CN114840684A - Map construction method, device and equipment based on medical entity and storage medium - Google Patents
Map construction method, device and equipment based on medical entity and storage medium Download PDFInfo
- Publication number
- CN114840684A CN114840684A CN202210439849.3A CN202210439849A CN114840684A CN 114840684 A CN114840684 A CN 114840684A CN 202210439849 A CN202210439849 A CN 202210439849A CN 114840684 A CN114840684 A CN 114840684A
- Authority
- CN
- China
- Prior art keywords
- medical
- entity
- medical entity
- model
- speech
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
Abstract
The invention relates to an intelligent decision technology, and discloses a map construction method based on a medical entity, which comprises the following steps: performing semi-supervised learning on a pre-constructed initialized medical entity recognition model by using a labeled training set and a label-free training set to obtain a medical entity recognition model, and performing word segmentation and part-of-speech tagging on a pre-constructed medical information text by using a Chinese word segmentation system to obtain a medical word segmentation set; vectorizing the medical word segmentation set by using a BERT network of a medical entity recognition model to obtain secondary quantization data; and extracting the features of the secondary quantized data to obtain a part-of-speech-semantic feature sequence, performing entity classification judgment on the part-of-speech-semantic feature sequence to obtain an entity set, and constructing a knowledge graph according to the entity set. The invention also provides a map construction device, equipment and a storage medium based on the medical entity. The invention improves the construction efficiency of the medical knowledge map by improving the accuracy of entity identification in the medical industry.
Description
Technical Field
The invention relates to the technical field of intelligent decision, in particular to a map construction method, a map construction device, map construction equipment and a computer-readable storage medium based on a medical entity.
Background
With the development of big data technology, all industries begin to construct knowledge maps of enterprise data so as to create more data values. However, due to the nature of the clinical and medical industries, data entities need to be more clear and accurate, such as "capsule" and "no capsule" entities, and therefore, the traditional named entity identification (NER) method cannot be used.
At present, in order to construct a knowledge graph in the medical industry, entities extracted by Chinese word segmentation system (NLPIR) segmentation can only be further integrated and labeled by an artificial labeling method in the industry to obtain accurate entity information, then the knowledge graph is constructed by the integrated entity information, a large amount of manpower and time are consumed, and with the progress of the medical level, the efficiency of artificial labeling gradually fails to meet the requirement of increasing medical data, so that a method for accurately and efficiently constructing the knowledge graph by using the entity information in the medical industry is urgently needed.
Disclosure of Invention
The invention provides a map construction method, a map construction device, map construction equipment and a storage medium based on medical entities, and mainly aims to improve the construction efficiency of medical knowledge maps by improving the identification accuracy of entities in the medical industry.
In order to achieve the above object, the present invention provides a map building method based on medical entities, comprising:
acquiring a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
acquiring a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, and performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set;
importing the medical word segmentation set into the medical entity recognition model, carrying out vectorization operation on the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a primary quantized data set, and carrying out attention weight calculation operation on the primary quantized data set to obtain a secondary quantized data set;
identifying part-of-speech relations and semantic relations among adjacent secondary quantized data in the secondary quantized data set by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence;
judging whether each secondary quantized data in the secondary quantized data set is an entity or a combined entity or not according to the part of speech-semantic feature sequence by utilizing an output layer network in the medical entity identification model, and outputting the entity and the combined entity according to a judgment result to obtain an entity set;
and according to a preset map construction rule, map construction is carried out on the entity set to obtain the medical knowledge map.
Optionally, the training the initialized medical entity recognition model by using the labeled training set and the unlabeled training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model includes:
training the initialized medical entity recognition model by using the labeled training set to obtain a primary medical entity recognition model;
performing entity recognition on the label-free training set by using the primary medical entity recognition model to obtain a recognition result set and confidence degrees corresponding to all recognition results in the recognition result set;
extracting the recognition result with the confidence coefficient larger than a preset threshold value in the recognition result set as a pseudo label, and defining a label-free training set corresponding to the pseudo label as a pseudo label training set;
calculating the model error of the primary medical entity recognition model according to the confidence degree corresponding to each recognition result;
judging whether the model error is greater than or equal to a preset standard error or not;
when the model error is larger than or equal to the standard error, judging that the primary medical entity recognition model is unqualified, merging the pseudo-labeled training set into the labeled training set, and returning to the step of training the initialized medical entity recognition model by using the labeled training set;
and when the model error is smaller than the standard error, judging that the primary medical entity identification model is qualified, and setting the primary medical entity identification model as a medical entity identification model.
Optionally, the training the initialized medical entity recognition model by using the labeled training set to obtain a primary medical entity recognition model includes:
sequentially acquiring a labeled training sample from the labeled training set, and performing forward propagation calculation on the labeled training sample by using the initialized medical entity recognition model to obtain a calculation result;
comparing the real label of the labeled training sample with the calculation result to obtain the identification error of the initialized medical entity identification model;
minimizing the identification error to obtain a model parameter of the initialized medical entity identification model when the identification error is minimum;
carrying out reverse updating operation by using the model parameters to obtain an updated medical entity identification model;
judging whether the marked training set has marked training samples or not;
when the marked training sample exists in the marked training set, returning to the step of sequentially acquiring one marked training sample from the marked training set, and performing iterative updating on the updated medical entity model;
and when no labeled training sample exists in the labeled training set, taking the updated medical entity model as a primary medical entity recognition model.
Optionally, the vectorizing operation is performed on the medical word segmentation set by using the BERT network of the medical entity recognition model to obtain a first-level quantization data set, including:
vectorizing the medical word segmentation text in the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a text vector set;
vectorizing the part of speech of each medical participle in the medical participle set to obtain a part of speech vector set;
vectorizing the word sequence of each medical word in the medical word set to obtain a word position vector set;
and correspondingly superposing the text vector set, the word vector set and the word position vector set to obtain the first-level quantitative data set.
Optionally, the recognizing, by using a part-of-speech-semantic recognition network in the medical entity recognition model, part-of-speech relationships and semantic relationships between each two adjacent pieces of secondary quantized data in the secondary quantized data set to obtain a part-of-speech-semantic feature sequence includes:
performing feature extraction on each secondary quantized data in the secondary quantized data set by using a convolution kernel set of the part-of-speech-semantic recognition network to obtain a convolution feature matrix set corresponding to each secondary quantized data;
carrying out average pooling operation on the convolution characteristic matrix set to obtain a pooled characteristic matrix set;
and splitting each pooling feature matrix in the pooling feature matrix set according to a preset recombination rule, and sequentially connecting the splitting results in a one-dimensional manner to obtain a part of speech-semantic feature sequence.
Optionally, the performing atlas construction on the entity set according to a preset atlas construction rule to obtain a medical knowledge atlas includes:
clustering the entity set to obtain a cluster set;
clustering the cluster set into a pre-constructed historical medical knowledge map according to the cluster type of the cluster set;
and according to a preset connection rule, constructing the entity set by using a minimum spanning tree in a cluster to obtain the medical knowledge map.
Optionally, before obtaining the pre-constructed initialized medical entity recognition model, the method further includes:
acquiring a pre-constructed BERT network and a part-of-speech-semantic recognition network containing recognition semantics and an activation function of the part-of-speech;
and connecting the BERT network serving as an input layer and the part-of-speech-semantic recognition network serving as a processing layer and an output layer to obtain an initialized medical entity recognition model.
In order to solve the above problems, the present invention also provides an atlas constructing apparatus based on a medical entity, the apparatus comprising:
the model training module is used for acquiring a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
the information vectorization module is used for obtaining a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set, introducing the medical word segmentation set into the medical entity recognition model, performing vectorization operation on the medical word segmentation set by using a BERT (best effort network) of the medical entity recognition model to obtain a first-level quantitative data set, and performing attention weight calculation operation on the first-level quantitative data set to obtain a second-level quantitative data set;
an entity identification module, configured to identify a part-of-speech relationship and a semantic relationship between each two adjacent sets of secondary quantized data in the set of secondary quantized data by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence, and determine, by using an output layer network in the medical entity identification model, whether each piece of secondary quantized data in the set of secondary quantized data is an entity or a combined entity according to the part-of-speech-semantic feature sequence, and output the entity and the combined entity according to a determination result to obtain an entity set;
and the map construction module is used for carrying out map construction on the entity set according to a preset map construction rule so as to obtain the medical knowledge map.
In order to solve the above problem, the present invention also provides an electronic device, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the medical entity-based atlas construction method described above.
In order to solve the above problem, the present invention further provides a computer-readable storage medium, in which at least one computer program is stored, the at least one computer program being executed by a processor in an electronic device to implement the medical entity-based atlas construction method described above.
According to the embodiment of the invention, through an attention mechanism in a BERT network, each word is quantized into second quantized data containing the characteristics of the word and the characteristics of adjacent words, so that a subsequent part-of-speech-semantic recognition network can judge whether an entity and adjacent entities can be combined or not according to the second quantized data, and the recognition accuracy is improved; in addition, the embodiment of the invention can effectively improve the training effect of the medical entity recognition model when the sample data is insufficient by the semi-supervised learning method with the labeled training set and the unlabeled training set. Therefore, the map construction method, device, equipment and storage medium based on the medical entity provided by the embodiment of the invention can improve the construction efficiency of the medical knowledge map by improving the identification accuracy of the medical industry entity.
Drawings
Fig. 1 is a schematic flow chart of a medical entity-based atlas construction method according to an embodiment of the present invention;
FIG. 2 is a detailed flowchart illustrating a step in a method for constructing a medical entity-based atlas according to an embodiment of the invention;
FIG. 3 is a detailed flowchart illustrating a step in a method for constructing a medical entity-based atlas according to an embodiment of the invention;
FIG. 4 is a detailed flowchart illustrating a step in a method for constructing a medical entity-based atlas according to an embodiment of the invention;
FIG. 5 is a detailed flowchart illustrating a step in a method for constructing an atlas based on medical entities according to an embodiment of the invention;
FIG. 6 is a detailed flowchart illustrating a step in a method for constructing a medical entity-based atlas according to an embodiment of the invention;
FIG. 7 is a functional block diagram of a medical entity-based atlas formation apparatus provided in an embodiment of the invention;
fig. 8 is a schematic structural diagram of an electronic device for implementing the medical entity-based atlas creating method according to an embodiment of the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The embodiment of the application provides a map construction method based on a medical entity. In the embodiment of the present application, the executing subject of the atlas construction method based on a medical entity includes, but is not limited to, at least one of electronic devices such as a server and a terminal that can be configured to execute the method provided in the embodiment of the present application. In other words, the medical entity-based atlas construction method may be performed by software or hardware installed in a terminal device or a server device, and the software may be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like. The server may be an independent server, or may be a cloud server that provides basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a Content Delivery Network (CDN), a big data and artificial intelligence platform, and the like.
Referring to fig. 1, a flowchart of a medical entity-based atlas construction method according to an embodiment of the present invention is shown. In this embodiment, the method for constructing a map based on a medical entity includes steps S1-S6:
s1, obtaining a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model.
In the embodiment of the invention, the initialized medical entity recognition model is a text classification model based on a neural network, and can be used for extracting entities or combined entities in text information.
In detail, the embodiment of the invention obtains a pre-constructed BERT network and a part-of-speech-semantic recognition network comprising recognition semantics and an activation function of the part-of-speech; and connecting the BERT network serving as an input layer and the part-of-speech-semantic recognition network serving as a processing layer and an output layer to obtain an initialized medical entity recognition model. The BERT network comprises a text quantization method and an attention mechanism, and text information can be subjected to vectorization coding.
Because the types of the trial medical entities are widely covered, labeled samples in the database magnitude are still too few in the training process, and the effect of accurately identifying the entities cannot be achieved, the embodiment of the invention utilizes two types of training sets with labels and without labels to carry out semi-supervised learning on the initialized medical entity identification model, and solves the problem of less training samples.
In detail, referring to fig. 2, in the embodiment of the present invention, according to a preset semi-supervised learning strategy, the training of the initialized medical entity recognition model by using the labeled training set and the unlabeled training set to obtain a trained medical entity recognition model includes steps S11 to S17:
s11, training the initialized medical entity recognition model by using the labeled training set to obtain a primary medical entity recognition model;
s12, performing entity recognition on the label-free training set by using the primary medical entity recognition model to obtain a recognition result set and confidence degrees corresponding to the recognition results in the recognition result set;
s13, extracting the recognition result with the confidence coefficient larger than a preset threshold value in the recognition result set as a pseudo label, and defining a label-free training set corresponding to the pseudo label as a pseudo label training set;
s14, calculating the model error of the primary medical entity recognition model according to the confidence corresponding to each recognition result;
s15, judging whether the model error is larger than or equal to a preset standard error or not;
when the model error is greater than or equal to the standard error, S16, judging that the primary medical entity recognition model is unqualified, merging the pseudo-labeled training set into the labeled training set, and returning to the step of training the initialized medical entity recognition model by using the labeled training set;
when the model error is less than the standard error, S17, determining that the primary medical entity identification model is qualified, and setting the primary medical entity identification model as a medical entity identification model.
According to the embodiment of the invention, firstly, the labeled training set is used for carrying out parameter assignment on the initialized medical entity recognition model, and then the unlabeled training set is used for testing the primary medical entity recognition model; in the embodiment of the invention, the magnitude of the recognition result with high confidence and low confidence is recorded every time the medical entity recognition model is trained, so that the model error is obtained, and when the model error is smaller than the standard error, the confidence of each recognition result is higher, so that the training process is completed, and the medical entity recognition model is obtained.
Further, referring to fig. 3, in the embodiment of the present invention, the training the initialized medical entity recognition model by using the labeled training set to obtain a primary medical entity recognition model includes S111-S116:
s111, sequentially obtaining a labeled training sample from the labeled training set, and performing forward propagation calculation on the labeled training sample by using the initialized medical entity recognition model to obtain a calculation result;
s112, comparing the real label of the labeled training sample with the calculation result to obtain the identification error of the initialized medical entity identification model;
s113, minimizing the identification error to obtain a model parameter of the initialized medical entity identification model when the identification error is minimum;
s114, performing reverse updating operation by using the model parameters to obtain an updated medical entity identification model;
s115, judging whether the marked training set has marked training samples or not;
when the marked training sample exists in the marked training set, returning to the step of sequentially obtaining one marked training sample from the marked training set, and performing iterative updating on the updated medical entity model;
and when no labeled training sample exists in the labeled training set, S116, using the updated medical entity model as a primary medical entity recognition model.
Specifically, in the training process, the initialized medical entity recognition model has the basic function of recognizing the entity but does not accord with the accuracy of the medical entity recognition, therefore, the initialized medical entity recognition model needs to be pre-trained by using the marked training set, the invention firstly uses the initialized medical entity recognition model to carry out forward propagation calculation on the marked training sample to obtain a calculation result, then compares the calculation result with the real mark or the pseudo mark of the marked training set by using a cross entropy loss function to obtain a recognition error, minimizes the recognition error according to a gradient descent method to obtain the model parameter of the initialized medical entity recognition model when the recognition error is minimum, and further carries out backward propagation on the model parameter to obtain an updated medical entity recognition model, and when the marked training set or the pseudo marked training set is completely trained, obtaining a primary medical entity recognition model.
S2, obtaining a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, and performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set.
In the embodiment of the invention, a jieba word segmentation tool is combined with a Chinese word segmentation system of NLPIR to perform word segmentation operation and part of speech recognition operation on the medical information text to obtain a medical word segmentation set. The jieba word segmentation tool is a special tool for Chinese word segmentation, and the NLPIR Chinese word segmentation system is a platform supporting various codes, various operating systems and various development languages and is mainly used for labeling the part of speech of the word segmentation.
Specifically, the medical information text of "fever and sore throat occurs less than 1 month and 24 days" is used as an example for word segmentation, and then the "Xiaoming", "within", "1 month and 24 days", "occurring", "fever", "throat", "pain" and "symptom" are obtained, and after the parts of speech are labeled, the "Xiaoming [ name ]", "preposition ]", "1 month and 24 days [ time ]", "verb ]" "fever [ common name ]", "common throat [ common name ]" pain [ common name ] "" symptom [ common name ] ") can be obtained.
S3, importing the medical word segmentation set into the medical entity recognition model, carrying out vectorization operation on the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a primary quantized data set, and carrying out attention weight calculation operation on the primary quantized data set to obtain a secondary quantized data set.
In detail, referring to fig. 4, in the embodiment of the present invention, the vectorizing operation is performed on the medical word segmentation set by using the BERT network of the medical entity recognition model to obtain a first-level quantized data set, which includes steps S31 to S34:
s31, vectorizing the medical word segmentation text in the medical word segmentation set by using the BERT network of the medical entity recognition model to obtain a text vector set;
s32, vectorizing the part of speech of each medical participle in the medical participle set to obtain a part of speech vector set;
s33, vectorizing the word sequence of each medical participle in the medical participle set to obtain a word position vector set;
and S34, correspondingly superposing the text vector set, the word vector set and the word position vector set to obtain the first-level quantitative data set.
Specifically, the embodiment of the invention carries out word segmentation on each medical text through quantization operationVectorizing coding to obtain a text vector set [ E ] Xiaoming liquor ”、“E In that ”、“E 1 month and 24 days "… …"; vectorizing and coding the part of speech of each medical word to obtain a part of speech vector set [ E ] Name of a person ”、“E Preposition word ”、“E Time "… …"; vectorizing the word segmentation sequence of each medical text to obtain a word position vector set (E) 0 ”、“E 1 ”、“E 2 "… …"; and finally, correspondingly superposing the text vector set, the word vector set and the word position vector set to obtain the first-level quantitative data set (E) 0 +E Xiaoming liquor +E Name(s) ”、“E 1 +E In that +E Preposition word ”、“E 2 +E 1 month and 24 days +E Time ”……】。
Further, the embodiment of the invention utilizes an attention mechanism, and passes through the preset weight coefficient W q 、W k 、W v Set [ E ] the one-level quantized data 0 +E Xiaoming liquor +E Name(s) ”、“E 1 +E In that +E Preposition word ”、“E 2 +E 1 month and 24 days +E Time "… …" is weighted to obtain "q 0 、k 0 、v 0 ”、“q 1 、k 1 、v 1 ”、“q 2 、k 2 、v 2 "… …", an embodiment of the present invention divides the 1 st word by q 0 With k (k) of each participle 1 、k 2 、k 3 、k 4 … …) are subjected to matrix multiplication to respectively obtain a 0，1 ，a 0，2 ，a 0，3 ,a 0，4 ，a 0，5 … …; according to the attention mechanism, will [ a ] 0，1 v 2 、a 0，2 v 2 、a 0，3 v 3 … …, adding to obtain the second-level quantized data of the 1 st participle.
Similarly, when the secondary quantization data of the 2 nd, 3 rd, 4 th and 5 … … th participles are completed, a secondary quantization data set is obtained.
And S4, recognizing the part-of-speech relationship and semantic relationship between every two adjacent secondary quantized data in the secondary quantized data set by using the part-of-speech-semantic recognition network in the medical entity recognition model to obtain a part-of-speech-semantic feature sequence.
In the embodiment of the invention, the part-of-speech-semantic recognition network comprises a convolution layer, a pooling layer and a flatten layer.
In detail, referring to fig. 5, in an embodiment of the present invention, the recognizing, by using a part-of-speech-semantic recognition network in the medical entity recognition model, part-of-speech relationships and semantic relationships between each adjacent two-level quantized data in the two-level quantized data set to obtain a part-of-speech-semantic feature sequence includes steps S41 to S43:
s41, extracting the characteristics of each secondary quantized data in the secondary quantized data set by using the convolution kernel set of the part-of-speech-semantic recognition network to obtain a convolution characteristic matrix set corresponding to each secondary quantized data;
s42, carrying out average pooling operation on the convolution characteristic matrix set to obtain a pooled characteristic matrix set;
s43, according to a preset recombination rule, splitting each pooled feature matrix in the pooled feature matrix set, and sequentially connecting the split results in a one-dimensional manner to obtain a part of speech-semantic feature sequence.
Specifically, in the embodiment of the present invention, each convolution core in the convolution layer is used to perform feature extraction on each secondary quantization data, so as to obtain a convolution feature matrix set corresponding to each secondary quantization data; taking one convolution feature matrix set of secondary quantization data as an example, in the embodiment of the present invention, the pooling layer is used to perform an average pooling operation on the convolution feature matrix set to obtain a pooled feature matrix set, where the average pooling operation is to extract a feature value set within a preset range, and replace the feature value within the preset range with an average feature of the feature value set, so as to perform dimension reduction on the feature matrix under the condition of preserving the features. In the embodiment of the invention, the flatten layer is utilized to split each pooling feature matrix in the pooling feature matrix set through flattening operation, and the split results are sequentially connected in a one-dimensional manner to obtain a part of speech-semantic feature sequence. And the flatten is also used for reducing the dimension of the characteristic matrix, and is favorable for reducing the calculated amount of the model.
And S5, judging whether each secondary quantized data in the secondary quantized data set is an entity or a combined entity by utilizing an output layer network in the medical entity identification model according to the part of speech-semantic feature sequence, and outputting the entity and the combined entity according to a judgment result to obtain an entity set.
In the embodiment of the invention, the part of speech-semantic feature sequence is analyzed by utilizing a decision tree forest in the output layer network, whether each participle is an entity or not is judged, whether each entity can be combined with front and back adverbs and prepositions to obtain a new entity or not is judged, whether adjacent entities can be combined to obtain a combined entity or not is judged, and finally, the entity and the combined entity are output according to the judgment result to obtain an entity set.
And S6, according to preset map construction rules, performing map construction on the entity set to obtain a medical knowledge map.
In detail, referring to fig. 6, in an embodiment of the present invention, the mapping the entity set according to a preset mapping rule to obtain a medical knowledge map includes steps S61 to S63:
s61, clustering the entity set to obtain a cluster set;
s62, clustering the cluster set into a pre-constructed historical medical knowledge map according to the cluster type of the cluster set;
and S63, constructing the entity set by using the minimum spanning tree in the cluster according to a preset connection rule to obtain the medical knowledge map.
The embodiment of the invention utilizes an Ordering identification Clustering algorithm (Ordering Points To identification the Clustering Structure, OPTICS for short) To cluster the entity set To obtain a cluster set; and connecting top end nodes in the clusters to a pre-constructed historical medical knowledge graph according to the cluster types of all the clusters in the cluster set, and arranging the clusters in a tree structure according to a minimum spanning tree method to obtain an updated medical knowledge graph.
According to the embodiment of the invention, through an attention mechanism in a BERT network, each word is quantized into second quantized data containing the characteristics of the word and the characteristics of adjacent words, so that a subsequent part-of-speech-semantic recognition network can judge whether an entity and adjacent entities can be combined or not according to the second quantized data, and the recognition accuracy is improved; in addition, the embodiment of the invention can effectively improve the training effect of the medical entity recognition model when the sample data is insufficient by the semi-supervised learning method with the labeled training set and the unlabeled training set. Therefore, the map construction method based on the medical entity provided by the embodiment of the invention can accurately and efficiently form the knowledge map by the entity information of the medical industry.
Fig. 7 is a functional block diagram of an atlas database according to an embodiment of the invention.
The medical entity-based atlas database system 100 of the present invention may be installed in an electronic device. According to the implemented functions, the medical entity-based atlas construction apparatus 100 may include a model training module 101, an information vectorization module 102, an entity identification module 103, and an atlas construction module 104. The module of the present invention, which may also be referred to as a unit, refers to a series of computer program segments that can be executed by a processor of an electronic device and that can perform a fixed function, and that are stored in a memory of the electronic device.
In the present embodiment, the functions regarding the respective modules/units are as follows:
the model training module 101 is configured to obtain a pre-constructed initialized medical entity recognition model, a labeled training set and a unlabeled training set, and train the initialized medical entity recognition model by using the labeled training set and the unlabeled training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
the information vectorization module 102 is configured to obtain a pre-constructed medical information text, perform word segmentation on the medical information text by using a chinese word segmentation system, perform part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set, introduce the medical word segmentation set into the medical entity identification model, perform vectorization on the medical word segmentation set by using a BERT network of the medical entity identification model to obtain a first-level quantized data set, and perform attention weight calculation on the first-level quantized data set to obtain a second-level quantized data set;
the entity identification module 103 is configured to identify a part-of-speech relationship and a semantic relationship between each two adjacent sets of secondary quantized data in the set of secondary quantized data by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence, determine, by using an output layer network in the medical entity identification model, whether each piece of secondary quantized data in the set of secondary quantized data is an entity or a combined entity according to the part-of-speech-semantic feature sequence, and output the entity and the combined entity according to a determination result to obtain an entity set;
the map construction module 104 is configured to perform map construction on the entity set according to preset map construction rules to obtain a medical knowledge map.
In detail, when the modules in the atlas configuration apparatus 100 based on a medical entity are used, the same technical means as the atlas configuration method based on a medical entity described in fig. 1 to fig. 3 are adopted, and the same technical effects can be produced, and details are not described here.
Fig. 8 is a schematic structural diagram of an electronic device for implementing a medical entity-based atlas construction method according to an embodiment of the present invention.
The electronic device 1 may comprise a processor 10, a memory 11, a communication bus 12 and a communication interface 13, and may further comprise a computer program, such as a map building program based on a medical entity, stored in the memory 11 and executable on the processor 10.
In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same function or different functions, and includes one or more Central Processing Units (CPUs), a microprocessor, a digital Processing chip, a graphics processor, a combination of various control chips, and the like. The processor 10 is a Control Unit (Control Unit) of the electronic device, connects various components of the electronic device by using various interfaces and lines, and executes various functions and processes data of the electronic device by running or executing programs or modules (e.g., executing a medical entity-based map building program, etc.) stored in the memory 11 and calling data stored in the memory 11.
The memory 11 includes at least one type of readable storage medium including flash memory, removable hard disks, multimedia cards, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, magnetic disks, optical disks, etc. The memory 11 may in some embodiments be an internal storage unit of the electronic device, for example a removable hard disk of the electronic device. The memory 11 may also be an external storage device of the electronic device in other embodiments, such as a plug-in mobile hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the electronic device. Further, the memory 11 may also include both an internal storage unit and an external storage device of the electronic device. The memory 11 may be used not only to store application software installed in the electronic device and various types of data, such as codes of a medical entity-based atlas creation program, etc., but also to temporarily store data that has been output or is to be output.
The communication bus 12 may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is arranged to enable connection communication between the memory 11 and at least one processor 10 or the like.
The communication interface 13 is used for communication between the electronic device and other devices, and includes a network interface and a user interface. Optionally, the network interface may include a wired interface and/or a wireless interface (e.g., WI-FI interface, bluetooth interface, etc.), which are typically used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a Display (Display), an input unit such as a Keyboard (Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable, among other things, for displaying information processed in the electronic device and for displaying a visualized user interface.
Fig. 8 only shows an electronic device with components, and it will be understood by those skilled in the art that the structure shown in fig. 8 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than those shown, or some components may be combined, or a different arrangement of components.
For example, although not shown, the electronic device may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 10 through a power management device, so that functions of charge management, discharge management, power consumption management and the like are realized through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
It is to be understood that the described embodiments are for purposes of illustration only and that the scope of the appended claims is not limited to such structures.
The memory 11 in the electronic device 1 stores a medical entity-based atlas building program that is a combination of instructions that, when executed in the processor 10, may implement:
acquiring a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
acquiring a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, and performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set;
importing the medical word segmentation set into the medical entity recognition model, carrying out vectorization operation on the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a primary quantized data set, and carrying out attention weight calculation operation on the primary quantized data set to obtain a secondary quantized data set;
identifying part-of-speech relations and semantic relations among adjacent secondary quantized data in the secondary quantized data set by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence;
judging whether each secondary quantized data in the secondary quantized data set is an entity or a combined entity or not according to the part of speech-semantic feature sequence by utilizing an output layer network in the medical entity identification model, and outputting the entity and the combined entity according to a judgment result to obtain an entity set;
and according to a preset map construction rule, map construction is carried out on the entity set to obtain the medical knowledge map.
Specifically, the specific implementation method of the instruction by the processor 10 may refer to the description of the relevant steps in the embodiment corresponding to the drawings, which is not described herein again.
Further, the integrated modules/units of the electronic device 1, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. The computer readable storage medium may be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying said computer program code, recording medium, U-disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM).
The present invention also provides a computer-readable storage medium, storing a computer program which, when executed by a processor of an electronic device, may implement:
acquiring a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
acquiring a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, and performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set;
importing the medical word segmentation set into the medical entity recognition model, carrying out vectorization operation on the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a primary quantized data set, and carrying out attention weight calculation operation on the primary quantized data set to obtain a secondary quantized data set;
identifying part-of-speech relations and semantic relations among adjacent secondary quantized data in the secondary quantized data set by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence;
judging whether each secondary quantized data in the secondary quantized data set is an entity or a combined entity or not according to the part of speech-semantic feature sequence by utilizing an output layer network in the medical entity identification model, and outputting the entity and the combined entity according to a judgment result to obtain an entity set;
and according to a preset map construction rule, map construction is carried out on the entity set to obtain the medical knowledge map.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method can be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof.
The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
The embodiment of the application can acquire and process related data based on an artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the system claims may also be implemented by one unit or means in software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (10)
1. A method for atlas construction based on medical entities, the method comprising:
acquiring a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
acquiring a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, and performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set;
importing the medical word segmentation set into the medical entity recognition model, carrying out vectorization operation on the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a primary quantized data set, and carrying out attention weight calculation operation on the primary quantized data set to obtain a secondary quantized data set;
identifying part-of-speech relations and semantic relations among adjacent secondary quantized data in the secondary quantized data set by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence;
judging whether each secondary quantized data in the secondary quantized data set is an entity or a combined entity or not according to the part of speech-semantic feature sequence by utilizing an output layer network in the medical entity identification model, and outputting the entity and the combined entity according to a judgment result to obtain an entity set;
and according to a preset map construction rule, map construction is carried out on the entity set to obtain the medical knowledge map.
2. The medical entity-based atlas construction method of claim 1, wherein the training the initialized medical entity recognition model by using the labeled training set and the unlabeled training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model comprises:
training the initialized medical entity recognition model by using the labeled training set to obtain a primary medical entity recognition model;
performing entity recognition on the label-free training set by using the primary medical entity recognition model to obtain a recognition result set and confidence degrees corresponding to all recognition results in the recognition result set;
extracting the recognition result with the confidence coefficient larger than a preset threshold value in the recognition result set as a pseudo label, and defining a label-free training set corresponding to the pseudo label as a pseudo label training set;
calculating the model error of the primary medical entity recognition model according to the confidence degree corresponding to each recognition result;
judging whether the model error is greater than or equal to a preset standard error or not;
when the model error is larger than or equal to the standard error, judging that the primary medical entity recognition model is unqualified, merging the pseudo-labeled training set into the labeled training set, and returning to the step of training the initialized medical entity recognition model by using the labeled training set;
and when the model error is smaller than the standard error, judging that the primary medical entity identification model is qualified, and setting the primary medical entity identification model as a medical entity identification model.
3. The medical entity-based atlas construction method of claim 2, wherein the training the initial medical entity identification model using the labeled training set to obtain a preliminary medical entity identification model comprises:
sequentially acquiring a labeled training sample from the labeled training set, and performing forward propagation calculation on the labeled training sample by using the initialized medical entity recognition model to obtain a calculation result;
comparing the real label of the labeled training sample with the calculation result to obtain the identification error of the initialized medical entity identification model;
minimizing the identification error to obtain a model parameter of the initialized medical entity identification model when the identification error is minimum;
carrying out reverse updating operation by using the model parameters to obtain an updated medical entity identification model;
judging whether the marked training set has marked training samples or not;
when the marked training sample exists in the marked training set, returning to the step of sequentially obtaining one marked training sample from the marked training set, and performing iterative updating on the updated medical entity model;
and when no labeled training sample exists in the labeled training set, taking the updated medical entity model as a primary medical entity recognition model.
4. The medical entity-based atlas construction method of claim 1, wherein the vectorizing operation on the medical participle set by using the BERT network of the medical entity recognition model to obtain a primary quantitative data set comprises:
vectorizing the medical word segmentation text in the medical word segmentation set by using a BERT network of the medical entity recognition model to obtain a text vector set;
vectorizing the part of speech of each medical participle in the medical participle set to obtain a part of speech vector set;
vectorizing the word sequence of each medical word in the medical word set to obtain a word position vector set;
and correspondingly superposing the text vector set, the word characteristic vector set and the word position vector set to obtain the primary quantized data set.
5. The method for constructing a medical entity-based atlas according to claim 1, wherein the identifying the part-of-speech relationship and the semantic relationship between each adjacent two-level quantized data in the two-level quantized data set by using the part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence comprises:
performing feature extraction on each secondary quantized data in the secondary quantized data set by using a convolution kernel set of the part-of-speech-semantic recognition network to obtain a convolution feature matrix set corresponding to each secondary quantized data;
carrying out average pooling operation on the convolution characteristic matrix set to obtain a pooled characteristic matrix set;
and splitting each pooling feature matrix in the pooling feature matrix set according to a preset recombination rule, and sequentially connecting the splitting results in a one-dimensional manner to obtain a part of speech-semantic feature sequence.
6. The medical entity-based atlas construction method of claim 1, wherein the atlas construction of the entity set according to preset atlas construction rules to obtain the medical knowledge atlas comprises:
clustering the entity set to obtain a cluster set;
clustering the cluster set into a pre-constructed historical medical knowledge map according to the cluster type of the cluster set;
and according to a preset connection rule, constructing the entity set by using a minimum spanning tree in a cluster to obtain the medical knowledge map.
7. The medical entity-based atlas construction method of claim 1, wherein prior to obtaining the pre-constructed initial medical entity identification model, the method further comprises:
acquiring a pre-constructed BERT network and a part-of-speech-semantic recognition network containing recognition semantics and an activation function of the part-of-speech;
and connecting the BERT network serving as an input layer and the part-of-speech-semantic recognition network serving as a processing layer and an output layer to obtain an initialized medical entity recognition model.
8. An apparatus for atlas construction based on a medical entity, the apparatus comprising:
the model training module is used for acquiring a pre-constructed initialized medical entity recognition model, a labeled training set and a label-free training set, and training the initialized medical entity recognition model by using the labeled training set and the label-free training set according to a preset semi-supervised learning strategy to obtain a trained medical entity recognition model;
the information vectorization module is used for obtaining a pre-constructed medical information text, performing word segmentation operation on the medical information text by using a Chinese word segmentation system, performing part-of-speech tagging on a word segmentation result to obtain a medical word segmentation set, introducing the medical word segmentation set into the medical entity recognition model, performing vectorization operation on the medical word segmentation set by using a BERT (best effort network) of the medical entity recognition model to obtain a first-level quantitative data set, and performing attention weight calculation operation on the first-level quantitative data set to obtain a second-level quantitative data set;
an entity identification module, configured to identify a part-of-speech relationship and a semantic relationship between each two adjacent sets of secondary quantized data in the set of secondary quantized data by using a part-of-speech-semantic identification network in the medical entity identification model to obtain a part-of-speech-semantic feature sequence, and determine, by using an output layer network in the medical entity identification model, whether each piece of secondary quantized data in the set of secondary quantized data is an entity or a combined entity according to the part-of-speech-semantic feature sequence, and output the entity and the combined entity according to a determination result to obtain an entity set;
and the map construction module is used for carrying out map construction on the entity set according to a preset map construction rule so as to obtain the medical knowledge map.
9. An electronic device, characterized in that the electronic device comprises:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the medical entity based atlas construction method of any of claims 1-7.
10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the medical entity-based atlas construction method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210439849.3A CN114840684A (en) | 2022-04-25 | 2022-04-25 | Map construction method, device and equipment based on medical entity and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210439849.3A CN114840684A (en) | 2022-04-25 | 2022-04-25 | Map construction method, device and equipment based on medical entity and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114840684A true CN114840684A (en) | 2022-08-02 |
Family
ID=82566142
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210439849.3A Pending CN114840684A (en) | 2022-04-25 | 2022-04-25 | Map construction method, device and equipment based on medical entity and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114840684A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115294426A (en) * | 2022-10-08 | 2022-11-04 | 深圳市益心达医学新技术有限公司 | Method, device and equipment for tracking interventional medical equipment and storage medium |
| CN116311539A (en) * | 2023-05-19 | 2023-06-23 | 亿慧云智能科技（深圳）股份有限公司 | Sleep motion capturing method, device, equipment and storage medium based on millimeter waves |
-
2022
- 2022-04-25 CN CN202210439849.3A patent/CN114840684A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115294426A (en) * | 2022-10-08 | 2022-11-04 | 深圳市益心达医学新技术有限公司 | Method, device and equipment for tracking interventional medical equipment and storage medium |
| CN115294426B (en) * | 2022-10-08 | 2022-12-06 | 深圳市益心达医学新技术有限公司 | Method, device and equipment for tracking interventional medical equipment and storage medium |
| CN116311539A (en) * | 2023-05-19 | 2023-06-23 | 亿慧云智能科技（深圳）股份有限公司 | Sleep motion capturing method, device, equipment and storage medium based on millimeter waves |
| CN116311539B (en) * | 2023-05-19 | 2023-07-28 | 亿慧云智能科技（深圳）股份有限公司 | Sleep motion capturing method, device, equipment and storage medium based on millimeter waves |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112883190A (en) | Text classification method and device, electronic equipment and storage medium | |
| CN113378970B (en) | Sentence similarity detection method and device, electronic equipment and storage medium | |
| CN114840684A (en) | Map construction method, device and equipment based on medical entity and storage medium | |
| CN113821622B (en) | Answer retrieval method and device based on artificial intelligence, electronic equipment and medium | |
| CN115238670A (en) | Information text extraction method, device, equipment and storage medium | |
| CN113344125B (en) | Long text matching recognition method and device, electronic equipment and storage medium | |
| CN113360654B (en) | Text classification method, apparatus, electronic device and readable storage medium | |
| CN114416939A (en) | Intelligent question and answer method, device, equipment and storage medium | |
| CN114461777A (en) | Intelligent question and answer method, device, equipment and storage medium | |
| CN113658002A (en) | Decision tree-based transaction result generation method and device, electronic equipment and medium | |
| CN116468025A (en) | Electronic medical record structuring method and device, electronic equipment and storage medium | |
| CN115346095A (en) | Visual question answering method, device, equipment and storage medium | |
| CN114780688A (en) | Text quality inspection method, device and equipment based on rule matching and storage medium | |
| CN114595321A (en) | Question marking method and device, electronic equipment and storage medium | |
| CN114219367A (en) | User scoring method, device, equipment and storage medium | |
| CN114943306A (en) | Intention classification method, device, equipment and storage medium | |
| CN114610854A (en) | Intelligent question and answer method, device, equipment and storage medium | |
| CN113806540A (en) | Text labeling method and device, electronic equipment and storage medium | |
| CN113706207A (en) | Order transaction rate analysis method, device, equipment and medium based on semantic analysis | |
| CN113515591A (en) | Text bad information identification method and device, electronic equipment and storage medium | |
| CN114462411B (en) | Named entity recognition method, device, equipment and storage medium | |
| CN111680513B (en) | Feature information identification method and device and computer readable storage medium | |
| CN114239595B (en) | Intelligent return visit list generation method, device, equipment and storage medium | |
| CN112988963B (en) | User intention prediction method, device, equipment and medium based on multi-flow nodes | |
| CN114970501A (en) | Text-based entity relationship extraction method, device, equipment and 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 |