CN112652386A - Triage data processing method and device, computer equipment and storage medium - Google Patents
Triage data processing method and device, computer equipment and storage mediumDownload PDFInfo
- Publication number
- CN112652386A CN112652386ACN202011565047.4ACN202011565047ACN112652386ACN 112652386 ACN112652386 ACN 112652386ACN 202011565047 ACN202011565047 ACN 202011565047ACN 112652386 ACN112652386 ACN 112652386A
- Authority
- CN
- China
- Prior art keywords
- data
- unstructured
- patient
- triage
- structured
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
Description
Technical Field
The invention relates to the technical field of artificial intelligence intelligent decision making, in particular to a diagnosis data processing method, a diagnosis data processing device, computer equipment and a storage medium.
Background
With the progress and development of medical science, hospitals are more specialized in department setting, the problem brought with the professional selection is that a user brings certain difficulty, and in order to solve the problem, each large hospital is additionally provided with a diagnosis guide link, including a diagnosis guide person and an autonomous diagnosis guide service, which mainly helps patients to recommend diagnosis departments.
At present, when a patient goes to a hospital for a diagnosis, the patient firstly needs to go to a diagnosis platform for manual diagnosis, a large amount of queuing time is consumed for the patient in the process, and the depth and the breadth of professional knowledge of a diagnosis guide staff of the diagnosis platform are higher, because the patient often carries the examination contents (or examination reports) of other hospitals for the diagnosis, in the prior art, when the patient is subjected to the diagnosis, the diagnosis guide staff usually needs to firstly read the examination contents (or the examination reports) and then carry out the diagnosis, at the moment, if the diagnosis guide staff gives the patient the wrong diagnosis, the patient needs to be subjected to the diagnosis again, the time of the patient is greatly wasted, the patient experience is seriously affected, and in addition, a reasonable diagnosis department or doctor is often difficult to give out when the diagnosis is wrong, and the patient experience can be further reduced.
Disclosure of Invention
The invention provides a triage data processing method, a device, computer equipment and a storage medium, which are used for realizing the purpose of accurately identifying a triage result by acquiring structured data and unstructured data of a patient, combining the structured data and the unstructured data, and performing triage prediction on the combined structured data and unstructured data.
A triage data processing method, comprising:
receiving a patient request of a patient, and acquiring a patient identification code and patient symptom information in the patient request; the patient symptom information includes symptom description and test data;
obtaining structured data and historical data associated with the patient identification code from a patient management database, and determining the patient symptom information and the historical data as unstructured data;
inputting the structured data and the unstructured data into a triage model; the triage model comprises an input preprocessing model and a language representation model;
performing code conversion on the structured data through the input preprocessing model to obtain structured embedded data, and performing non-structure conversion processing on the non-structured data to obtain non-structured embedded data;
splicing the structured embedded data and the unstructured embedded data to obtain data to be processed;
and performing triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
A triage data processing apparatus comprising:
the receiving module is used for receiving a patient request of a patient, and acquiring a patient identification code and patient symptom information in the patient request; the patient symptom information includes symptom description and test data;
an acquisition module for acquiring structured data and historical data associated with the patient identification code from a patient management database, and determining the patient symptom information and the historical data as unstructured data;
an input module for inputting the structured data and the unstructured data into a triage model; the triage model comprises an input preprocessing model and a language representation model;
the conversion module is used for performing code conversion on the structured data through the input preprocessing model to obtain structured embedded data and performing unstructured conversion processing on the unstructured data to obtain unstructured embedded data;
the splicing module is used for splicing the structured embedded data and the unstructured embedded data to obtain data to be processed;
and the prediction module is used for performing triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
A computer device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, said processor implementing the steps of the triage data processing method described above when executing said computer program.
A computer-readable storage medium, which stores a computer program that, when executed by a processor, implements the steps of the triage data processing method described above.
According to the triage data processing method, the triage data processing device, the computer equipment and the storage medium, the patient identification code and the patient symptom information in the patient request are obtained by receiving the patient request of the patient; the patient symptom information includes symptom description and test data; obtaining structured data and historical data associated with the patient identification code from a patient management database, and determining the patient symptom information and the historical data as unstructured data; inputting the structured data and the unstructured data into a triage model; the triage model comprises an input preprocessing model and a language representation model; performing code conversion on the structured data through the input preprocessing model to obtain structured embedded data, and performing non-structure conversion processing on the non-structured data to obtain non-structured embedded data; splicing the structured embedded data and the unstructured embedded data to obtain data to be processed; and performing triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
Thus, the present invention achieves this by obtaining the patient identification code and patient symptom information containing symptom description and verification data in the patient request; obtaining structured data and historical data from a patient management database, and determining the patient symptom information and the historical data as unstructured data; inputting the structured data and the unstructured data into a triage model; performing code conversion on the structured data to obtain structured embedded data, and performing non-structural conversion processing on the non-structured data to obtain non-structured embedded data; splicing the structured embedded data and the unstructured embedded data to obtain data to be processed; the data to be processed is subjected to triage prediction, and the triage result of the patient is predicted, so that the structured data and the unstructured data of the patient are obtained through the patient identification code and the patient symptom information provided by the patient, the structured data and the unstructured data of the patient are converted and spliced, and the spliced data are subjected to triage prediction, so that useful information is extracted through the common relation between the structured data and the unstructured data, the triage result of the patient is predicted, the department of the triage of the patient can be rapidly and accurately determined according to the information provided by the patient, the diagnosis accuracy is improved, and the patient experience is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments of the present invention will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without inventive labor.
FIG. 1 is a schematic diagram of an application environment of a triage data processing method according to an embodiment of the present invention;
FIG. 2 is a flow chart of a triage data processing method according to an embodiment of the present invention;
FIG. 3 is a flowchart of step S40 of the triage data processing method according to an embodiment of the present invention;
FIG. 4 is a flowchart of step S40 of the triage data processing method according to another embodiment of the present invention;
FIG. 5 is a flowchart of step S404 of a triage data processing method according to an embodiment of the invention;
FIG. 6 is a flowchart of step S405 of a triage data processing method according to an embodiment of the invention;
FIG. 7 is a functional block diagram of a triage data processing apparatus according to an embodiment of the present invention;
FIG. 8 is a functional block diagram of a conversion module of the triage data processing apparatus in an embodiment of the present invention;
FIG. 9 is a schematic diagram of a computer device in an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The triage data processing method provided by the invention can be applied to the application environment shown in fig. 1, wherein a client (computer device) communicates with a server through a network. The client (computer device) includes, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, cameras, and portable wearable devices. The server may be implemented as a stand-alone server or as a server cluster consisting of a plurality of servers.
In an embodiment, as shown in fig. 2, a method for processing triage data is provided, which mainly includes the following steps S10-S60:
s10, receiving a patient request of a patient, and acquiring the patient identification code and the patient symptom information in the patient request; the patient symptom information includes symptom description and test data.
Understandably, the patient identification code is a unique identification code registered and logged by a patient at an application platform, the patient needs to log at the application platform before issuing the patient request, the patient request is a request issued by the patient at the application platform, the patient symptom information is information which is input by the patient and is related to patient symptoms and test reports, the patient can input the symptom description and the test data on an application interface provided by the application platform, or select related symptoms from all symptom sets provided on the application interface as the symptom description of the patient, take a picture of a paper file report to be tested as the test data of the patient, and thus trigger the patient request after inputting the symptom description and the test data.
Wherein the symptom description is a description related to the symptom of the patient, and the test data is electronic data of a test report after the patient completes the related test, namely imaging data of a paper file after the patient completes the related test and the test report is photographed.
S20, acquiring the structured data and the historical data associated with the patient identification code from the patient management database, and determining the patient symptom information and the historical data as unstructured data.
Understandably, the patient management database is used for storing and managing structured data and historical data of all patients, the patient management database establishes an association relationship between a patient identification code corresponding to a patient and the structured data and the historical data of the patient, the structured data and the historical data associated with the patient identification code are searched from the patient management database, and the patient symptom information and the historical data are marked as unstructured data.
Wherein the structured data, also called row data, is data logically expressed and implemented by a two-dimensional table structure, the structured data is mainly stored and managed by a relational database, and the structured data is basic information of a patient corresponding to a patient identification code associated therewith, such as: the patient's age, sex, height, weight, etc., and the historical data is the data of the historical visit related to the patient corresponding to the patient identification code associated with the patient, that is, the historical data is the data stored according to the visit time node of the patient and is associated with the patient's chief complaint information, the test result, the past history identification, etc., such as: the unstructured data are data which are irregular in data structure, have no predefined data model, are inconvenient to represent by a database two-dimensional logic table, can be office documents, texts, reports, images, audio/video information and the like in all formats, and comprise patient symptom information and historical data.
S30, inputting the structured data and the unstructured data into a triage model; the triage model includes an input preprocessing model and a language representation model.
Understandably, the triage model is a trained neural network model, the triage model can be used for carrying out triage prediction to obtain a triage result by combining the structured data and the unstructured data, and the triage model comprises an input preprocessing model and a language representation model.
And S40, performing code conversion on the structured data through the input preprocessing model to obtain structured embedded data, and performing unstructured conversion processing on the unstructured data to obtain unstructured embedded data.
Understandably, the preprocessing model is a model for converting input structured data and unstructured data into a uniform executable data format, the structured data is subjected to one-hot coding conversion through the input preprocessing model, the one-hot coding conversion is to convert the structured data into the structured embedded data by using a one-hot coding technique, the one-hot coding technique is a technique for allocating a vector value corresponding to one-to-one mapping of each type of information content in the structured data, namely, coding each type of information, and then converting the structured data into an array vector, and the unstructured data is subjected to unstructured conversion processing so as to convert the unstructured data into the unstructured embedded data, wherein the unstructured conversion processing is to identify a data type in the unstructured data, and performing image-text conversion and chief complaint analysis according to the data types of the images, analyzing text data useful for diagnosis, and identifying the conversion process of the text data of the symptom description and the historical data.
Wherein the structured embedded data is a one-dimensional array vector, the unstructured embedded data is a multi-dimensional array vector, and the unstructured embedded data includes a multi-dimensional array vector associated with the patient symptom information and a multi-dimensional array vector associated with the historical data.
In an embodiment, as shown in fig. 3, the step S40, namely performing transcoding on the structured data to obtain structured embedded data, includes:
s401, encoding basic information by using a one-hot encoding technology through the input preprocessing model to obtain basic embedded data; the structured data includes the basic information and physical examination data.
Understandably, the one-hot encoding technique is a technique that allocates a vector value corresponding to a one-to-one mapping for each type of information content in the structured data, that is, encodes each type of information, and then converts the structured data into an array vector, where the structured data includes the basic information and physical examination data, and the basic information is information on basic aspects of a patient, such as: the patient's age, sex, height, weight, etc., and the physical examination data is data of a general physical examination that the patient regularly or occasionally performs, such as: blood pressure, vision, chest X-ray and the like of the patient.
Wherein, the basic information is encoded by using the one-hot encoding technique to generate the basic embedded data, for example: in the process of applying the one-hot encoding technology, the input preprocessing model comprises embedded layers with structures such as agembeddings (age embedded layers), sexembeddings (gender embedded layers), weightembeddings (weight embedded layers), highembeddings (height embedded layers), and the like, and the age of a patient is represented by encoding the age of the patient into an age embedded layer corresponding to the age embedded layer, for example, the age of one patient is 25 years, and the corresponding age embedded layer is A25; the gender of the patient is indicated by the gender embedding layer encoding the gender of the patient into the gender embedding corresponding thereto, e.g., one patient's gender is male and the corresponding gender embedding is Smale.
S402, extracting useful data corresponding to preset parameters from the physical examination data through the input preprocessing model, and converting and generating peripheral embedded data corresponding to the useful data by applying a rule mapping conversion technology.
Understandably, the preset parameter may be set according to a requirement, for example, the preset parameter may be a field parameter in some preset physical examination data, the preset parameter may also be a field parameter that is out of a standard range of the field parameter in the physical examination data corresponding to the patient identification code, the useful data corresponding to the preset parameter is screened from the physical examination data, the useful data is data that is screened for subsequent diagnosis prediction, the rule mapping conversion technique is to map a value vector corresponding to a data content in the useful data according to a rule range that the data content in the useful data conforms to, for example, the content of a blood pressure parameter in the useful data conforms to a definition rule of hypertension, then map the content of the blood pressure parameter into a value vector corresponding to hypertension (for example, 10), and the content of the blood pressure parameter in the useful data conforms to a definition rule of hypotension, the content of the blood pressure parameters is mapped to a numerical vector (for example, 1) corresponding to the hypotension, the content of the electrocardio parameters in the useful data conforms to the definition rule of the arrhythmia, the content of the electrocardio parameters is mapped to a numerical vector (for example, 8) corresponding to the arrhythmia, and the like, so that the peripheral embedded data corresponding to the useful data is generated through conversion.
And the peripheral embedded data is a one-dimensional array vector.
And S403, adding the basic embedded data and the peripheral embedded data through the input preprocessing model to obtain the structured embedded data.
Understandably, the basic embedded data and the peripheral embedded data are subjected to array one-dimensional addition to obtain one-dimensional structured embedded data.
The invention realizes that the basic information in the structured data is coded by using one-hot coding technology to obtain basic embedded data; comprises the basic information and physical examination data; extracting useful data corresponding to preset parameters from the physical examination data in the structured data, and converting to generate peripheral embedded data corresponding to the useful data by using a rule mapping conversion technology; and adding the basic embedded data and the peripheral embedded data to obtain the structured embedded data, so that the structured data is converted into the structured embedded data by using a one-hot coding technology and a rule mapping conversion technology, a data source is provided for subsequent triage prediction, and the structured embedded data is converted into an executable array vector, and the efficiency is improved for subsequent identification.
In an embodiment, as shown in fig. 4, the step S40, namely performing an unstructured conversion process on the unstructured data to obtain unstructured embedded data, includes:
s404, text conversion and combination are carried out on the unstructured data through the input preprocessing model, and unstructured text data are obtained.
Understandably, the processing procedure of text conversion and merging is a procedure of performing image-text conversion on the content in the non-text type picture format in the unstructured data and analyzing the main complaints to obtain useful text data, performing word vector conversion on the content in the text format to obtain text data, and merging all the obtained text data, wherein the non-structural text data is a set of all the text data obtained after conversion.
In an embodiment, as shown in fig. 5, in the step S404, that is, performing text conversion and merging on the unstructured data through the input preprocessing model to obtain unstructured text data includes:
s4041, identifying the type of the inspection data in the unstructured data through the preprocessing model, identifying the data type corresponding to the inspection data, performing word vector conversion on the symptom description in the unstructured data to obtain first text data, and performing word vector conversion on the historical data in the unstructured data to obtain second text data.
Understandably, the type is identified as a type identifying which file format the test data is, such as a picture type in jpg format, an electronic file type in pdf format, and the like, and the test data is electronized data of a test report after the patient completes the relevant test, that is, imaging data in which a test report of a paper file after the patient completes the relevant test is photographed, or data of an electronic file of a test report of a scanned paper file.
The data type may be set according to a requirement, such as a video type in an mp4 format, a picture type in a jpg format, an electronic document type in a pdf format, and the like, the Word vector is converted into a conversion process that maps each unit character or Word into an N-dimensional Word vector by using the Word2Vec algorithm, and the Word2Vec algorithm predicts a vector corresponding to the unit character or Word by using a shallow Word2Vec model.
S4042, according to the identified data type, performing image-text conversion on the inspection data to obtain intermediate data.
Understandably, the teletext conversion is a conversion process corresponding to the identified data type, such as: and if the identified data type is a picture type in a jpg format, identifying the conversion process of characters in the picture by using an OCR technology, and if the identified data type is an electronic file type in a pdf format, performing snapshot by using a pdf editing tool and identifying the characters in the electronic file by using the OCR technology to obtain the intermediate data, wherein the intermediate data is text data.
S4043, performing chief complaint analysis on the intermediate data through the preprocessing model to obtain third text data.
Understandably, the chief complaint analysis is to extract text content of the conclusion of the test from the identified intermediate data, identify words with result semantics, such as "result", "conclusion", in the intermediate data by applying NLP text identification technology, analyze the paragraph content after the word, that is, identify the paragraph by text identification technology, identify keywords, and determine the identified keywords as the third text data, where the keywords are the frequency of occurrence in the whole paragraph and the words or words with medical character features or medical term features of the words.
S4044, determining the first text data, the second text data, and the third text data as the non-structural text data.
Understandably, the first text data, the second text data, and the third text data are labeled as the unstructured text data.
The method and the device realize the identification of the type of the inspection data in the unstructured data, identify the data type corresponding to the inspection data, perform word vector conversion on the symptom description in the unstructured data to obtain first text data, and perform word vector conversion on the historical data in the unstructured data to obtain second text data; performing image-text conversion on the inspection data according to the identified data type to obtain intermediate data; performing chief complaint analysis on the intermediate data through the preprocessing model to obtain third text data; the first text data, the second text data and the third text data are determined as the unstructured text data, so that after the type of the inspection data in the unstructured data is identified, image-text conversion and chief analysis are carried out on the data type to obtain useful text data in the inspection data, and symptom description and historical data are converted into text data through word vector conversion, so that the unstructured text data are obtained in a summary mode, the useful text data in the unstructured data are provided for subsequent triage prediction, and accuracy and efficiency of the triage prediction are improved.
S405, carrying out embedding representation processing on the unstructured text data through the input preprocessing model to obtain unstructured embedded data.
Understandably, the embedded representation processing adopts different unstructured embedding layers to represent the unstructured text data, and the embedded representation processing comprises splitting and coding of a word embedding layer, text representation of a type embedding layer and sequential representation of a position embedding layer, so as to obtain the unstructured embedded data.
In an embodiment, as shown in fig. 6, in step S405, that is, performing an embedding representation process on the text data through the input preprocessing model to obtain the unstructured embedded data, includes:
s4051, splitting and coding the text data through a word embedding layer to obtain first embedded data; the input pre-processing model comprises a word embedding layer, a type embedding layer and a position embedding layer.
Understandably, the splitting and encoding process is a process of splitting the text data into each word through the word embedding layer, encoding each word into one word embedding representation data corresponding to each word, and finally summarizing all the word embedding representation data to obtain the first embedding data, for example: "cough three days", split into 4 words, 4 words correspond to 4 embedded representation data respectively: e (cough), E (three), E (day).
Wherein the input pre-processing model comprises a word embedding layer, a type embedding layer and a position embedding layer.
S4052, performing text representation on the text data through the type embedding layer to obtain second embedding data.
Understandably, the text identification process is a process of converting the text type of the text data through the type embedding layer, and includes the type embedding layer representing the text of the chief complaint information, for example, the type embedding layer represents the text of the chief complaint information as t (cc), and represents the text of the past history identification information as t (pmh).
S4053, sequentially representing the text data through the position embedding layer to obtain third embedding data.
Understandably, the sequence representing process is a process of representing the sequence between the words in the text data through the position embedding layer, namely the sequence number of each word in the sentence to which the word belongs.
S4054, performing unstructured addition processing on the first embedded data, the second embedded data, and the third embedded data to obtain unstructured embedded data.
Understandably, the unstructured adding process is to superimpose a plurality of one-dimensional arrays of the first embedded data, the second embedded data and the third embedded data into a multidimensional array to obtain the multidimensional unstructured embedded data.
The invention realizes the splitting and coding of the text data through the word embedding layer to obtain the first embedded data; the input preprocessing model comprises a word embedding layer, a type embedding layer and a position embedding layer; performing text representation on the text data through the type embedding layer to obtain second embedding data; sequentially representing the text data through the position embedding layer to obtain third embedding data; and performing unstructured addition processing on the first embedded data, the second embedded data and the third embedded data to obtain unstructured embedded data, so that text data are converted into embedded data with different dimensions through different unstructured embedded layers to obtain unstructured embedded data with multiple dimensions, multi-dimensional useful information can be provided for subsequent triage prediction, and the identification accuracy is improved.
And S50, splicing the structured embedded data and the unstructured embedded data to obtain data to be processed.
Understandably, the one-dimensional structured embedded data and the multidimensional unstructured embedded data are spliced into the multidimensional data to be processed, and the splicing process is vertical splicing, namely, an array of the one-dimensional structured embedded data is added below the unstructured embedded data, so that the data to be processed is obtained.
And S60, performing triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
Understandably, the language representation model is a pre-trained language model completed through supervised learning training, the language representation model is used for carrying out symptom feature extraction on the data to be processed after combining structured data and unstructured data, a triage result corresponding to the patient is identified according to the symptom features, the triage result is a predicted triage class with the highest probability, the triage result comprises a department class, namely a class of a department for the patient to see a doctor, and the triage result provides an accurate basis for the patient to make an appointment, so that the patient can conveniently select an accurate department to make an appointment for seeing.
Thus, the present invention achieves this by obtaining the patient identification code and patient symptom information containing symptom description and verification data in the patient request; obtaining structured data and historical data from a patient management database, and determining the patient symptom information and the historical data as unstructured data; inputting the structured data and the unstructured data into a triage model; performing code conversion on the structured data to obtain structured embedded data, and performing non-structural conversion processing on the non-structured data to obtain non-structured embedded data; splicing the structured embedded data and the unstructured embedded data to obtain data to be processed; the data to be processed is subjected to triage prediction, and the triage result of the patient is predicted, so that the structured data and the unstructured data of the patient are obtained through the patient identification code and the patient symptom information provided by the patient, the structured data and the unstructured data of the patient are converted and spliced, and the spliced data are subjected to triage prediction, so that useful information is extracted through the common relation between the structured data and the unstructured data, the triage result of the patient is predicted, the department of the triage of the patient can be rapidly and accurately determined according to the information provided by the patient, the diagnosis accuracy is improved, and the patient experience is improved.
In an embodiment, before the step S60, that is, before the triage prediction is performed on the data to be processed through the language representation model to obtain a triage result corresponding to the patient, the method includes:
s601, acquiring a training sample set; the training sample set includes a plurality of training samples, the training samples associated with triage labels.
Understandably, the training sample set is the set of training samples, the training samples are historically collected information of the treatment of the patient associated with the patient identification code, and the information is processed by the method of steps S20 to S50, the triage labels are labels of the existing department categories, and one training sample is associated with one triage label.
S602, inputting the training sample into a pre-training language model containing initial parameters.
Understandably, the pre-training language model may be a model constructed based on RoBERTa or a model constructed based on BERT, the pre-training language model includes the initial parameters, and the initial parameters may be preset parameters or parameters migrated from other models.
The roberta (a Robustly Optimized BERT predicting approach) model performs a multi-task classification recognition model on input contents mainly in a Masking (Masking) manner, and the BERT (bidirectional Encoder responses from transformers) model is a multi-task recognition model recognized by jointly adjusting bidirectional transformers in all layers.
S603, carrying out classification and identification on the training samples through the pre-training language model to obtain sample results corresponding to the training samples.
Understandably, the classification recognition is a process of extracting symptom features of the training samples and performing full-connection recognition on the extracted symptom features, the symptom features are features of related symptoms of commonalities between the structured data and the unstructured data, and the sample results are departments of the treatment identified in the training process.
S604, determining a loss value according to the sample result corresponding to the training sample and the triage label associated with the training sample.
Understandably, the sample result corresponding to the training sample and the triage tag associated with the training sample are input into a loss function, and the loss value is calculated by the loss function, which is preferably a cross entropy (cross entropy) loss function.
And S605, when the loss value does not reach a preset convergence condition, iteratively updating the initial parameters of the pre-training language model until the loss value reaches the convergence condition, and recording the pre-training language model after convergence as a language representation model.
Understandably, the convergence condition may be a condition that the loss value is small and does not decrease again after 20000 times of calculation, that is, when the loss value is small and does not decrease again after 20000 times of calculation, stopping training, and recording the pre-training language model after convergence as a language representation model; the convergence condition may also be a condition that the loss value is smaller than a set threshold, that is, when the loss value is smaller than the set threshold, the training is stopped, and the pre-training language model after convergence is recorded as a language representation model, so that when the loss value does not reach the preset convergence condition, the initial parameters of the pre-training language model are continuously adjusted, and the step of performing classification recognition on the training sample through the pre-training language model to obtain a sample result corresponding to the training sample is triggered, so that the accurate result can be continuously drawn close to the accurate result, and the recognition accuracy is higher and higher. Therefore, triage identification can be optimized, and accuracy and reliability of triage identification are improved.
In an embodiment, a triage data processing apparatus is provided, and the triage data processing apparatus corresponds to the triage data processing method in the above embodiment one to one. As shown in fig. 7, the triage data processing apparatus includes a receiving module 11, an obtaining module 12, an input module 13, a converting module 14, a splicing model library 15, and a predicting module 16. The functional modules are explained in detail as follows:
the receiving module 11 is used for receiving a patient request of a patient, and acquiring a patient identification code and patient symptom information in the patient request; the patient symptom information includes symptom description and test data;
an obtaining module 12, configured to obtain structured data and historical data associated with the patient identification code from a patient management database, and determine the patient symptom information and the historical data as unstructured data;
an input module 13, configured to input the structured data and the unstructured data into a triage model; the triage model comprises an input preprocessing model and a language representation model;
a conversion module 14, configured to perform code conversion on the structured data through the input preprocessing model to obtain structured embedded data, and perform unstructured conversion processing on the unstructured data to obtain unstructured embedded data;
the splicing module 15 is configured to splice the structured embedded data and the unstructured embedded data to obtain data to be processed;
and the prediction module 16 is configured to perform triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
In one embodiment, as shown in fig. 8, the conversion module 14 includes:
a merging submodule 41, configured to perform text conversion and merging on the unstructured data through the input preprocessing model to obtain unstructured text data;
and the embedding submodule 42 is configured to perform embedding representation processing on the unstructured text data through the input preprocessing model to obtain the unstructured embedded data.
For specific limitations of the triage data processing apparatus, reference may be made to the above limitations on the triage data processing method, which are not described herein again. The modules in the triage data processing device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 9. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of triage data processing.
In one embodiment, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the triage data processing method in the above embodiments is implemented.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the triage data processing method in the above-described embodiments.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, databases, or other media used in embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can 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), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.
Claims (10)
1. A method for processing triage data, comprising:
receiving a patient request of a patient, and acquiring a patient identification code and patient symptom information in the patient request; the patient symptom information includes symptom description and test data;
obtaining structured data and historical data associated with the patient identification code from a patient management database, and determining the patient symptom information and the historical data as unstructured data;
inputting the structured data and the unstructured data into a triage model; the triage model comprises an input preprocessing model and a language representation model;
performing code conversion on the structured data through the input preprocessing model to obtain structured embedded data, and performing non-structure conversion processing on the non-structured data to obtain non-structured embedded data;
splicing the structured embedded data and the unstructured embedded data to obtain data to be processed;
and performing triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
2. The triage data processing method of claim 1, wherein the transcoding the structured data to obtain structured embedded data comprises:
encoding the basic information by using a one-hot encoding technology through the input preprocessing model to obtain basic embedded data; the structured data comprises the basic information and physical examination data;
extracting useful data corresponding to preset parameters from the physical examination data through the input preprocessing model, and converting and generating peripheral embedded data corresponding to the useful data by applying a rule mapping conversion technology;
and adding the basic embedded data and the peripheral embedded data through the input preprocessing model to obtain the structured embedded data.
3. The triage data processing method of claim 1, wherein the performing unstructured conversion processing on the unstructured data to obtain unstructured embedded data comprises:
performing text conversion and combination on the unstructured data through the input preprocessing model to obtain unstructured text data;
and carrying out embedding representation processing on the unstructured text data through the input preprocessing model to obtain the unstructured embedded data.
4. The triage data processing method of claim 3, wherein the text-converting and merging the unstructured data through the input preprocessing model to obtain unstructured text data comprises:
performing type recognition on the inspection data in the unstructured data through the preprocessing model, recognizing a data type corresponding to the inspection data, performing word vector conversion on the symptom description in the unstructured data to obtain first text data, and performing word vector conversion on the historical data in the unstructured data to obtain second text data;
performing image-text conversion on the inspection data according to the identified data type to obtain intermediate data;
performing chief complaint analysis on the intermediate data through the preprocessing model to obtain third text data;
determining the first text data, the second text data, and the third text data as the unstructured text data.
5. The triage data processing method of claim 3, wherein the embedding representation processing of the text data through the input preprocessing model to obtain the unstructured embedded data comprises:
splitting and coding the text data through a word embedding layer to obtain first embedded data; the input preprocessing model comprises a word embedding layer, a type embedding layer and a position embedding layer;
performing text representation on the text data through the type embedding layer to obtain second embedding data;
sequentially representing the text data through the position embedding layer to obtain third embedding data;
and carrying out unstructured addition processing on the first embedded data, the second embedded data and the third embedded data to obtain the unstructured embedded data.
6. The triage data processing method of claim 1, wherein before the triage prediction of the data to be processed by the language representation model to obtain the triage result corresponding to the patient, the method comprises:
acquiring a training sample set; the training sample set comprises a plurality of training samples, and the training samples are associated with triage labels;
inputting the training sample into a pre-training language model containing initial parameters;
classifying and identifying the training samples through the pre-training language model to obtain sample results corresponding to the training samples;
determining a loss value according to the sample result corresponding to the training sample and the triage label associated with the training sample;
and when the loss value does not reach a preset convergence condition, iteratively updating the initial parameters of the pre-training language model until the loss value reaches the convergence condition, and recording the pre-training language model after convergence as a language representation model.
7. A triage data processing apparatus, comprising:
the receiving module is used for receiving a patient request of a patient, and acquiring a patient identification code and patient symptom information in the patient request; the patient symptom information includes symptom description and test data;
an acquisition module for acquiring structured data and historical data associated with the patient identification code from a patient management database, and determining the patient symptom information and the historical data as unstructured data;
an input module for inputting the structured data and the unstructured data into a triage model; the triage model comprises an input preprocessing model and a language representation model;
the conversion module is used for performing code conversion on the structured data through the input preprocessing model to obtain structured embedded data and performing unstructured conversion processing on the unstructured data to obtain unstructured embedded data;
the splicing module is used for splicing the structured embedded data and the unstructured embedded data to obtain data to be processed;
and the prediction module is used for performing triage prediction on the data to be processed through the language representation model to obtain a triage result corresponding to the patient.
8. The triage data processing apparatus according to claim 7, wherein the conversion module includes:
the merging submodule is used for performing text conversion and merging on the unstructured data through the input preprocessing model to obtain unstructured text data;
and the embedding submodule is used for carrying out embedding representation processing on the unstructured text data through the input preprocessing model to obtain the unstructured embedded data.
9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the triage data processing method according to any one of claims 1 to 6 when executing the computer program.
10. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out the triage data processing method according to any one of claims 1 to 6.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011565047.4ACN112652386A (en) | 2020-12-25 | 2020-12-25 | Triage data processing method and device, computer equipment and storage medium |
| PCT/CN2021/084240WO2022134357A1 (en) | 2020-12-25 | 2021-03-31 | Triage data processing method and apparatus, and computer device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011565047.4ACN112652386A (en) | 2020-12-25 | 2020-12-25 | Triage data processing method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112652386Atrue CN112652386A (en) | 2021-04-13 |
Family
ID=75363030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011565047.4APendingCN112652386A (en) | 2020-12-25 | 2020-12-25 | Triage data processing method and device, computer equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112652386A (en) |
| WO (1) | WO2022134357A1 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113241135A (en)* | 2021-04-30 | 2021-08-10 | 山东大学 | Disease risk prediction method and system based on multi-mode fusion |
| CN113642312A (en)* | 2021-08-19 | 2021-11-12 | 平安医疗健康管理股份有限公司 | Medical examination data processing method, device, equipment and storage medium |
| CN113724854A (en)* | 2021-07-27 | 2021-11-30 | 广州医科大学附属第二医院 | Machine learning-based grading triage method and system and computer equipment |
| CN114834466A (en)* | 2022-06-08 | 2022-08-02 | 润光智能(深圳)有限公司 | Vehicle interior system diagnostic device, method, vehicle, and storage medium |
| CN115101186A (en)* | 2022-07-25 | 2022-09-23 | 武汉大学人民医院（湖北省人民医院） | Method and device for hospital consultation information management based on big data |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117556265B (en)* | 2024-01-11 | 2024-09-06 | 中国医学科学院医学信息研究所 | A detection index prediction method, device, equipment and storage medium |
| CN118690333A (en)* | 2024-08-26 | 2024-09-24 | 中国科学院自动化研究所 | Disease severity prediction method based on heterogeneous table information fusion |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109378064A (en)* | 2018-10-29 | 2019-02-22 | 南京医基云医疗数据研究院有限公司 | Medical data processing method, device electronic equipment and computer-readable medium |
| EP3564964A1 (en)* | 2018-05-04 | 2019-11-06 | Avaintec Oy | Method for utilising natural language processing technology in decision-making support of abnormal state of object |
| CN110867231A (en)* | 2019-11-18 | 2020-03-06 | 中山大学 | Disease prediction method, device, computer equipment and medium based on text classification |
| CN112017789A (en)* | 2020-09-09 | 2020-12-01 | 平安科技（深圳）有限公司 | Triage data processing method, device, equipment and medium |
| CN112016318A (en)* | 2020-09-08 | 2020-12-01 | 平安科技（深圳）有限公司 | Triage information recommendation method, device, equipment and medium based on interpretation model |
| CN112101550A (en)* | 2020-09-25 | 2020-12-18 | 平安科技（深圳）有限公司 | Triage fusion model training method, triage method, device, equipment and medium |
- 2020
- 2020-12-25CNCN202011565047.4Apatent/CN112652386A/enactivePending
- 2021
- 2021-03-31WOPCT/CN2021/084240patent/WO2022134357A1/ennot_activeCeased
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3564964A1 (en)* | 2018-05-04 | 2019-11-06 | Avaintec Oy | Method for utilising natural language processing technology in decision-making support of abnormal state of object |
| CN109378064A (en)* | 2018-10-29 | 2019-02-22 | 南京医基云医疗数据研究院有限公司 | Medical data processing method, device electronic equipment and computer-readable medium |
| CN110867231A (en)* | 2019-11-18 | 2020-03-06 | 中山大学 | Disease prediction method, device, computer equipment and medium based on text classification |
| CN112016318A (en)* | 2020-09-08 | 2020-12-01 | 平安科技（深圳）有限公司 | Triage information recommendation method, device, equipment and medium based on interpretation model |
| CN112017789A (en)* | 2020-09-09 | 2020-12-01 | 平安科技（深圳）有限公司 | Triage data processing method, device, equipment and medium |
| CN112101550A (en)* | 2020-09-25 | 2020-12-18 | 平安科技（深圳）有限公司 | Triage fusion model training method, triage method, device, equipment and medium |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113241135A (en)* | 2021-04-30 | 2021-08-10 | 山东大学 | Disease risk prediction method and system based on multi-mode fusion |
| CN113724854A (en)* | 2021-07-27 | 2021-11-30 | 广州医科大学附属第二医院 | Machine learning-based grading triage method and system and computer equipment |
| CN113642312A (en)* | 2021-08-19 | 2021-11-12 | 平安医疗健康管理股份有限公司 | Medical examination data processing method, device, equipment and storage medium |
| CN114834466A (en)* | 2022-06-08 | 2022-08-02 | 润光智能(深圳)有限公司 | Vehicle interior system diagnostic device, method, vehicle, and storage medium |
| CN115101186A (en)* | 2022-07-25 | 2022-09-23 | 武汉大学人民医院（湖北省人民医院） | Method and device for hospital consultation information management based on big data |
| CN115101186B (en)* | 2022-07-25 | 2022-11-08 | 武汉大学人民医院（湖北省人民医院） | Hospital treatment information management method and device based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022134357A1 (en) | 2022-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112652386A (en) | Triage data processing method and device, computer equipment and storage medium | |
| CN118627628B (en) | A large language model knowledge question answering method and system integrating multimodal knowledge graph | |
| CN112149414B (en) | Text similarity determination method, device, equipment and storage medium | |
| CN112016295B (en) | Symptom data processing method, symptom data processing device, computer equipment and storage medium | |
| CN112101550B (en) | Triage fusion model training method, triage device, triage equipment and medium | |
| CN108899064A (en) | Electronic health record generation method, device, computer equipment and storage medium | |
| CN112908473B (en) | Model-based data processing method, device, computer equipment and storage medium | |
| CN119293194B (en) | Medical record quality management system and method | |
| CN116343980B (en) | Intelligent medical review follow-up data processing method and system | |
| CN113707304B (en) | Triage data processing method, triage data processing device, triage data processing equipment and storage medium | |
| KR102642488B1 (en) | Data providing device, method and computer program generating answer using artificial intelligence technology | |
| US20230177267A1 (en) | Automated classification and interpretation of life science documents | |
| CN114153995B (en) | Processing method and device of medical terms, computer equipment and storage medium | |
| CN113707296B (en) | Medical scheme data processing method, device, equipment and storage medium | |
| CN117809794A (en) | Medical report generation method, system and terminal based on cross-view semantic alignment | |
| CN118093626A (en) | Information processing method and device, computer storage medium and electronic equipment | |
| CN119494347B (en) | A named entity recognition method and system based on knowledge-enhanced pre-training model | |
| CN119541750B (en) | Automatic generation method and system for inspection report | |
| CN114821613A (en) | A method and system for extracting table information in PDF | |
| US20250285718A1 (en) | Systems and methods for automatic medical report generation | |
| CN117314658A (en) | Risk prediction method and device, electronic equipment and storage medium | |
| HK40041499A (en) | Triage data processing method, device, computer equipment and storage medium | |
| CN112580309B (en) | Document data processing method, device, computer equipment and storage medium | |
| CN119296817B (en) | Method, device, computer equipment and storage medium for recommending consultation response techniques | |
| CN118692661B (en) | An auxiliary diagnosis method and system based on multi-view knowledge graph |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| REG | Reference to a national code | Ref country code:HK Ref legal event code:DE Ref document number:40041499 Country of ref document:HK | |
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |