CN115719628A - Traditional Chinese medicine prescription generation method, device, equipment and storage medium - Google Patents
Traditional Chinese medicine prescription generation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN115719628A CN115719628A CN202211435070.0A CN202211435070A CN115719628A CN 115719628 A CN115719628 A CN 115719628A CN 202211435070 A CN202211435070 A CN 202211435070A CN 115719628 A CN115719628 A CN 115719628A
- Authority
- CN
- China
- Prior art keywords
- chinese medicine
- character
- traditional chinese
- sequence
- sample
- 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.)
- Granted
Links
- 239000003814 drug Substances 0.000 title claims abstract description 315
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000012549 training Methods 0.000 claims abstract description 73
- 208000024891 symptom Diseases 0.000 claims abstract description 72
- 230000000873 masking effect Effects 0.000 claims abstract description 12
- 239000013598 vector Substances 0.000 claims description 74
- 238000004590 computer program Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 230000007246 mechanism Effects 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 7
- 230000011218 segmentation Effects 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 238000003745 diagnosis Methods 0.000 abstract description 12
- 230000008569 process Effects 0.000 abstract description 10
- 230000000694 effects Effects 0.000 abstract description 6
- 238000004891 communication Methods 0.000 description 8
- 229910052709 silver Inorganic materials 0.000 description 7
- 239000004332 silver Substances 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 240000002853 Nelumbo nucifera Species 0.000 description 5
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 5
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 5
- 229910052737 gold Inorganic materials 0.000 description 5
- 239000010931 gold Substances 0.000 description 5
- 238000013507 mapping Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 229940079593 drug Drugs 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 210000003800 pharynx Anatomy 0.000 description 3
- 240000000233 Melia azedarach Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 229940126678 chinese medicines Drugs 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 210000000867 larynx Anatomy 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for generating a traditional Chinese medicine prescription, wherein the method for generating the traditional Chinese medicine prescription comprises the following steps: acquiring a target symptom character sequence of a target patient; inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model; and determining a target traditional Chinese medicine prescription of the target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model. According to the technical scheme of the embodiment of the invention, the more accurate traditional Chinese medicine prescription can be automatically generated by utilizing the traditional Chinese medicine prediction network model obtained by finely adjusting the pre-training model, and a more scientific prescription reference is provided for the traditional Chinese medicine in the process of configuring the prescription for the patient by the traditional Chinese medicine, so that the diagnosis and treatment efficiency is improved, and the diagnosis and treatment effect is ensured.
Description
Technical Field
The invention relates to the technical field of medical treatment, in particular to a method, a device, equipment and a storage medium for generating a traditional Chinese medicine prescription.
Background
As the oldest traditional medicine in China, the traditional Chinese medicine plays an important role in the medical system in China. The Chinese medicine has strong subjectivity in treatment and uneven diagnosis and treatment level. At present, traditional Chinese medicine often configures a traditional Chinese medicine prescription for a patient according to self experience and professional knowledge based on analysis of user symptoms and examination information. However, this method strictly requires the level of traditional Chinese medicine, and there may be cases of wrong use and missed use of medicine, which affects the diagnosis and treatment effect, so it is of great significance to construct an auxiliary system capable of generating a corresponding prescription for the patient's disease.
Disclosure of Invention
The invention provides a method, a device, equipment and a storage medium for generating a traditional Chinese medicine prescription, which are used for realizing the automatic generation of the traditional Chinese medicine prescription, providing prescription reference for the traditional Chinese medicine, improving the diagnosis and treatment efficiency and ensuring the diagnosis and treatment effect.
In a first aspect, an embodiment of the present invention provides a method for generating a prescription of a chinese medical science, including:
acquiring a target symptom character sequence of a target patient;
inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model;
and determining a target traditional Chinese medicine prescription of the target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model.
In a second aspect, an embodiment of the present invention further provides a device for generating a chinese medical prescription, which is used for generating a chinese medical prescription, and includes:
the symptom character sequence acquisition module is used for acquiring a target symptom character sequence of a target patient;
the traditional Chinese medicine prediction module is used for inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, and the traditional Chinese medicine prediction network model is obtained by carrying out mask masking on a sample traditional Chinese medicine character sequence and carrying out fine adjustment on a pre-training model;
and the target traditional Chinese medicine prescription determining module is used for determining a target traditional Chinese medicine prescription of a target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model.
In a third aspect, the present invention also provides an electronic device, including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to enable the at least one processor to perform the method of generating a prescription for chinese medicine of any of the embodiments of the present invention.
According to another aspect of the present invention, there is provided a computer-readable storage medium storing computer instructions for causing a processor to implement the method for generating a prescription of chinese medicine according to any one of the embodiments of the present invention when executed.
According to the technical scheme of the embodiment of the invention, a target symptom character sequence of a target patient is obtained; inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model; according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model, the target traditional Chinese medicine prescription of the target patient is determined, so that a more accurate traditional Chinese medicine prescription can be automatically generated by using the traditional Chinese medicine prediction network model obtained by fine tuning the pre-training model, and in the process of allocating the prescription for the patient by traditional Chinese medicine, a more scientific prescription reference is provided for traditional Chinese medicine, so that the diagnosis and treatment efficiency is improved, and the diagnosis and treatment effect is ensured.
It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present invention, nor do they necessarily limit the scope of the invention. Other features of the present invention will become apparent from the following description.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments 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 to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a flowchart of a method for generating a prescription of chinese medicine according to an embodiment of the present invention;
FIG. 2 is a flowchart of a method for obtaining a Chinese medicine prediction network model according to a second embodiment of the present invention;
fig. 3 is a schematic structural diagram of a chinese medical prescription generating apparatus according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device implementing the method for generating a prescription of chinese medicine according to the embodiment of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solutions of the present invention, 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 only a part of the embodiments of the present invention, and not all of the embodiments. 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.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in other sequences than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example one
Fig. 1 is a flowchart of a method for generating a chinese medical prescription according to an embodiment of the present invention, where the method is applicable to a situation where a corresponding chinese medical prescription is generated according to patient symptom information, and the method can be executed by a chinese medical prescription generating apparatus, which can be implemented in hardware and/or software, and the chinese medical prescription generating apparatus can be configured in an electronic device.
As shown in fig. 1, the method includes:
and S110, acquiring a target symptom character sequence of the target patient.
Wherein, the target symptom character sequence refers to a character sequence composed of related symptoms, health information, examination and examination information, etc. of the target patient, such as [ waist, space, disk, protrusion, nutrition, poor quality, high, blood, pressure ].
Specifically, the data such as basic information, symptom information, inspection and examination information of the target patient are acquired from the electronic medical record of traditional Chinese medicine and/or the visit record of the target patient, the data are subjected to duplication removal and cleaning treatment, and terms with the same meaning are subjected to unification treatment to obtain a target symptom character sequence corresponding to the target patient.
And S120, inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by finely adjusting a pre-training model in a mode of masking the sample traditional Chinese medicine character sequence.
The traditional Chinese medicine prediction network model is a network model which is obtained by pre-training and can predict corresponding traditional Chinese medicine character sequences according to the target symptom character sequences of the target patients. The sample Chinese medicine character sequence refers to a character sequence for training Chinese medicine names of a sample Chinese medicine prediction network model, such as [ Lian, qiao, chen, pi, feng, mi ]. The Mask is to replace partial characters in the Chinese medicine character sequence with Mask, such as [ Lian, wo, mask, mi ]. The pre-training model is obtained by training a model through large-scale corpora. Further, training may be resumed or otherwise used based on the pre-trained model. Fine tuning refers to the use of pre-trained models to custom train certain tasks, such as the following predictions from the above. Specifically, mask fine adjustment is performed on the pre-training model in advance, so that the task of predicting the corresponding Chinese medicine character sequence according to the target symptom character sequence of the target patient can be completed. Inputting the target symptom character sequence into the traditional Chinese medicine prediction network model, and outputting a corresponding traditional Chinese medicine character sequence with the maximum probability.
And S130, determining a target traditional Chinese medicine prescription of the target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model.
The target Chinese medicine prescription is obtained by converting a target Chinese medicine character sequence output by the Chinese medicine prediction network model.
Specifically, the Chinese medicine prediction network model predicts a Chinese medicine character sequence with the maximum probability corresponding to the patient symptom character sequence through data processing according to the input patient symptom character sequence, removes useless symbols from the Chinese medicine character sequence to obtain a target Chinese medicine prescription in a text format, and displays the obtained target Chinese medicine prescription to provide prescription reference for Chinese medicine.
According to the technical scheme of the embodiment of the invention, a target symptom character sequence of a target patient is obtained; inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model; according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model, the target traditional Chinese medicine prescription of the target patient is determined, so that the traditional Chinese medicine prediction network model obtained by fine tuning through the pre-training model can automatically generate a more accurate traditional Chinese medicine prescription, and in the process of allocating the prescription for the patient by traditional Chinese medicine, a more scientific prescription reference is provided for the traditional Chinese medicine, the diagnosis and treatment efficiency is improved, and the diagnosis and treatment effect is ensured.
On the basis of the technical scheme, the pre-training model is finely adjusted by performing a mask mode on the sample traditional Chinese medicine character sequence to obtain a traditional Chinese medicine prediction network model, and the method can comprise the following steps of S101-S105:
s101, obtaining a sample symptom character sequence and a sample traditional Chinese medicine character sequence corresponding to a sample patient.
Specifically, assume a sample symptom character sequence X = (X1, X2, …, xm) for a patient; sample chinese character sequence Y = (Y1, Y2, …, yn) as one sample. Wherein xi is the ith character in the sample symptom character sequence, m is the number of characters in the sample symptom character sequence of the patient, yj is the jth character in the sample traditional Chinese medicine character sequence, and n is the number of characters in the sample traditional Chinese medicine character sequence. Further, for better model training, as many samples as possible may be obtained for model training.
Illustratively, a sample symptom character sequence [ hair, heat, nose, plug, throat, pain ] of a patient and a corresponding sample Chinese medicine character sequence [ gold, silver, flower, silver, perk, thin, lotus ] are taken as a set of samples.
And S102, carrying out random mask on the traditional Chinese medicine characters in the sample traditional Chinese medicine character sequence based on a preset mask character proportion to obtain a masked traditional Chinese medicine character sequence.
Wherein, the mask character proportion refers to the proportion of the masked character in the Chinese medicine character sequence of the sample. Further, the mask character ratio may be set according to actual situations, and the embodiment is not limited herein.
Specifically, a sample chinese medicine character sequence Y = [ Y1, Y2, Y3, Y4, Y5, Y6] is randomly masked to obtain a masked chinese medicine character sequence Y1= [ Y1, Y2, Y3, mask, Y6].
On the basis of the above example, a sample Chinese medicine character sequence [ gold, silver, flower, silver, warp, thin and lotus ] is randomly masked to obtain a Mask Chinese medicine character sequence [ gold, silver, flower, mask and lotus ]
S103, splicing the sample symptom character sequence and the mask traditional Chinese medicine character sequence to obtain a sample splicing sequence.
On the basis of the above example, the sample symptom character sequence and the Mask Chinese medicine character sequence are spliced to obtain [ hair, heat, nose, plug, pharynx, larynx, pain, gold, silver, flower, mask, lotus ]
Optionally, the sequence start character, the sample symptom character sequence, the sequence segmentation character, the mask Chinese medicine character sequence and the sequence end character are spliced in sequence to obtain a sample splicing sequence.
Wherein, the start sequence character is 'SOS' to indicate the start of the sequence, and the sequence segmentation character and the sequence end character are 'EOS' to indicate the segmentation of the sample symptom character sequence and the mask Chinese medicine character sequence and the end of the sequence.
Specifically, the start sequence character is "SOS", the sample symptom character sequence X of the patient, the sequence segmentation character "EOS", the sample chinese medicine character sequence Y and the sequence end character "EOS" are concatenated to obtain a sample concatenated sequence [ SOS, X, EOS, Y, EOS ].
On the basis of the above example, the sample splicing sequence is obtained by splicing according to the sequence starting character, the sample symptom character sequence, the sequence segmentation character, the mask Chinese medicine character sequence and the sequence ending character: [ SOS, hair, heat, nose, plug, throat, pain, EOS, gold, silver, flower, mask, lotus, EOS ].
S104, inputting the sample splicing sequence into a pre-training model to predict masked Chinese medicine characters, and obtaining predicted Chinese medicine characters based on the output of the pre-training model.
The Chinese medicine character prediction is the masked Chinese medicine character obtained by performing the following prediction through semantic analysis of a sample symptom character sequence by a pre-training model.
Specifically, the sample splicing sequence is input into a pre-training model, the pre-training model outputs predicted Chinese medicine characters, and further, certain difference may exist between the predicted Chinese medicine characters and actual masked characters.
S105, determining a training error based on the predicted traditional Chinese medicine characters and the actual traditional Chinese medicine characters which are masked, reversely transmitting the training error to the pre-training model, adjusting model parameters in the pre-training model until fine tuning is finished when a preset convergence condition is reached, and obtaining a traditional Chinese medicine prediction network model.
Specifically, each sample symptom character sequence is input into a pre-training model, corresponding predicted traditional Chinese medicine characters are obtained based on the output of the pre-training model, training errors are determined according to the predicted traditional Chinese medicine characters and actual traditional Chinese medicine characters covered by masks based on a loss function, and the training errors can be understood as loss values which play a role in correcting parameters of the pre-training model. And reversely transmitting the training error to the pre-training model, and adjusting the model parameters in the pre-training model until a preset convergence condition is reached, for example, when the iteration times are equal to the preset times or the variation of the training error tends to be stable, determining that the fine tuning of the pre-training model is finished.
Wherein the loss function may be, but is not limited to, a cross-entropy loss function L, which may be represented as follows:
wherein X is a sample symptom character sequence, y j Is the jth Chinese medicine character, and the output is y when p is the Chinese medicine character before the input is X and the jth Chinese medicine character j N is the number of characters of the predicted traditional Chinese medicine.
Example two
Fig. 2 is a flowchart of a method for obtaining a traditional Chinese medicine prediction network model according to a second embodiment of the present invention, where on the basis of the second embodiment, the pre-training model in this embodiment may include: the process of obtaining the traditional Chinese medicine prediction network model is described in detail by carrying out fine adjustment on the pre-training model through an input layer, an encoding layer and a decoding layer. The technical terms that are the same as or corresponding to the above embodiments are not repeated herein.
As shown in fig. 2, the method specifically includes the following steps:
s210, obtaining a sample symptom character sequence and a sample traditional Chinese medicine character sequence corresponding to the sample patient.
S220, carrying out random mask on the traditional Chinese medicine characters in the sample traditional Chinese medicine character sequence based on a preset mask character proportion to obtain a masked traditional Chinese medicine character sequence.
And S230, splicing the sample symptom character sequence and the mask traditional Chinese medicine character sequence to obtain a sample splicing sequence.
S240, inputting the sample splicing sequence into an input layer in a pre-training model, vectorizing each character in the sample splicing sequence in the input layer, determining a character vector corresponding to each character, and obtaining a character vector sequence.
The input layer is a structure in which the neural network model receives external information data and maps the external information data into corresponding vectors. Vectorization is the process of mapping a sample-spliced sequence into a vector form. The character vector is a vector representing characters, and the vector sequence is obtained by sequencing and splicing vectors corresponding to the character sequence.
Specifically, the sample splicing sequence is input to the input layer, each character in the sample splicing sequence is mapped into a vector, and the character vectors corresponding to each character are sequenced and spliced to obtain a character vector sequence.
Optionally, the "vectorizing each character in the sample concatenation sequence and determining a character vector corresponding to each character" in S240 may include: determining a word vector, a position vector and a segment vector corresponding to each character in the sample splicing sequence; adding the word vector, the position vector and the segment vector corresponding to each character according to the position to obtain a character vector corresponding to each character; wherein, the segment vector corresponding to the symptom character is zero vector, and the segment vector corresponding to the Chinese medicine character is unit vector.
The position vector refers to a vector capable of representing the position of each character in the sample concatenation sequence, and the model is made aware that the input of the character sequence has a time attribute. A segment vector refers to a vector representation that helps the model distinguish between different input sequences.
Specifically, the sample stitching sequence S = [ SOS, S1, EOS, S2, EOS =]Wherein S1 is the sample symptom character sequence [ x1, x2, …, xm]S2 is a sample Chinese medicine character sequence [ y1, y2, …, yn]. Inputting the sample splicing sequence S into an input layer, and mapping the sample splicing sequence S into a character vectorSequence H 0 Then, then
Wherein,for each character corresponding vector representation, s n For samples, splicing individual characters in the sequence S, E emb For a word vector of each character, E pos (s n ) For each character's position vector, E seg (s n ) A segment vector for each character. Further, the segment vectors corresponding to all symptom characters are set to be 0, and the segment vectors corresponding to the Chinese medicine characters are set to be 1, so as to distinguish the two vectors.
And S250, inputting the character vector sequence into an encoding layer in a pre-training model, extracting features in the encoding layer based on a multi-head attention mechanism, and determining a target hidden state corresponding to each masked Chinese medicine character.
The coding layer is a structure in the translator responsible for mapping the character vector sequence into a hidden state. Furthermore, the multi-head attention mechanism can be understood as a vector matrix for performing weight distribution on the front and rear character vectors when the Chinese medicine characters to be masked are predicted, and distributing different degrees of attention to the front and rear information to participate in the prediction of the Chinese medicine characters to be masked. The target hidden state is a hidden state of a final character vector sequence obtained after mapping of a plurality of coding layers. Hidden state can be understood as multiple mappings through the coding layer in the model, and is called hidden state since the user cannot see the data.
Specifically, a character vector sequence H 0 Inputting the data into the coding layers of the pre-training model, and obtaining H through characteristic extraction of each coding layer 0 Hidden state H corresponding to each layer of coding layer l :
H l =Transformer l (H l-1 ),l∈[1,L]
Where L denotes the L-th coding layer and L denotes the last coding layer.
Optionally, the coding layer may include multiple translator coding layers. Correspondingly, in S250, "performing feature extraction based on a multi-head attention mechanism in the coding layer, and determining a target hidden state corresponding to each masked chinese medicine character" may include: in each translator coding layer, coding a hidden state corresponding to each character output by the last translator coding layer based on the mask attention matrix to obtain a current hidden state output by the current translator coding layer; and obtaining a target hidden state corresponding to each Chinese medicine character which is masked based on the output of the last translator encoding layer.
Wherein, the target hidden state refers to the target hidden state corresponding to each Chinese medicine character output by the last layer of coding layerThe mask attention matrix M is a matrix obtained by masking the attention head. Further, the multi-head attention mechanism and the mask attention matrix M may be expressed as:
Q=H l-1 W l Q ,K=H l-1 W l K ,V=H l-1 W l V
wherein A is l For attention head, it is understood that the attention matrix QK is first found T And then weighted by V with the attention matrix,the aim is to change the attention moment array into a standard normal distribution, so that the result after the softmax normalization is more stable, and an equilibrium gradient is obtained during back propagation. Q, K, V, it is understood that this is 3 different linear transformations to represent 3 different states for the same input.The matrix that has been trained in the model is pre-trained. The core process of the self-attention mechanism is to obtain attention weight through Q and K calculation, and then apply the attention weight to V to obtain the whole weight and output.
Further, the portion of the mask attention matrix M representing the sample symptom character sequence is set to 0, indicating that the portion on the right of the predicted chinese medicine character sequence is set to- ∞, so that all sample symptom characters can be subjected to self-attention calculation, while the portion on the right of the predicted chinese medicine character does not participate in attention calculation.
Specifically, when the current translator coding layer outputs the current hidden state corresponding to the sample symptom character, the current translator coding layer refers to the semantic information of the hidden states of all characters of the sample symptom output by the previous translator coding layer, and only refers to the hidden state corresponding to the current Chinese medicine character output by the previous translator coding layer and the semantic information of the hidden states of all characters on the left of the current Chinese medicine character in the current hidden state of the Chinese medicine character output by the previous translator coding layer. Finally, outputting a target hidden state corresponding to each Chinese medicine character which is masked on the last translator encoding layer.
S260, inputting the hidden state corresponding to each Chinese medicine character to be masked into a decoding layer, and determining the predicted Chinese medicine character corresponding to each Chinese medicine character to be masked in the decoding layer based on the hidden state corresponding to each Chinese medicine character to be masked.
The decoding layer maps the target hidden state into a structure for predicting the Chinese medicine characters.
Specifically, the hidden state corresponding to each Chinese medicine character to be masked is input to the decoding layer, and the predicted Chinese medicine character corresponding to each Chinese medicine character to be masked is obtained through feature extraction and classification of the target hidden state by the decoding layer.
Optionally, the decoding layer may include: a full link layer and an active layer. In S260, "determining, in the decoding layer, the predicted chinese medicine character corresponding to each masked chinese medicine character based on the hidden state corresponding to each masked chinese medicine character" may include inputting the hidden state corresponding to each masked chinese medicine character to the full-connection layer for full-connection processing, and obtaining full-connection information corresponding to each masked chinese medicine character; inputting the full-connection information corresponding to each masked Chinese medicine character into the activation layer for normalization, determining the probability value of each masked Chinese medicine character being each Chinese medicine character in the Chinese medicine dictionary, and determining the Chinese medicine character with the highest probability value as the predicted Chinese medicine character.
The full connection layer is a structure for extracting features of the hidden state corresponding to each Chinese medicine character which is masked, and dimension reduction processing can be performed on the hidden state corresponding to each Chinese medicine character. The active layer is to map the input of the fully-connected layer to a probability value output using an activation function. The formula is as follows:
where P represents the probability distribution matrix of each masked character in the chinese prescription dictionary.
Specifically, inputting the hidden state corresponding to each Chinese medicine character to be masked into the full-connection layer for feature extraction, and obtaining a feature vector corresponding to each Chinese medicine character to be masked; inputting the feature vector corresponding to each masked Chinese medicine character into an activation layer for normalization, determining the probability value of each masked Chinese medicine character being each Chinese medicine character in a Chinese medicine dictionary, and selecting the word in the Chinese medicine prescription dictionary corresponding to the maximum probability value as the prediction result of the current masked Chinese medicine character for each masked Chinese medicine character.
And S270, obtaining the predicted Chinese medicine characters based on the output of the pre-training model.
S280, determining a training error based on the predicted traditional Chinese medicine characters and the actual traditional Chinese medicine characters which are masked, reversely transmitting the training error to a pre-training model, adjusting model parameters in the pre-training model until fine tuning is finished when a preset convergence condition is reached, and obtaining a traditional Chinese medicine prediction network model.
Optionally, two indexes of Recall (Recall) and Precision (Precision) are selected to evaluate the traditional Chinese medicine prediction network model:
wherein R (X) is the model-generated prescription after entry of symptom sequence X, and T (X) is the patient's original prescription with symptoms X. Recall (Recall) is the proportion of the number of the Chinese medicines overlapping the prescription generated by the model and the prescription actually used by the patient to the total number of the Chinese medicines contained in the prescription actually used by the patient. The higher the recall rate, the more the recipe is generated to cover the actual medication, and the better the recipe is generated.
The accuracy rate is the proportion of the number of medicines of the prescription generated by the model and the prescription actually used by the patient to the total number of medicines of the prescription generated by the model. The higher the accuracy rate, the more accurate the recipe is generated.
According to the technical scheme of the embodiment of the invention, a sample symptom character sequence and a sample traditional Chinese medicine character sequence corresponding to a sample patient are obtained; based on a preset mask character proportion, carrying out random mask on traditional Chinese medicine characters in the sample traditional Chinese medicine character sequence to obtain a mask traditional Chinese medicine character sequence after mask; splicing the sample symptom character sequence and the mask traditional Chinese medicine character sequence to obtain a sample splicing sequence; inputting the sample splicing sequence into an input layer, vectorizing each character in the sample splicing sequence in the input layer, determining a character vector corresponding to each character, and obtaining a character vector sequence; inputting the character vector sequence into an encoding layer, extracting features in the encoding layer based on a multi-head attention mechanism, and determining a target hidden state corresponding to each Chinese medicine character to be masked; inputting the hidden state corresponding to each Chinese medicine character to be masked into a decoding layer, and determining a predicted Chinese medicine character corresponding to each Chinese medicine character to be masked in the decoding layer based on the hidden state corresponding to each Chinese medicine character to be masked; obtaining a predicted Chinese medicine character based on the output of the pre-training model; determining a training error based on the predicted traditional Chinese medicine characters and the actual traditional Chinese medicine characters which are masked, reversely transmitting the training error to a pre-training model, adjusting model parameters in the pre-training model until fine adjustment is finished when a preset convergence condition is reached, obtaining a traditional Chinese medicine prediction network model, performing mask fine adjustment processing on the basis of the pre-training model, obtaining the capacity of predicting a traditional Chinese medicine prescription, and providing a reference basis for the traditional Chinese medicine to determine the prescription according to the symptom information of a patient.
EXAMPLE III
Fig. 3 is a schematic structural diagram of a device for generating a prescription of chinese medicine according to a third embodiment of the present invention. As shown in fig. 3, the apparatus includes: a symptom character sequence acquisition module 310, a traditional Chinese medicine prediction module 320 and a target traditional Chinese medicine prescription determination module 330.
Wherein, the symptom character sequence acquiring module 310 is configured to acquire a target symptom character sequence of a target patient; the traditional Chinese medicine prediction module 320 is used for inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, and the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model; and the target traditional Chinese medicine prescription determining module 330 is used for determining the target traditional Chinese medicine prescription of the target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model.
On the basis of the above technical solutions, the Chinese medicine prediction module 320 further comprises,
the character sequence acquisition unit is used for acquiring a sample symptom character sequence and a sample traditional Chinese medicine character sequence corresponding to a sample patient;
the traditional Chinese medicine character mask unit is used for carrying out random mask on traditional Chinese medicine characters in the sample traditional Chinese medicine character sequence based on a preset mask character proportion to obtain a mask traditional Chinese medicine character sequence after mask;
the sequence splicing unit is used for splicing the sample symptom character sequence and the mask traditional Chinese medicine character sequence to obtain a sample splicing sequence;
the predicted traditional Chinese medicine character acquisition unit is used for inputting the sample splicing sequence into a pre-training model to predict masked traditional Chinese medicine characters and acquiring predicted traditional Chinese medicine characters based on the output of the pre-training model;
and the model parameter adjusting unit is used for determining a training error based on the predicted traditional Chinese medicine characters and the actual traditional Chinese medicine characters subjected to mask, reversely transmitting the training error to the pre-training model, adjusting model parameters in the pre-training model until fine adjustment is finished when a preset convergence condition is reached, and obtaining the traditional Chinese medicine prediction network model.
On the basis of the technical schemes, the sequence splicing unit is specifically used for,
and splicing according to the sequence starting character, the sample symptom character sequence, the sequence segmentation character, the mask Chinese medicine character sequence and the sequence ending character to obtain a sample splicing sequence.
On the basis of the technical schemes, the unit for acquiring the predicted Chinese medicine characters specifically comprises,
the character vectorization subunit is used for inputting the sample splicing sequence into the input layer, vectorizing each character in the sample splicing sequence in the input layer, determining a character vector corresponding to each character, and obtaining a character vector sequence;
the mask character target hidden state determining subunit is used for inputting the character vector sequence into the coding layer, performing feature extraction based on a multi-head attention mechanism in the coding layer, and determining a target hidden state corresponding to each Chinese medicine character to be masked;
the predicted traditional Chinese medicine character determining subunit inputs the hidden state corresponding to each traditional Chinese medicine character to be masked to the decoding layer, and determines the predicted traditional Chinese medicine character corresponding to each traditional Chinese medicine character to be masked in the decoding layer based on the hidden state corresponding to each traditional Chinese medicine character to be masked.
On the basis of the technical schemes, the character vectorization subunit is specifically used for,
determining a word vector, a position vector and a segment vector corresponding to each character in the sample splicing sequence; adding the word vector, the position vector and the segment vector corresponding to each character according to the position to obtain a character vector corresponding to each character; wherein, the segment vector corresponding to the symptom character is zero vector, and the segment vector corresponding to the Chinese medicine character is unit vector.
On the basis of the technical schemes, the mask character target hidden state determining subunit is specifically used for,
in each translator coding layer, coding a hidden state corresponding to each character output by the last translator coding layer based on the mask attention matrix to obtain a current hidden state output by the current translator coding layer; and obtaining a target hidden state corresponding to each Chinese medicine character which is masked based on the output of the last translator encoding layer.
On the basis of the technical schemes, the Chinese medicine character prediction determination subunit is specifically used for,
inputting the hidden state corresponding to each Chinese medicine character to be masked to a full-connection layer for full-connection processing to obtain full-connection information corresponding to each Chinese medicine character to be masked; inputting the full-connection information corresponding to each masked Chinese medicine character into the activation layer for normalization, determining the probability value of each masked Chinese medicine character being each Chinese medicine character in the Chinese medicine dictionary, and determining the Chinese medicine character with the highest probability value as the predicted Chinese medicine character.
According to the technical scheme of the embodiment of the invention, a target symptom character sequence of a target patient is obtained; inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model; according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model, the target traditional Chinese medicine prescription of the target patient is determined, so that the more accurate traditional Chinese medicine prescription can be automatically generated by the traditional Chinese medicine prediction network model obtained by fine tuning through the pre-training model, a more scientific prescription reference is provided for traditional Chinese medicine in the process of allocating the prescription for the patient by the traditional Chinese medicine, the diagnosis and treatment efficiency is improved, and the diagnosis and treatment effect is ensured.
The traditional Chinese medicine prescription generating device provided by the embodiment of the invention can execute the traditional Chinese medicine prescription generating method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the executing method.
Example four
FIG. 4 shows a schematic block diagram of an electronic device 10 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 4, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, and the like, wherein the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various suitable actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data necessary for the operation of the electronic apparatus 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
A number of components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, or the like; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and so forth. The processor 11 performs the various methods and processes described above, such as the chinese medical prescription generation method.
In some embodiments, the chinese medical prescription generation method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the above-described chinese medical prescription generating method may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the chinese medical prescription generation method by any other suitable means (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. A computer program can execute entirely on a machine, partly on a machine, as a stand-alone software package partly on a machine and partly on a remote machine or entirely on a remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service are overcome.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present invention may be executed in parallel, sequentially, or in different orders, and are not limited herein as long as the desired results of the technical solution of the present invention can be achieved.
The above-described embodiments should not be construed as limiting the scope of the invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A traditional Chinese medicine prescription generating method is characterized by comprising the following steps:
acquiring a target symptom character sequence of a target patient;
inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, wherein the traditional Chinese medicine prediction network model is obtained by performing mask masking on a sample traditional Chinese medicine character sequence and performing fine adjustment on a pre-training model;
and determining a target traditional Chinese medicine prescription of the target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model.
2. The method of claim 1, wherein the fine-tuning of the pre-training model by masking the sample chinese medicine character sequence to obtain the chinese medicine prediction network model comprises:
acquiring a sample symptom character sequence and a sample traditional Chinese medicine character sequence corresponding to a sample patient;
based on a preset mask character proportion, carrying out random mask on traditional Chinese medicine characters in the sample traditional Chinese medicine character sequence to obtain a masked traditional Chinese medicine character sequence;
splicing the sample symptom character sequence and the mask traditional Chinese medicine character sequence to obtain a sample splicing sequence;
inputting the sample splicing sequence into a pre-training model to predict masked Chinese medicine characters, and obtaining predicted Chinese medicine characters based on the output of the pre-training model;
and determining a training error based on the predicted traditional Chinese medicine characters and the actual traditional Chinese medicine characters subjected to mask, reversely transmitting the training error to a pre-training model, adjusting model parameters in the pre-training model until fine tuning is finished when a preset convergence condition is reached, and obtaining a traditional Chinese medicine prediction network model.
3. The method of claim 2, wherein the concatenating the sample symptom character sequence and the mask chinese medicine character sequence to determine a sample concatenation sequence comprises:
and splicing according to the sequence starting character, the sample symptom character sequence, the sequence segmentation character, the mask Chinese medicine character sequence and the sequence ending character to obtain a sample splicing sequence.
4. The method of claim 2, wherein the pre-training model comprises: an input layer, an encoding layer, and a decoding layer;
inputting the sample splicing sequence into a pre-training model to predict masked Chinese medicine characters, wherein the method comprises the following steps:
inputting the sample splicing sequence into the input layer, vectorizing each character in the sample splicing sequence in the input layer, determining a character vector corresponding to each character, and obtaining a character vector sequence;
inputting the character vector sequence into the coding layer, performing feature extraction in the coding layer based on a multi-head attention mechanism, and determining a target hidden state corresponding to each Chinese medicine character subjected to mask masking;
inputting the hidden state corresponding to each Chinese medicine character to be masked to the decoding layer, and determining the predicted Chinese medicine character corresponding to each Chinese medicine character to be masked in the decoding layer based on the hidden state corresponding to each Chinese medicine character to be masked.
5. The method of claim 4, wherein vectorizing each character in the sample stitching sequence and determining a character vector corresponding to each character comprises:
determining a word vector, a position vector and a segment vector corresponding to each character in the sample splicing sequence;
adding the word vector, the position vector and the segment vector corresponding to each character according to the position to obtain a character vector corresponding to each character;
wherein, the segment vector corresponding to the symptom character is zero vector, and the segment vector corresponding to the Chinese medicine character is unit vector.
6. The method of claim 4, wherein the coding layer comprises a plurality of translator coding layers;
the method comprises the following steps of performing feature extraction in the coding layer based on a multi-head attention mechanism, and determining a target hidden state corresponding to each masked Chinese medicine character, wherein the step of:
in each translator coding layer, coding a hidden state corresponding to each character output by the last translator coding layer based on the mask attention matrix to obtain a current hidden state output by the current translator coding layer;
and obtaining a target hidden state corresponding to each Chinese medicine character which is masked based on the output of the last translator encoding layer.
7. The method of claim 4, wherein the decoding layer comprises: a fully-connected layer and an active layer;
determining a predicted traditional Chinese medicine character corresponding to each masked traditional Chinese medicine character based on a hidden state corresponding to each masked traditional Chinese medicine character in a decoding layer, wherein the method comprises the following steps:
inputting the hidden state corresponding to each Chinese medicine character to be masked to the full-connection layer for full-connection processing, and obtaining full-connection information corresponding to each Chinese medicine character to be masked;
inputting the full-connection information corresponding to each Chinese medicine character to be masked into the activation layer for normalization, determining the probability value of each Chinese medicine character to be masked as each Chinese medicine character in the Chinese medicine dictionary, and determining the Chinese medicine character with the highest probability value as a predicted Chinese medicine character.
8. A chinese medical prescription generating apparatus, comprising:
the symptom character sequence acquisition module is used for acquiring a target symptom character sequence of a target patient;
the traditional Chinese medicine prediction module is used for inputting the target symptom character sequence into a traditional Chinese medicine prediction network model for traditional Chinese medicine prediction, and the traditional Chinese medicine prediction network model is obtained by carrying out fine adjustment on a pre-training model in a mask mode on a sample traditional Chinese medicine character sequence;
and the target traditional Chinese medicine prescription determining module is used for determining the target traditional Chinese medicine prescription of the target patient according to the target traditional Chinese medicine character sequence output by the traditional Chinese medicine prediction network model.
9. An electronic device, characterized in that the electronic device comprises:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of generating a chinese medical prescription according to any one of claims 1 to 7.
10. A computer-readable storage medium storing computer instructions for causing a processor to perform the method of generating a prescription of chinese medical science according to any one of claims 1 to 7 when executed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211435070.0A CN115719628B (en) | 2022-11-16 | 2022-11-16 | Method, device, equipment and storage medium for generating traditional Chinese medicine prescription |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211435070.0A CN115719628B (en) | 2022-11-16 | 2022-11-16 | Method, device, equipment and storage medium for generating traditional Chinese medicine prescription |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115719628A true CN115719628A (en) | 2023-02-28 |
CN115719628B CN115719628B (en) | 2024-07-23 |
Family
ID=85256764
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211435070.0A Active CN115719628B (en) | 2022-11-16 | 2022-11-16 | Method, device, equipment and storage medium for generating traditional Chinese medicine prescription |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115719628B (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111312356A (en) * | 2020-01-17 | 2020-06-19 | 四川大学 | Traditional Chinese medicine prescription generation method based on BERT and integration efficacy information |
CN111897955A (en) * | 2020-07-13 | 2020-11-06 | 广州视源电子科技股份有限公司 | Comment generation method, device and equipment based on coding and decoding and storage medium |
CN113239705A (en) * | 2021-07-12 | 2021-08-10 | 北京百度网讯科技有限公司 | Pre-training method and device of semantic representation model, electronic equipment and storage medium |
JP2021152963A (en) * | 2020-12-25 | 2021-09-30 | 北京百度網訊科技有限公司 | Word meaning feature generating method, model training method, apparatus, device, medium, and program |
CN114300122A (en) * | 2021-12-17 | 2022-04-08 | 中国农业大学 | Crop disease diagnosis method and system based on electronic medical record |
CN114550870A (en) * | 2022-03-23 | 2022-05-27 | 康键信息技术(深圳)有限公司 | Prescription auditing method, device, equipment and medium based on artificial intelligence |
CN114783618A (en) * | 2022-05-17 | 2022-07-22 | 平安科技(深圳)有限公司 | Traditional Chinese medicine model training method and device based on knowledge structure and four-diagnosis character |
CN114783571A (en) * | 2022-04-06 | 2022-07-22 | 北京交通大学 | Traditional Chinese medicine dynamic diagnosis and treatment scheme optimization method and system based on deep reinforcement learning |
CN115310446A (en) * | 2022-08-03 | 2022-11-08 | 湖南中医药大学 | Traditional Chinese medicine ancient book named entity identification method and device, electronic equipment and memory |
-
2022
- 2022-11-16 CN CN202211435070.0A patent/CN115719628B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111312356A (en) * | 2020-01-17 | 2020-06-19 | 四川大学 | Traditional Chinese medicine prescription generation method based on BERT and integration efficacy information |
CN111897955A (en) * | 2020-07-13 | 2020-11-06 | 广州视源电子科技股份有限公司 | Comment generation method, device and equipment based on coding and decoding and storage medium |
JP2021152963A (en) * | 2020-12-25 | 2021-09-30 | 北京百度網訊科技有限公司 | Word meaning feature generating method, model training method, apparatus, device, medium, and program |
CN113239705A (en) * | 2021-07-12 | 2021-08-10 | 北京百度网讯科技有限公司 | Pre-training method and device of semantic representation model, electronic equipment and storage medium |
CN114300122A (en) * | 2021-12-17 | 2022-04-08 | 中国农业大学 | Crop disease diagnosis method and system based on electronic medical record |
CN114550870A (en) * | 2022-03-23 | 2022-05-27 | 康键信息技术(深圳)有限公司 | Prescription auditing method, device, equipment and medium based on artificial intelligence |
CN114783571A (en) * | 2022-04-06 | 2022-07-22 | 北京交通大学 | Traditional Chinese medicine dynamic diagnosis and treatment scheme optimization method and system based on deep reinforcement learning |
CN114783618A (en) * | 2022-05-17 | 2022-07-22 | 平安科技(深圳)有限公司 | Traditional Chinese medicine model training method and device based on knowledge structure and four-diagnosis character |
CN115310446A (en) * | 2022-08-03 | 2022-11-08 | 湖南中医药大学 | Traditional Chinese medicine ancient book named entity identification method and device, electronic equipment and memory |
Also Published As
Publication number | Publication date |
---|---|
CN115719628B (en) | 2024-07-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110427486B (en) | Body condition text classification method, device and equipment | |
CN111681726A (en) | Method, device, equipment and medium for processing electronic medical record data | |
JP7457125B2 (en) | Translation methods, devices, electronic equipment and computer programs | |
JPWO2019187372A1 (en) | Prediction system, model generation system, method and program | |
CN114613523A (en) | Doctor allocation method, device, storage medium and equipment for on-line medical inquiry | |
CN112560985A (en) | Neural network searching method and device and electronic equipment | |
WO2021139432A1 (en) | Artificial intelligence-based user rating prediction method and apparatus, terminal, and medium | |
CN116010586A (en) | Method, device, equipment and storage medium for generating health advice | |
CN113642727B (en) | Training method of neural network model and processing method and device of multimedia information | |
CN112949433B (en) | Method, device and equipment for generating video classification model and storage medium | |
CN113590777A (en) | Text information processing method and device, electronic equipment and storage medium | |
CN115719628B (en) | Method, device, equipment and storage medium for generating traditional Chinese medicine prescription | |
CN116521860A (en) | Medical problem abstract generation method and device based on contrast learning | |
JP7390442B2 (en) | Training method, device, device, storage medium and program for document processing model | |
CN115130470B (en) | Method, device, equipment and medium for generating text keywords | |
CN115577705A (en) | Method, device and equipment for generating text processing model and storage medium | |
CN115665363A (en) | Video conference method, device, equipment and storage medium | |
CN114461085A (en) | Medical input recommendation method, device, equipment and storage medium | |
CN114723045A (en) | Model training method, device, system, apparatus, medium, and program product | |
CN112560467A (en) | Method, device, equipment and medium for determining element relationship in text | |
CN114429787B (en) | Omics data processing method and device, electronic device and storage medium | |
CN114418063B (en) | Method and device for distributing network layer in neural network model | |
CN115331811A (en) | Disease type prediction method and device, electronic equipment and storage medium | |
CN117933385A (en) | Text description generation method and device of graph data, storage medium and electronic equipment | |
CN118230971A (en) | Medical information extraction method, device, electronic equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant |