corresponding to the information completion mode to obtain the first consultation script set;

[0035] Doctors can enter information in the input box; they can enter text directly or voice.

[0036] The "Doctor Information to be Completed" field contains the text corresponding to the information entered by the doctor in the input box.

[0037] Specifically, based on the doctor information to be completed corresponding to the information completion signal, the most similar consultation scripts are searched from the consultation script library corresponding to the information completion mode, and all the searched consultation scripts are combined into a first consultation script set.

[0038] S3: Based on the doctor information to be supplemented, perform bad recall result deletion processing on the first consultation script set to obtain the second consultation script set;

[0039] Bad recall results are diagnostic scripts that yield significant discrepancies and / or offer little benefit in terms of information completion.

[0040] The consultation scripts in the first consultation script set may contain bad recall results. In order to prevent bad recall results from entering the next stage (i.e., sorting), it is necessary to delete bad recall results from the first consultation script set according to the doctor information to be supplemented, and then use the first consultation script set after deletion as the second consultation script set.

[0041] S4: Sort the second set of consultation scripts to obtain the third set of consultation scripts;

[0042] Specifically, based on a preset sorting rule, the consultation scripts in the second consultation script set are sorted, and the sorted second consultation script set is used as the third consultation script set.

[0043] Optionally, the default sorting rule is the number of characters.

[0044] Optionally, the default sorting rule is the latest usage time.

[0045] S5: Generate the information completion result corresponding to the doctor information to be completed based on the highest-ranked consultation script in the third consultation script set.

[0046] Specifically, the consultation script with the highest ranking in the third consultation script set is taken as the hit script; the doctor information to be supplemented is supplemented according to the hit script to obtain the information supplementation result.

[0047] Optionally, the highest-ranked consultation script in the third consultation script set is selected as the hit script; the hit script is used as the information completion result.

[0048] In reverse sorting, the highest-ranking element is the first element; in ascending sorting, the highest-ranking element is the last element.

[0049] Compared to sending pre-configured quick information via keyword triggering or a single mouse click, this embodiment utilizes a consultation script library corresponding to the information completion mode, catering to doctors' communication habits and meeting patients' personalized needs. By deleting bad recall results from the first consultation script set, it avoids bad recall results from entering the sorting stage, improving the accuracy of the determined information completion results. By responding to information completion signals to generate information completion results in real time, doctors only need to input partial information for automatic information completion, reducing the amount of text input by doctors and effectively improving consultation efficiency.

[0050] In one embodiment, before the step of acquiring the information completion signal based on a pre-configured information completion mode, the method further includes:

[0051] S11: Get the mode configuration request;

[0052] A mode configuration request is a request to pre-configure the information completion mode.

[0053] Specifically, it can obtain pattern configuration requests input by the user or pattern configuration requests sent by third-party applications.

[0054] S12: Configure the information completion mode according to the mode configuration data carried in the mode configuration request;

[0055] The mode configuration data includes: script library range configuration, and the value range of the script library range configuration includes: self, same institution and same department, and same department.

[0056] The scope of the dialogue script library is configured to be the doctor himself / herself, that is, the personal dialogue script library corresponding to the doctor whose information needs to be supplemented. This personal dialogue script library is the consultation dialogue script library corresponding to the doctor himself / herself. The consultation dialogue script library corresponding to the doctor himself / herself is generated based on all messages sent by the doctor in the historical consultations within the preset window.

[0057] The script database is configured to be within the same institution and department. This means it uses the script database for the department of the doctor whose information needs to be supplemented, within the same institution and department. This departmental script database is a consultation script database for the same institution and department. The consultation script database for the same institution and department is generated based on all messages sent by all doctors in the same institution and department within the preset window during their historical consultations.

[0058] The script database is configured to be within the same department, meaning it uses the script database for the department of the doctor whose information to be supplemented is located across all hospitals. This script database is the consultation script database for the same department. The consultation script database for the same department is generated based on all messages sent by all doctors in the same department across all institutions during their historical consultations within a preset window.

[0059] Remove messages without padding from all messages sent in the history of consultations within the preset window, and then use the remaining messages as the consultation script library.

[0060] Message removal without fill space includes: removing system-directed messages, removing quick reply messages, and removing duplicate messages.

[0061] To remove system-generated messages, which are part of the doctor's consultation process, some messages are directly presented by the system, such as when a prescription is issued and the doctor selects the prescription drug, and the system directly presents the drug name and dosage. There is no space to complete these messages, so they are removed.

[0062] The removal of quick reply messages refers to messages that some doctors have set up, such as "Please pay close attention to changes in your health. Please give me a five-star review after the consultation, thank you very much." These messages are sent by doctors with a single keystroke using shortcuts or special keys, and there is no space to complete them, so they have been removed.

[0063] To remove duplicate messages, which is because doctors habitually send spaces when sending messages, causing differences between messages and greatly increasing the number of messages in the script library, we need to remove spaces from messages and then remove completely identical messages.

[0064] This embodiment uses a pre-configured information completion mode. The range of values ​​configured for the dialogue script library includes: the patient, the same institution and department, and the same department. The subsequent use of the consultation dialogue script library corresponding to the information completion mode caters to the doctor's dialogue habits and meets the patient's need for personalization.

[0065] In one embodiment, the step of searching for consultation scripts from the consultation script library corresponding to the information completion mode based on the doctor information to be completed corresponding to the information completion signal to obtain a first consultation script set includes:

[0066] S21: Perform word segmentation on the doctor information to be completed to obtain the first word segmentation result;

[0067] Specifically, the doctor information to be completed is segmented into words, and all the phrases obtained from the segmentation are used as the first segmentation result.

[0068] S22: Calculate the TF score for the second word segmentation result corresponding to each of the consultation dialogues in the consultation dialogue library corresponding to the first word segmentation result and the information completion mode;

[0069] TF score, short for Term Frequency score, is also known as word frequency score.

[0070] The consultation script database stores the consultation scripts and the associated data corresponding to the second word segmentation results.

[0071] The methods and steps for calculating TF scores can be chosen from existing technologies, and will not be elaborated here.

[0072] S23: Take the M consultation scripts with the largest TF scores in the consultation script library corresponding to the information completion mode as the first consultation script set;

[0073] The second word segmentation result is data obtained by segmenting the word index of the consultation script. The word index is data obtained by indexing the first K characters of the consultation script using Elasticsearch's index building method.

[0074] Elasticsearch is a distributed search and analytics engine.

[0075] It is understood that the number of characters in the doctor's information to be completed is less than or equal to K, where K is an integer greater than 1.

[0076] The steps for indexing the first K characters of the consultation script using Elasticsearch's indexing method can be chosen from existing technologies and will not be elaborated here.

[0077] In this embodiment, the M consultation scripts with the highest TF scores in the consultation script library corresponding to the information completion mode are used as the first consultation script set, thereby achieving the search for the M most similar consultation scripts as the first consultation script set; through the consultation script library corresponding to the information completion mode, the doctor's script habits are catered to, and the patient's need for personalization is met.

[0078] In one embodiment, the step of performing bad recall result deletion processing on the first consultation script set based on the doctor information to be supplemented, to obtain the second consultation script set, includes:

[0079] S31: Based on the dynamic programming method, align the doctor information to be completed with each of the consultation scripts in the first consultation script set to obtain the aligned text and the completed text corresponding to each of the consultation scripts in the first consultation script set;

[0080] Dynamic programming, also known as rolling programming, uses the result of the previous calculation as the input for the current calculation to achieve dynamic planning.

[0081] Specifically, based on the dynamic programming method, the doctor information to be supplemented is aligned with each of the consultation scripts in the first consultation script set. The part of the consultation script that is similar to the doctor information to be supplemented is used as the aligned text, and the part of the consultation script that is not aligned with the aligned text is used as the supplemented text.

[0082] For example, if the doctor information to be completed is "The causes of eczema are complex and varied", and the consultation script is "The causes of eczema are quite complex and no one can say for sure", then the aligned text is "The causes of eczema are quite complex", and the completed text is "No one can say for sure".

[0083] For example, if the doctor information to be completed is "The causes of eczema are complex and diverse", and the consultation script is "The causes of dermatitis are complex and diverse and related to multiple factors", then the aligned text is "The causes of dermatitis are complex and diverse", and the completed text is "related to multiple factors".

[0084] For example, if the doctor information to be completed is "The causes of eczema are complex and diverse", and the consultation script is "The causes of eczema include internal and external factors", then the aligned text is "The causes of eczema", and the completed text is "Including internal and external factors".

[0085] S32: Based on the doctor information to be completed, each of the aligned texts and each of the completed texts, perform bad recall result deletion processing on the first consultation script set;

[0086] Bad recall results include at least the consultation scripts in which the doctor information to be supplemented differs significantly from the aligned text.

[0087] Specifically, the similarity is calculated between the doctor information to be completed and each of the aligned texts. Consultation scripts with a similarity less than a preset similarity threshold are deleted from the first consultation script set. Consultation scripts with a similarity less than the preset similarity threshold (i.e., with large differences) are bad recall results. The similarity can be cosine similarity.

[0088] S33: The first set of consultation scripts that has been deleted is used as the second set of consultation scripts.

[0089] Specifically, the first set of consultation scripts that has completed the deletion process, which is the first set of consultation scripts that has completed the deletion process of bad recall results, is a high-quality search result related to the doctor information to be supplemented. Therefore, the first set of consultation scripts that has completed the deletion process is used as the second set of consultation scripts.

[0090] This embodiment performs bad recall result deletion processing on the first consultation script set according to each of the aligned texts and each of the completed texts, and uses the first consultation script set after deletion processing as the second consultation script set, thereby avoiding bad recall results from entering the sorting stage and improving the accuracy of the determined information completion results.

[0091] In one embodiment, the step of aligning the doctor information to be completed with each of the consultation scripts in the first consultation script set based on a dynamic programming method to obtain the aligned text and completed text corresponding to each of the consultation scripts in the first consultation script set includes:

[0092] S311: Perform word segmentation on the doctor information to be completed to obtain the first word segmentation result;

[0093] Specifically, the doctor information to be completed is segmented into words, and all the phrases obtained from the segmentation are used as the first segmentation result.

[0094] S312: Perform punctuation-ignoring processing and word segmentation on each of the first consultation scripts in the first consultation script set to obtain the third word segmentation result;

[0095] Specifically, each of the consultation scripts in the first consultation script set is processed by ignoring punctuation marks, and the consultation scripts that have completed the punctuation mark ignoring processing are segmented into words. All phrases obtained by segmenting a consultation script are used as the third segmentation result.

[0096] S313: Based on the dynamic programming method and the principle that synonyms belong to the same word, the first word segmentation result and each of the third word segmentation results are aligned to obtain the aligned text and the completed text corresponding to each of the third word segmentation results.

[0097] The principle that synonyms belong to the same word is that synonyms are considered the same word.

[0098] Specifically, based on dynamic programming and the principle that synonyms belong to the same word, the first word segmentation result and each of the third word segmentation results are aligned. The part of the consultation script that is similar to the doctor information to be completed is taken as the aligned text, and the part of the consultation script that is not aligned is taken as the completed text.

[0099] This embodiment performs punctuation-ignoring and word segmentation processing on each of the first consultation scripts in the set of consultation scripts, thereby avoiding the influence of punctuation on alignment and improving the accuracy of the determined aligned text and the completed text. By using the principle that synonyms belong to the same word, alignment is achieved based on semantic similarity, further improving the accuracy of the determined aligned text and the completed text.

[0100] In one embodiment, the step of performing bad recall result deletion processing on the first consultation script set based on the doctor information to be completed, each of the aligned texts, and each of the completed texts includes:

[0101] S321: Using a preset similarity function, calculate the similarity between the doctor information to be completed and each of the aligned texts to obtain candidate similarities;

[0102] Specifically, a preset similarity function is used to calculate the similarity between the doctor information to be completed and each of the aligned texts, and each calculated similarity is taken as a candidate similarity.

[0103] S322: Delete the consultation script corresponding to each aligned text whose candidate similarity is less than a preset similarity threshold from the first consultation script set;

[0104] Specifically, the consultation script corresponding to each aligned text whose candidate similarity is less than the preset similarity threshold is a bad recall result because the aligned part (i.e., the aligned text) of the consultation script differs too much from the doctor information to be completed. Therefore, the consultation script corresponding to each aligned text whose candidate similarity is greater than the preset similarity threshold is deleted from the first consultation script set.

[0105] S323: Delete the consultation script corresponding to the completed text with a character count greater than the first character count threshold from the first consultation script set;

[0106] Specifically, the consultation scripts corresponding to the completed texts with a word count greater than the first word count threshold have large sections of redundant content, which are considered bad recall results. Therefore, the consultation scripts corresponding to each aligned text with a candidate similarity greater than the preset similarity threshold are deleted from the first consultation script set.

[0107] S324: Delete the consultation script corresponding to the completed text whose word count is less than the second word count threshold from the first consultation script set;

[0108] Wherein, the second character count threshold is less than the first character count threshold;

[0109] The formula for calculating the similarity function, sim, is:

[0110]

[0111]

[0112] q and p are constants. It is the Jaccard distance between the doctor information to be completed and the aligned text. The doctor information that needs to be supplemented is as described. It is the aligned text, `max` calculates the edit distance at the character level, `len` calculates the maximum value, and `len` calculates the character length.

[0113] Specifically, the consultation scripts corresponding to the completed text with a character count less than the second character count threshold are mostly meaningless characters. In order not to cause interference to doctors, these consultation scripts need to be deleted. Therefore, the consultation scripts corresponding to the completed text with a character count less than the second character count threshold are deleted from the first consultation script set.

[0114] It is the character sequence corresponding to the doctor information to be completed. It is the character sequence corresponding to the aligned text. It calculates the character-level edit distance between the character sequence corresponding to the doctor's information to be completed and the character sequence corresponding to the aligned text.

[0115] Calculate the edit distance at the character level. It is the difference between words in two sequences that enables the similarity function to measure the similarity between two sequences at both the character and word levels, thereby improving the accuracy of the determined candidate similarity.

[0116] This embodiment removes the consultation scripts corresponding to each aligned text with a candidate similarity greater than a preset similarity threshold from the first consultation script set; removes the consultation scripts corresponding to the completed texts with a word count greater than a first word count threshold from the first consultation script set; and removes the consultation scripts corresponding to the completed texts with a word count less than a second word count threshold from the first consultation script set. This achieves the removal of bad recall results from the first consultation script set, avoids bad recall results from entering the sorting stage, and improves the accuracy of the determined information completion results.

[0117] In one embodiment, the step of sorting the second set of consultation scripts to obtain a third set of consultation scripts includes:

[0118] S41: Using a preset sorting model, sort the second set of consultation scripts to obtain the third set of consultation scripts;

[0119] The ranking model is a model trained based on a preset ranking algorithm, and the ranking algorithm's rank formula is:

[0120]

[0121] , and The data is obtained through training, sim is the similarity function, and log is the logarithmic function. This refers to the number of times the consultation script is used within a preset time window.

[0122] Specifically, the preset sorting model is a model trained based on the rank sorting algorithm. , and These are the parameters that need to be trained during model training, therefore , and This is the data obtained through training. sim is the similarity function described in step S321.

[0123] This embodiment sorts the second set of medical consultation scripts using a model trained based on the ranking algorithm rank, thereby improving the efficiency and accuracy of the sorting.

[0124] Please see Figure 3 As shown, in one embodiment, an information completion device suitable for medical consultation is provided, the device comprising:

[0125] The signal acquisition module 801 is used to acquire information completion signals based on a pre-configured information completion mode.

[0126] Search module 802 is used to search for consultation scripts from the consultation script library corresponding to the information completion mode based on the doctor information to be completed corresponding to the information completion signal, and obtain a first consultation script set.

[0127] The deletion processing module 803 is used to perform bad recall result deletion processing on the first consultation script set according to the doctor information to be supplemented, so as to obtain the second consultation script set.

[0128] The sorting module 804 is used to sort the second set of consultation scripts to obtain the third set of consultation scripts;

[0129] The information completion result determination module 805 is used to generate the information completion result corresponding to the doctor information to be completed based on the highest-ranked consultation script in the third consultation script set.

[0130] Compared to sending pre-configured quick information via keyword triggering or a single mouse click, this embodiment utilizes a consultation script library corresponding to the information completion mode, catering to doctors' communication habits and meeting patients' personalized needs. By deleting bad recall results from the first consultation script set, it avoids bad recall results from entering the sorting stage, improving the accuracy of the determined information completion results. By responding to information completion signals to generate information completion results in real time, doctors only need to input partial information for automatic information completion, reducing the amount of text input by doctors and effectively improving consultation efficiency.

[0131] In one embodiment, the apparatus further includes:

[0132] The mode configuration module is used to obtain a mode configuration request and configure the information completion mode according to the mode configuration data carried in the mode configuration request.

[0133] The mode configuration data includes: script library range configuration, and the value range of the script library range configuration includes: self, same institution and same department, and same department.

[0134] In one embodiment, the step of the search module 802 searching for consultation scripts from the consultation script library corresponding to the information completion mode based on the doctor information to be completed corresponding to the information completion signal, and obtaining a first consultation script set, includes:

[0135] The doctor information to be completed is segmented into words to obtain the first segmentation result;

[0136] TF score is calculated for the second word segmentation result corresponding to each of the consultation dialogues in the consultation dialogue library corresponding to the first word segmentation result and the information completion mode;

[0137] The M consultation scripts with the highest TF scores in the consultation script library corresponding to the information completion mode are taken as the first consultation script set;

[0138] The second word segmentation result is data obtained by segmenting the word index of the consultation script. The word index is data obtained by indexing the first K characters of the consultation script using Elasticsearch's index building method.

[0139] In one embodiment, the step of the deletion processing module 803 performing bad recall result deletion processing on the first consultation script set based on the doctor information to be supplemented, to obtain the second consultation script set, includes:

[0140] Based on the dynamic programming method, the doctor information to be completed is aligned with each of the consultation scripts in the first consultation script set to obtain the aligned text and the completed text corresponding to each of the consultation scripts in the first consultation script set.

[0141] Based on each of the aligned texts and each of the completed texts, perform bad recall result deletion processing on the first consultation script set;

[0142] The first set of consultation scripts that has been deleted is used as the second set of consultation scripts.

[0143] In one embodiment, the step of the deletion processing module 803 aligning the doctor information to be supplemented with each of the consultation scripts in the first consultation script set using the dynamic programming method to obtain the aligned text and supplemented text corresponding to each of the consultation scripts in the first consultation script set includes:

[0144] The doctor information to be completed is segmented into words to obtain the first segmentation result;

[0145] Each of the consultation scripts in the first consultation script set is processed by ignoring punctuation marks and by word segmentation to obtain a third word segmentation result;

[0146] Based on the dynamic programming method and the principle that synonyms belong to the same word, the first word segmentation result and each of the third word segmentation results are aligned to obtain the aligned text and the completed text corresponding to each of the third word segmentation results.

[0147] In one embodiment, the step of the deletion processing module 803 performing bad recall result deletion processing on the first consultation script set based on each of the aligned texts and each of the completed texts includes:

[0148] A preset similarity function is used to calculate the similarity between the doctor information to be completed and each of the aligned texts to obtain candidate similarities;

[0149] The consultation script corresponding to each aligned text whose candidate similarity is greater than a preset similarity threshold is deleted from the first consultation script set;

[0150] The consultation scripts corresponding to the completed texts with a character count greater than the first character count threshold are deleted from the first consultation script set.

[0151] The consultation scripts corresponding to the completed texts with a character count less than the second character count threshold shall be deleted from the first consultation script set.

[0152] Wherein, the second character count threshold is less than the first character count threshold;

[0153] The formula for calculating the similarity function, sim, is:

[0154]

[0155]

[0156] q and p are constants. It is the Jaccard distance between the doctor information to be completed and the aligned text. The doctor information that needs to be supplemented is as described. It is the aligned text, `max` calculates the edit distance at the character level, `len` calculates the maximum value, and `len` calculates the character length.

[0157] In one embodiment, the step of sorting the second set of medical consultation scripts to obtain a third set of medical consultation scripts by the sorting module 804 includes:

[0158] The second set of consultation scripts is sorted using a preset sorting model to obtain the third set of consultation scripts.

[0159] The ranking model is a model trained based on a preset ranking algorithm, and the ranking algorithm's rank formula is:

[0160]

[0161] , and The data is obtained through training, sim is the similarity function, and log is the logarithmic function. This refers to the number of times the consultation script is used within a preset time window.

[0162] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements functions or steps on the client side of an information completion method suitable for online medical consultation.

[0163] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, performs the following steps:

[0164] Based on the pre-configured information completion mode, obtain the information completion signal;

[0165] Based on the doctor information to be completed corresponding to the information completion signal, search for consultation scripts from the consultation script library corresponding to the information completion mode to obtain the first consultation script set;

[0166] Based on the doctor information to be supplemented, the first consultation script set is processed by deleting bad recall results to obtain the second consultation script set.

[0167] The second set of consultation scripts is sorted to obtain the third set of consultation scripts.

[0168] Based on the highest-ranked consultation script in the third consultation script set, generate the information completion result corresponding to the doctor information to be completed.

[0169] Compared to sending pre-configured quick information via keyword triggering or a single mouse click, this embodiment utilizes a consultation script library corresponding to the information completion mode, catering to doctors' communication habits and meeting patients' personalized needs. By deleting bad recall results from the first consultation script set, it avoids bad recall results from entering the sorting stage, improving the accuracy of the determined information completion results. By responding to information completion signals to generate information completion results in real time, doctors only need to input partial information for automatic information completion, reducing the amount of text input by doctors and effectively improving consultation efficiency.

[0170] In one embodiment, a computer-readable storage medium is provided that stores a computer program, which, when executed by a processor, performs the following steps:

[0171] Based on the pre-configured information completion mode, obtain the information completion signal;

[0172] Based on the doctor information to be completed corresponding to the information completion signal, search for consultation scripts from the consultation script library corresponding to the information completion mode to obtain the first consultation script set;

[0173] Based on the doctor information to be supplemented, the first consultation script set is processed by deleting bad recall results to obtain the second consultation script set.

[0174] The second set of consultation scripts is sorted to obtain the third set of consultation scripts.

[0175] Based on the highest-ranked consultation script in the third consultation script set, generate the information completion result corresponding to the doctor information to be completed.

[0176] Compared to sending pre-configured quick information via keyword triggering or a single mouse click, this embodiment utilizes a consultation script library corresponding to the information completion mode, catering to doctors' communication habits and meeting patients' personalized needs. By deleting bad recall results from the first consultation script set, it avoids bad recall results from entering the sorting stage, improving the accuracy of the determined information completion results. By responding to information completion signals to generate information completion results in real time, doctors only need to input partial information for automatic information completion, reducing the amount of text input by doctors and effectively improving consultation efficiency.

[0177] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0178] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0179] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0180] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for completing information during medical consultation, the method comprising: Based on the pre-configured information completion mode, obtain the information completion signal; Based on the doctor information to be completed corresponding to the information completion signal, search for consultation scripts from the consultation script library corresponding to the information completion mode to obtain the first consultation script set; Based on the doctor information to be supplemented, the first consultation script set is processed by deleting bad recall results to obtain the second consultation script set. The second set of consultation scripts is sorted to obtain the third set of consultation scripts. Based on the highest-ranked consultation script in the third consultation script set, generate the information completion result corresponding to the doctor information to be completed; The step of performing bad recall deletion processing on the first consultation script set based on the doctor information to be supplemented, to obtain the second consultation script set, includes: Based on the dynamic programming method, the doctor information to be completed is aligned with each of the consultation scripts in the first consultation script set to obtain the aligned text and the completed text corresponding to each of the consultation scripts in the first consultation script set. Based on the doctor information to be completed, each of the aligned texts, and each of the completed texts, the first consultation script set is processed to delete bad recall results. The first set of consultation scripts that has been deleted is used as the second set of consultation scripts. The step of aligning the doctor information to be completed with each of the consultation scripts in the first consultation script set based on the dynamic programming method to obtain the aligned text and completed text corresponding to each of the consultation scripts in the first consultation script set includes: The doctor information to be completed is segmented into words to obtain the first segmentation result; Each of the consultation scripts in the first consultation script set is processed by ignoring punctuation marks and by word segmentation to obtain a third word segmentation result; Based on the dynamic programming method and the principle that synonyms belong to the same word, the first word segmentation result and each of the third word segmentation results are aligned to obtain the aligned text and the completed text corresponding to each of the third word segmentation results.

2. The information completion method for medical consultation according to claim 1, characterized in that, Before the step of obtaining the information completion signal based on the pre-configured information completion mode, the method further includes: Get the mode configuration request; Configure the information completion mode according to the mode configuration data carried in the mode configuration request; The mode configuration data includes: script library range configuration, and the value range of the script library range configuration includes: self, same institution and same department, and same department.

3. The information completion method for medical consultation according to claim 1, characterized in that, The step of searching for consultation scripts from the consultation script library corresponding to the information completion mode based on the doctor information to be completed according to the information completion signal, and obtaining the first consultation script set, includes: The doctor information to be completed is segmented into words to obtain the first segmentation result; TF score is calculated for the second word segmentation result corresponding to each of the consultation dialogues in the consultation dialogue library corresponding to the first word segmentation result and the information completion mode; The M consultation scripts with the highest TF scores in the consultation script library corresponding to the information completion mode are taken as the first consultation script set; The second word segmentation result is data obtained by segmenting the word index of the consultation script. The word index is data obtained by indexing the first K characters of the consultation script using Elasticsearch's index building method.

4. The information completion method for medical consultation according to claim 1, characterized in that, The step of performing bad recall result deletion processing on the first consultation script set based on the doctor information to be completed, each of the aligned texts, and each of the completed texts includes: A preset similarity function is used to calculate the similarity between the doctor information to be completed and each of the aligned texts to obtain candidate similarities; The consultation script corresponding to each aligned text whose candidate similarity is less than a preset similarity threshold is deleted from the first consultation script set; The consultation scripts corresponding to the completed texts with a character count greater than the first character count threshold are deleted from the first consultation script set. The consultation scripts corresponding to the completed texts with a character count less than the second character count threshold shall be deleted from the first consultation script set. Wherein, the second character count threshold is less than the first character count threshold; The formula for calculating the similarity function is: q and p are constants. It is the Jaccard distance between the doctor information to be completed and the aligned text. The doctor information that needs to be supplemented is as described. It is the aligned text, `max` calculates the edit distance at the character level, `len` calculates the maximum value, and `len` calculates the character length.

5. The information completion method for medical consultation according to claim 4, characterized in that, The step of sorting the second set of consultation scripts to obtain the third set of consultation scripts includes: The second set of consultation scripts is sorted using a preset sorting model to obtain the third set of consultation scripts. The ranking model is a model trained based on a preset ranking algorithm, and the ranking algorithm's rank formula is: , and The data is obtained through training, sim is the similarity function, and log is the logarithmic function. This refers to the number of times the consultation script is used within a preset time window.

6. An information completion device suitable for medical consultation, characterized in that, The device includes: The signal acquisition module is used to acquire information completion signals based on a pre-configured information completion mode; The search module is used to search for consultation scripts from the consultation script library corresponding to the information completion mode based on the doctor information to be completed corresponding to the information completion signal, and obtain the first consultation script set. The deletion processing module is used to perform bad recall result deletion processing on the first consultation script set according to the doctor information to be supplemented, so as to obtain the second consultation script set. The sorting module is used to sort the second set of consultation scripts to obtain the third set of consultation scripts; The information completion result determination module is used to generate the information completion result corresponding to the doctor information to be completed based on the highest-ranked consultation script in the third consultation script set; The step of performing bad recall deletion processing on the first consultation script set based on the doctor information to be supplemented, to obtain the second consultation script set, includes: Based on the dynamic programming method, the doctor information to be completed is aligned with each of the consultation scripts in the first consultation script set to obtain the aligned text and the completed text corresponding to each of the consultation scripts in the first consultation script set. Based on the doctor information to be completed, each of the aligned texts, and each of the completed texts, the first consultation script set is processed to delete bad recall results. The first set of consultation scripts that has been deleted is used as the second set of consultation scripts. The step of aligning the doctor information to be completed with each of the consultation scripts in the first consultation script set based on the dynamic programming method to obtain the aligned text and completed text corresponding to each of the consultation scripts in the first consultation script set includes: The doctor information to be completed is segmented into words to obtain the first segmentation result; Each of the consultation scripts in the first consultation script set is processed by ignoring punctuation marks and by word segmentation to obtain a third word segmentation result; Based on the dynamic programming method and the principle that synonyms belong to the same word, the first word segmentation result and each of the third word segmentation results are aligned to obtain the aligned text and the completed text corresponding to each of the third word segmentation results.

7. 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 steps of the information completion method for medical consultation as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the information completion method for medical consultation as described in any one of claims 1 to 5.

Images