Method, device, equipment and medium for generating epidemiological investigation report based on semantic recognition

Abstract

The present application relates to a method, device, equipment and storage medium for generating an epidemic investigation report based on semantic recognition. The method comprises: performing a missing value filling operation on the original epidemic investigation statistical data to obtain target epidemic investigation statistical data, generating a composite document based on the target epidemic investigation statistical data and the corresponding document, calculating the text similarity between each document in the composite document and performing a deduplication operation to obtain the target composite document, performing a syntactic analysis on each document to obtain the syntactic analysis vector of each document and then inputting it into a semantic recognition model to obtain the semantic recognition result of each document, summarizing the target epidemic investigation statistical data and semantic recognition results and filling them into a template file to generate a target epidemic investigation report. The present application can obtain a target epidemic investigation report with semantic information related to the epidemic investigation document for reference by relevant decision makers, and the content of the target epidemic investigation report is concise.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, equipment and storage medium for generating an epidemiological investigation report based on semantic recognition. Background Art

[0002] Epidemiological investigation is the abbreviation of epidemiological survey. It is the key to controlling epidemics. The purpose of epidemiological investigation is to know the places that the relevant people have visited and the people they have contacted in a certain period of time in the past, so as to determine whether the people have the possibility of spreading the disease or the risk of being infected with the disease.

[0003] At present, when conducting epidemiological investigations on the relevant investigated population, epidemiological investigators will fill in relevant epidemiological investigation statistics and epidemiological investigation documents. Afterwards, epidemiological investigation reports need to be generated based on the epidemiological investigation statistics and epidemiological investigation documents for decision-makers in disease prevention and control to provide decision-making references. Although there are related solutions for automatically generating reports in the prior art, most of these solutions simply fill the content directly into a blank template, which cannot reflect the relevant semantics of the epidemiological investigation documents in the report. In addition, since the number of investigated persons is usually large and the places where the investigated persons have appeared are also large, epidemiological investigators in different regions may conduct epidemiological investigations on the same investigated person and generate multiple epidemiological investigation documents. The epidemiological investigation reports generated by these solutions are too complicated. Summary of the invention

[0004] In view of the above, the present application provides a method, device, equipment and storage medium for generating an epidemiological investigation report based on semantic recognition, which aims to solve the above technical problems.

[0005] In a first aspect, the present application provides a method for generating an epidemiological investigation report based on semantic recognition, the method comprising:

[0006] Reading original flow survey statistical data from a preset flow survey database, performing a missing value filling operation on the original flow survey statistical data, and obtaining target flow survey statistical data;

[0007] Generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data;

[0008] Calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document;

[0009] Performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document;

[0010] The target flow survey statistical data and the semantic recognition results are summarized and filled into a preset template file to generate a target flow survey report.

[0011] Preferably, performing a missing value filling operation on the original flow survey statistical data includes:

[0012] The following formula is used to perform missing value filling operation on the data in the numerical format in the original flow survey statistical data:

[0013]

[0014] Among them, K represents the fill value, i represents the number of variables, and X it represents the value of a variable at a point in time, X il represents the value of a variable at another point in time. Represents the standard deviation determined based on the observed data of the i-th variable.

[0015] Preferably, the calculating of the text similarity between each document in the compound document comprises:

[0016] Segmenting each document in the compound document to obtain at least one text segment corresponding to each document;

[0017] Perform a word segmentation operation on each text segment to obtain at least one word segment of each text segment, and convert each word segment into a word vector;

[0018] Based on the similarity of each word vector between each text segment, determine the similarity between each text segment;

[0019] Based on the similarity between each text segment, the text similarity between each document in the compound document is calculated.

[0020] Preferably, performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document includes:

[0021] Selecting at least two documents in the composite document whose text similarity is greater than a preset threshold;

[0022] Reading identification information of the at least two documents, and determining whether there is a document with the same identification information among the at least two documents;

[0023] When it is determined that there is a document with the same identification information among the at least two documents, any one of the documents with the same identification information is retained.

[0024] Preferably, performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector includes:

[0025] Performing a sentence splitting operation on each document in the target compound document to obtain at least one sentence corresponding to each document;

[0026] Performing syntactic content analysis on the at least one sentence to obtain a syntactic content vector corresponding to the at least one sentence;

[0027] Performing dependency syntactic analysis on the at least one sentence to obtain a syntactic dependency vector corresponding to the at least one sentence;

[0028] Performing syntactic structure analysis on the at least one sentence to obtain a syntactic structure vector corresponding to the at least one sentence;

[0029] The syntactic content vector, the syntactic dependency vector and the syntactic structure vector are integrated to obtain a syntactic analysis vector for each sentence, and a syntactic analysis vector for each document is obtained based on the syntactic analysis vector for each sentence.

[0030] Preferably, the fusion of the syntactic content vector, the syntactic dependency vector and the syntactic structure vector to obtain a syntactic analysis vector for each sentence includes:

[0031] The syntactic analysis vector of each sentence is obtained using the following formula:

[0032] V sum =aV1+bV2+(1-ab)V3

[0033] Among them, V sum represents the syntactic analysis vector, V1 represents the syntactic dependency vector, V2 represents the syntactic structure vector, V3 represents the syntactic content vector, and a and b represent preset weights respectively.

[0034] Preferably, after generating the target epidemiological survey report, the method further includes:

[0035] The target flow survey report is sent to a preset terminal and stored in a preset template library.

[0036] In a second aspect, the present application provides a device for generating an epidemiological investigation report based on semantic recognition, and the device for generating an epidemiological investigation report based on semantic recognition includes:

[0037] Preprocessing module: used to read original flow survey statistical data from a preset flow survey database, perform missing value filling operation on the original flow survey statistical data, and obtain target flow survey statistical data;

[0038] The first generating module is used to generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data;

[0039] Deduplication module: used for calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document;

[0040] Recognition module: used for performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document;

[0041] The second generation module is used to summarize the target flow survey statistical data and the semantic recognition results and fill them into the preset template file to generate a target flow survey report.

[0042] In a third aspect, the present application provides an electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0043] Memory, used to store computer programs;

[0044] The processor is used to implement the steps of the method for generating an epidemiological investigation report based on semantic recognition as described in any one of the embodiments of the first aspect when executing the program stored in the memory.

[0045] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for generating an epidemiological investigation report based on semantic recognition as described in any embodiment of the first aspect are implemented.

[0046] The above technical solution provided by the embodiment of the present application has the following advantages compared with the prior art:

[0047] The method, device, equipment and storage medium for generating an epidemic investigation report based on semantic recognition proposed in the present application can obtain more accurate target epidemic investigation statistical data by performing a missing value filling operation on the original epidemic investigation statistical data. The composite document created based on the target epidemic investigation statistical data and the documents corresponding to the target epidemic investigation data can improve the adaptability of the epidemic investigation report when identifying the semantics of the epidemic investigation document. The composite document can be deduplicated by calculating the text similarity between each document to obtain the target composite document. The duplicate epidemic investigation documents are deleted to make the content of the subsequently generated epidemic investigation report clearer and more concise. Perform syntactic analysis on each document in the target composite document to obtain a syntactic analysis vector for each document. Since the syntactic analysis vector can more accurately represent the semantics expressed by the document, the syntactic analysis vector is input into a pre-trained semantic recognition model to accurately obtain the semantic recognition result of each document. The target epidemic investigation statistical data and the semantic recognition result are summarized and filled into a preset template file to obtain a target epidemic investigation report with semantic information related to the epidemic investigation document. The target epidemic investigation report can be provided for reference by relevant decision makers. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0050] Figure 1 This is a flowchart diagram of a preferred embodiment of the method for generating an epidemiological investigation report based on semantic recognition of the present application;

[0051] Figure 2 This is a module diagram of a preferred embodiment of the apparatus for generating an epidemiological investigation report based on semantic recognition in this application;

[0052] Figure 3 A schematic diagram of a preferred embodiment of the electronic device of the present application;

[0053] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0055] This application provides a method for generating an epidemiological investigation report based on semantic recognition. Figure 1 The figure is a method flow chart of an embodiment of the method for generating an epidemiological investigation report based on semantic recognition of the present application. The method can be executed by an electronic device, and the electronic device can be implemented by software and / or hardware. The method for generating an epidemiological investigation report based on semantic recognition includes:

[0056] Step S10: reading original flow survey statistical data from a preset flow survey database, performing a missing value filling operation on the original flow survey statistical data, and obtaining target flow survey statistical data;

[0057] Step S20: Generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data;

[0058] Step S30: calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document;

[0059] Step S40: performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document;

[0060] Step S50: Summarize the target flow survey statistical data and the semantic recognition results and fill them into a preset template file to generate a target flow survey report.

[0061] In this embodiment, the preset epidemiological investigation database can be a database that stores epidemiological investigation data, which includes original epidemiological investigation statistical data, epidemiological investigation documents, epidemiological investigation voice data, epidemiological investigation video data, etc. collected by epidemiological investigators during the epidemiological investigation of the investigated persons. The epidemiological investigation documents record text information useful for disease prevention and control that the epidemiological investigators learned during the epidemiological investigation process.

[0062] Since the data in the preset flow survey database comes from different applications, the original flow survey statistical data read from the preset flow survey database may contain missing data, so it is also necessary to perform data filling operations on the original flow survey statistical data, and use the flow survey statistical data after the data filling operation as the target flow survey statistical data. The missing value filling operation can be at least one of global constant automatic filling, central metric automatic filling, and group mean automatic filling. Furthermore, data cleaning operations such as data consistency detection and error data detection can also be performed on the original flow survey statistical data.

[0063] After obtaining the target epidemiological survey statistical data with the missing values ​​filled in, obtain the epidemiological survey documents corresponding to the target epidemiological survey statistical data. For example, if there are 30 persons under investigation involved in the target epidemiological survey statistical data, obtain the epidemiological survey documents corresponding to the 30 persons under investigation, and generate a composite document based on the target epidemiological survey statistical data and the epidemiological survey documents corresponding to the target epidemiological survey data. Composite documents can contain data such as text, graphics, spreadsheet data, sound, and video. Composite documents change the way users organize documents and collaborate with other users on the network. The trend in establishing composite documents is to use object-oriented technology, and non-standard information such as images and sounds can be included in the document as independent, self-contained objects.

[0064] After obtaining the compound document, the text similarity between each flow-adjusting document in the compound document can be calculated. For example, after segmenting the two documents, the cosine similarity distance or Euclidean distance between the sentences is calculated to determine the similarity between the sentences. The similarities between the sentences of the two documents are summed as the text similarity of the documents, or the text similarity between the two documents is obtained by comparing the Hamming distance of the simhash values ​​of the two documents. After obtaining the text similarity between the documents, the compound document is deduplicated according to the text similarity to obtain the target compound document, and only one document is retained for the documents in the compound document whose similarity is greater than a preset threshold (for example, 95%). For example, if there are three documents with a similarity greater than 95% between each other, two documents are deleted and only one document is retained.

[0065] Perform syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector for each document. Syntactic analysis includes syntactic structure analysis and syntactic dependency analysis. Syntactic analysis can eliminate ambiguity in terms of lexical and structural aspects of document sentences and can analyze the internal structure of sentences, such as component composition, contextual relationship, etc. Syntactic structure analysis can be performed using rule-based analysis methods or statistical-based analysis methods, and syntactic dependency analysis can be performed using generative dependency analysis methods, discriminative dependency analysis methods, or decision-making dependency analysis methods.

[0066] The obtained syntactic analysis vector can accurately represent the semantics expressed by the document, and the syntactic analysis vector is input into the pre-trained semantic recognition model to obtain the semantic recognition result of each document. Among them, the pre-trained semantic recognition model can be trained based on the hidden Markov model. Model training is a relatively mature technology and will not be described in detail here.

[0067] Read the relevant template files from the template library. The template files are pre-configured according to the standard format of the relevant reports. The target epidemiological survey statistics and the semantic recognition results of the documents are summarized and filled into the preset template files to generate the target epidemiological survey report. The target epidemiological survey report can provide relevant personnel with a detailed report of the epidemiological survey, for example, which places are places where close contacts of more diseases have appeared, how many people have arrived in this area from the risk areas of the disease in the recent period, etc., so as to provide a decision-making reference for disease prevention and control.

[0068] Furthermore, after the target flow survey report is generated, the target flow survey report can also be sent to a preset terminal and stored in a preset template library. The preset terminal can be a terminal corresponding to a relevant decision-maker who receives the target flow survey report.

[0069] This application can obtain more accurate target epidemic investigation statistics by performing missing value filling operations on the original epidemic investigation statistics data. The composite document created based on the target epidemic investigation statistics data and the documents corresponding to the target epidemic investigation data can improve the adaptability of the epidemic investigation report when identifying the semantics of the epidemic investigation document. The composite document can be deduplicated by calculating the text similarity between each document to obtain the target composite document. The duplicate epidemic investigation documents are deleted to make the content of the subsequently generated epidemic investigation report clearer and more concise. Perform syntactic analysis on each document in the target composite document to obtain a syntactic analysis vector for each document. Since the syntactic analysis vector can more accurately represent the semantics expressed by the document, the syntactic analysis vector is input into a pre-trained semantic recognition model to accurately obtain the semantic recognition result of each document. The target epidemic investigation statistics data and semantic recognition results are summarized and filled into a preset template file to obtain a target epidemic investigation report with semantic information related to the epidemic investigation document. The target epidemic investigation report can be provided for reference by relevant decision makers.

[0070] In one embodiment, performing a missing value filling operation on the original flow adjustment statistical data includes:

[0071] The following formula is used to perform missing value filling operation on the data in the numerical format in the original flow survey statistical data:

[0072]

[0073] Among them, K represents the fill value, i represents the number of variables, and X it represents the value of a variable at a point in time, X il represents the value of a variable at another point in time. represents the standard deviation determined based on the observed data of the i-th variable. The difference between the variables at different time points is divided by The purpose is to eliminate the dimensional differences between different variables.

[0074] Since there is a certain correlation between different variables in the management process of epidemiological survey statistical data, and considering the variable characteristics of the longitudinal data set of the original epidemiological survey statistical data, the above formula is used to fill in the missing values, that is, in the time point records containing complete data, the time point record with the highest match degree with the non-missing variable value score in the time point record of missing data is determined, and based on the corresponding variable value in the complete time point record, the data can be supplemented in combination with the actual variables.

[0075] In one embodiment, calculating the text similarity between each document in the compound document includes:

[0076] Segmenting each document in the compound document to obtain at least one text segment corresponding to each document;

[0077] Perform a word segmentation operation on each text segment to obtain at least one word segment of each text segment, and convert each word segment into a word vector;

[0078] Based on the similarity of each word vector between each text segment, determine the similarity between each text segment;

[0079] Based on the similarity between each text segment, the text similarity between each document in the compound document is calculated.

[0080] After segmenting each document in the compound document, the text segment corresponding to each document can be obtained. The word segmentation algorithm is used to perform word segmentation operations on each text segment to obtain the word segmentation of each text segment. The word segmentation algorithm can be a maximum forward matching algorithm or a maximum reverse matching algorithm, or a bidirectional maximum matching algorithm. Then, the word vector model word2vec model is used to convert each word segmentation into a word vector. The word2vec model can process word segmentation as a vector operation in a vector space, which can be used to calculate the similarity between word segments.

[0081] Since each text segment has word vectors corresponding to multiple word segments, the vector obtained by concatenating the word vectors corresponding to each text segment is used as the vector of the text segment. The cosine similarity algorithm can be used to calculate the similarity between two text segments. After obtaining the similarity between the two text segments, the average similarity between the text segments corresponding to the two documents is calculated as the text similarity between the two documents. For example, document D has text segments d1, d2 and d3, and document E has text segments e1, e2 and e3. The average of the similarities between d1 and e1, the similarity between d2 and e2, and the similarity between d3 and e3 is used as the similarity between document D and document E.

[0082] In one embodiment, performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document includes:

[0083] Selecting at least two documents in the composite document whose text similarity is greater than a preset threshold;

[0084] Reading identification information of the at least two documents, and determining whether there is a document with the same identification information among the at least two documents;

[0085] When it is determined that there is a document with the same identification information among the at least two documents, any one of the documents with the same identification information is retained.

[0086] According to the text similarity between documents, at least two documents with text similarity greater than a preset threshold (for example, 95%) are selected, and the identification information of at least two documents is read. The identification information can be the name or mobile phone number of the person under investigation corresponding to the document. It is determined whether there are documents with the same identification information among the at least two documents with a similarity greater than the preset threshold. That is, when there are documents with high similarity, it is determined whether the documents with high similarity are documents corresponding to the same person under investigation. Since the flow investigation documents in the flow investigation database come from different applications, it is necessary to deduplicate the documents of the same person under investigation. When it is determined that there are documents with the same identification information (that is, when there are documents of the same person under investigation), any one of the documents with the same identification information is retained.

[0087] In one embodiment, performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector includes:

[0088] Performing a sentence splitting operation on each document in the target compound document to obtain at least one sentence corresponding to each document;

[0089] Performing syntactic content analysis on the at least one sentence to obtain a syntactic content vector corresponding to the at least one sentence;

[0090] Performing dependency syntactic analysis on the at least one sentence to obtain a syntactic dependency vector corresponding to the at least one sentence;

[0091] Performing syntactic structure analysis on the at least one sentence to obtain a syntactic structure vector corresponding to the at least one sentence;

[0092] The syntactic content vector, the syntactic dependency vector and the syntactic structure vector are integrated to obtain a syntactic analysis vector for each sentence, and a syntactic analysis vector for each document is obtained based on the syntactic analysis vector for each sentence.

[0093] By performing a sentence segmentation operation on each document in the target composite document, at least one sentence corresponding to each document can be obtained, and the syntactic content vector, syntactic dependency vector and syntactic structure vector of the sentence are obtained respectively, and the syntactic content vector, syntactic dependency vector and syntactic structure vector are merged to obtain the syntactic analysis vector of each sentence, and the syntactic analysis vector of each sentence is concatenated to obtain the syntactic analysis vector of each document. Specifically, the syntactic content vector of the sentence can be obtained by inputting the word segmentation of the sentence into the Bert model to obtain the word vector of each word segmentation, averaging the word vector corresponding to each word segmentation to obtain the syntactic content vector of the sentence, and calculating the syntactic content vector of the sentence using the following formula:

[0094] V3=(p1+p2+p3+…+p m ) / m

[0095] Among them, V3 represents the syntactic content vector, p m Represents the word vector of the mth word segmentation of the sentence, where m represents the total number of word segments in the sentence.

[0096] The syntactic structure vector can be obtained based on syntactic structure analysis. Syntactic structure analysis refers to the analysis of the grammatical functions of words in a sentence. Through analysis, the syntactic structure in the sentence and the relationship between the components of the sentence can be discovered. For example, for the sentence "Who am I", a syntactic structure analysis can be performed on it, for example, it can be found that "I" is the subject, "I" is the predicate, and "who" is the object. Syntactic structure analysis can be based on a probabilistic context-free model, based on a center word driven syntactic analysis, etc., which is not limited here.

[0097] The syntactic dependency vector can be obtained by using a graph-based dependency syntactic analysis or a transfer-based dependency syntactic analysis. Both analysis methods can extract the syntactic dependency vector.

[0098] In one embodiment, the fusion of the syntactic content vector, the syntactic dependency vector and the syntactic structure vector to obtain a syntactic analysis vector for each sentence includes:

[0099] The syntactic analysis vector of each sentence is obtained using the following formula:

[0100] V sum =aV1+bV2+(1-ab)V3

[0101] Among them, V sum represents the syntactic analysis vector, V1 represents the syntactic dependency vector, V2 represents the syntactic structure vector, V3 represents the syntactic content vector, and a and b represent preset weights respectively.

[0102] Since the syntactic content vector, syntactic dependency vector and syntactic structure vector of a sentence all affect the semantics of the sentence, by assigning relevant weights to the syntactic content vector, syntactic dependency vector and syntactic structure vector of the sentence, and then concatenating and fusing them to obtain the syntactic analysis vector of the sentence, the semantics expressed by the sentence can be more accurately represented. a and b can take the same value or different values. For example, the values ​​of a and b are both 0.3, or a takes 0.2 and b takes 0.3. The values ​​of a and b are usually less than 0.4.

[0103] Reference Figure 2 As shown, it is a schematic diagram of the functional modules of the epidemiological investigation report generation device 100 based on semantic recognition of the present application.

[0104] The flow survey report generation device 100 based on semantic recognition described in the present application can be installed in an electronic device. According to the functions implemented, the flow survey report generation device 100 based on semantic recognition may include a preprocessing module 110, a first generation module 120, a deduplication module 130, an identification module 140 and a second generation module 150. The module described in the present application may also be referred to as a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0105] In this embodiment, the functions of each module / unit are as follows:

[0106] Preprocessing module 110: used to read original flow survey statistical data from a preset flow survey database, perform missing value filling operation on the original flow survey statistical data, and obtain target flow survey statistical data;

[0107] The first generating module 120 is used to generate a composite document based on the target flow adjustment statistical data and the document corresponding to the target flow adjustment data;

[0108] Deduplication module 130: used for calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document;

[0109] Recognition module 140: used to perform syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and input the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document;

[0110] The second generation module 150 is used to summarize the target flow survey statistical data and the semantic recognition results and fill them into a preset template file to generate a target flow survey report.

[0111] In one embodiment, performing a missing value filling operation on the original flow adjustment statistical data includes:

[0112] The following formula is used to perform missing value filling operation on the data in the numerical format in the original flow survey statistical data:

[0113]

[0114] Among them, K represents the fill value, i represents the number of variables, and X it represents the value of a variable at a point in time, X il represents the value of a variable at another point in time. Represents the standard deviation determined based on the observed data of the i-th variable.

[0115] In one embodiment, calculating the text similarity between each document in the compound document includes:

[0116] Segmenting each document in the compound document to obtain at least one text segment corresponding to each document;

[0117] Perform a word segmentation operation on each text segment to obtain at least one word segment of each text segment, and convert each word segment into a word vector;

[0118] Based on the similarity of each word vector between each text segment, determine the similarity between each text segment;

[0119] Based on the similarity between each text segment, the text similarity between each document in the compound document is calculated.

[0120] In one embodiment, performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document includes:

[0121] Selecting at least two documents in the composite document whose text similarity is greater than a preset threshold;

[0122] Reading identification information of the at least two documents, and determining whether there is a document with the same identification information among the at least two documents;

[0123] When it is determined that there is a document with the same identification information among the at least two documents, any one of the documents with the same identification information is retained.

[0124] In one embodiment, performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector includes:

[0125] Performing a sentence splitting operation on each document in the target compound document to obtain at least one sentence corresponding to each document;

[0126] Performing syntactic content analysis on the at least one sentence to obtain a syntactic content vector corresponding to the at least one sentence;

[0127] Performing dependency syntactic analysis on the at least one sentence to obtain a syntactic dependency vector corresponding to the at least one sentence;

[0128] Performing syntactic structure analysis on the at least one sentence to obtain a syntactic structure vector corresponding to the at least one sentence;

[0129] The syntactic content vector, the syntactic dependency vector and the syntactic structure vector are integrated to obtain a syntactic analysis vector for each sentence, and a syntactic analysis vector for each document is obtained based on the syntactic analysis vector for each sentence.

[0130] In one embodiment, the fusion of the syntactic content vector, the syntactic dependency vector and the syntactic structure vector to obtain a syntactic analysis vector for each sentence includes:

[0131] The syntactic analysis vector of each sentence is obtained using the following formula:

[0132] V sum =aV1+bV2+(1-ab)V3

[0133] Among them, V sum represents the syntactic analysis vector, V1 represents the syntactic dependency vector, V2 represents the syntactic structure vector, V3 represents the syntactic content vector, and a and b represent preset weights respectively.

[0134] In one embodiment, the semantic recognition-based epidemiological investigation report generation device 100 further includes a sending module 160, and the sending module 160 is used to send the target epidemiological investigation report to a preset terminal and store it in a preset template library.

[0135] Reference Figure 3 , which is a schematic diagram of a preferred embodiment of the electronic device 1 of the present application.

[0136] The electronic device 1 includes but is not limited to: a memory 11, a processor 12, a display 13 and a communication interface 14. The electronic device 1 is connected to a network via the communication interface 14. The network may be a wireless or wired network such as an intranet, the Internet, a global system of mobile communication (GSM), a wideband code division multiple access (WCDMA), a 4G network, a 5G network, Bluetooth, Wi-Fi, a call network, etc.

[0137] Among them, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a hard disk or memory of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the electronic device 1. Of course, the memory 11 can also include both the internal storage unit of the electronic device 1 and its external storage device. In this embodiment, the memory 11 is generally used to store the operating system and various application software installed on the electronic device 1, such as the program code of the flow survey report generation program 10 based on semantic recognition. In addition, the memory 11 can also be used to temporarily store various types of data that have been output or are to be output.

[0138] In some embodiments, the processor 12 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 12 is generally used to control the overall operation of the electronic device 1, such as performing control and processing related to data interaction or communication. In this embodiment, the processor 12 is used to run the program code stored in the memory 11 or process data, such as running the program code of the flow survey report generation program 10 based on semantic recognition.

[0139] The display 13 may be referred to as a display screen or a display unit. In some embodiments, the display 13 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an organic light-emitting diode (OLED) touch device, etc. The display 13 is used to display information processed in the electronic device 1 and to display a visual working interface.

[0140] The communication interface 14 may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), and the communication interface 14 is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0141] Figure 3Only an electronic device 1 having components 11-14 and a flow survey report generation program 10 based on semantic recognition is shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0142] Optionally, the electronic device 1 may further include a user interface, which may include a display (Display), an input unit such as a keyboard (Keyboard), and the optional user interface may also include a standard wired interface and a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an organic light-emitting diode (Organic Light-Emitting Diode, OLED) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0143] The electronic device 1 may further include a radio frequency (RF) circuit, a sensor, an audio circuit, etc., which will not be described in detail here.

[0144] In the above embodiment, when the processor 12 executes the semantic recognition-based flow survey report generation program 10 stored in the memory 11, the following steps can be implemented:

[0145] Reading original flow survey statistical data from a preset flow survey database, performing a missing value filling operation on the original flow survey statistical data, and obtaining target flow survey statistical data;

[0146] Generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data;

[0147] Calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document;

[0148] Performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document;

[0149] The target flow survey statistical data and the semantic recognition results are summarized and filled into a preset template file to generate a target flow survey report.

[0150] The storage device may be the memory 11 of the electronic device 1 , or may be another storage device that is communicatively connected to the electronic device 1 .

[0151] For a detailed description of the above steps, please refer to the above Figure 2Functional module diagram of an embodiment of an epidemiological investigation report generation device 100 based on semantic recognition and Figure 1 Description of a flowchart of an embodiment of a method for generating an epidemiological investigation report based on semantic recognition.

[0152] In addition, the embodiment of the present application also proposes a computer-readable storage medium, which can be non-volatile or volatile. The computer-readable storage medium can be any one or any combination of a hard disk, a multimedia card, an SD card, a flash memory card, an SMC, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disk read-only memory (CD-ROM), a USB memory, etc. The computer-readable storage medium includes a data storage area and a program storage area. The program storage area stores a flow survey report generation program 10 based on semantic recognition. When the flow survey report generation program 10 based on semantic recognition is executed by the processor, the following operations are implemented:

[0153] Reading original flow survey statistical data from a preset flow survey database, performing a missing value filling operation on the original flow survey statistical data, and obtaining target flow survey statistical data;

[0154] Generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data;

[0155] Calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document;

[0156] Performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document;

[0157] The target flow survey statistical data and the semantic recognition results are summarized and filled into a preset template file to generate a target flow survey report.

[0158] The specific implementation of the computer-readable storage medium of the present application is roughly the same as the specific implementation of the above-mentioned semantic recognition-based epidemiological investigation report generation method, which will not be repeated here.

[0159] It should be noted that the serial numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments. And the terms "including", "comprising" or any other variants thereof herein are intended to cover non-exclusive inclusion, so that the process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the presence of other identical elements in the process, device, article or method including the element.

[0160] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, an electronic device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0161] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for generating an epidemic investigation report based on semantic recognition, characterized in that: The method comprises: Reading original flow survey statistical data from a preset flow survey database, performing a missing value filling operation on the original flow survey statistical data, and obtaining target flow survey statistical data; Generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data; Calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document; Performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document; Summarize the target flow survey statistical data and the semantic recognition results and fill them into a preset template file to generate a target flow survey report; The performing of the missing value filling operation on the original flow survey statistical data includes: The following formula is used to perform missing value filling operation on the data in the numerical format in the original flow survey statistical data: Among them, K represents the fill value, i represents the number of variables, and X it represents the value of a variable at a point in time, X il represents the value of a variable at another point in time. Represents the standard deviation determined based on the observed data of the i-th variable.

2. The method for generating an epidemic investigation report based on semantic recognition according to claim 1, characterized in that: The calculating the text similarity between each document in the compound document comprises: Segmenting each document in the compound document to obtain at least one text segment corresponding to each document; Perform a word segmentation operation on each text segment to obtain at least one word segment of each text segment, and convert each word segment into a word vector; Based on the similarity of each word vector between each text segment, determine the similarity between each text segment; Based on the similarity between each text segment, the text similarity between each document in the compound document is calculated.

3. The method for generating an epidemic investigation report based on semantic recognition according to claim 1 or 2, characterized in that: The performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document includes: Selecting at least two documents in the composite document whose text similarity is greater than a preset threshold; Reading identification information of the at least two documents, and determining whether there is a document with the same identification information among the at least two documents; When it is determined that there is a document with the same identification information among the at least two documents, any one of the documents with the same identification information is retained.

4. The method for generating an epidemic investigation report based on semantic recognition according to claim 1, characterized in that: The performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector includes: Performing a sentence splitting operation on each document in the target compound document to obtain at least one sentence corresponding to each document; Performing syntactic content analysis on the at least one sentence to obtain a syntactic content vector corresponding to the at least one sentence; Performing dependency syntactic analysis on the at least one sentence to obtain a syntactic dependency vector corresponding to the at least one sentence; Performing syntactic structure analysis on the at least one sentence to obtain a syntactic structure vector corresponding to the at least one sentence; The syntactic content vector, the syntactic dependency vector and the syntactic structure vector are integrated to obtain a syntactic analysis vector for each sentence, and a syntactic analysis vector for each document is obtained based on the syntactic analysis vector for each sentence.

5. The method for generating an epidemic investigation report based on semantic recognition according to claim 4, characterized in that: The fusing the syntactic content vector, the syntactic dependency vector and the syntactic structure vector to obtain a syntactic analysis vector for each sentence includes: The syntactic analysis vector of each sentence is obtained using the following formula: V sum =aV1+bV2+(1-ab)V3 Among them, V sum represents the syntactic analysis vector, V1 represents the syntactic dependency vector, V2 represents the syntactic structure vector, V3 represents the syntactic content vector, and a and b represent preset weights respectively.

6. The method for generating an epidemic investigation report based on semantic recognition according to claim 1, characterized in that: After generating the target flow survey report, the method further includes: The target flow survey report is sent to a preset terminal and stored in a preset template library.

7. A device for generating an epidemic investigation report based on semantic recognition, characterized in that: The device comprises: Preprocessing module: used to read original flow survey statistical data from a preset flow survey database, perform missing value filling operation on the original flow survey statistical data, and obtain target flow survey statistical data; The first generating module is used to generate a composite document based on the target flow survey statistical data and the document corresponding to the target flow survey data; Deduplication module: used for calculating the text similarity between each document in the compound document, and performing a deduplication operation on the compound document based on the text similarity to obtain a target compound document; Recognition module: used for performing syntactic analysis on each document in the target compound document to obtain a syntactic analysis vector of each document, and inputting the syntactic analysis vector into a pre-trained semantic recognition model to obtain a semantic recognition result of each document; The second generation module is used to summarize the target flow survey statistical data and the semantic recognition results and fill them into a preset template file to generate a target flow survey report; The performing of the missing value filling operation on the original flow survey statistical data includes: The following formula is used to perform missing value filling operation on the data in the numerical format in the original flow survey statistical data: Among them, K represents the fill value, i represents the number of variables, and X it represents the value of a variable at a point in time, X il represents the value of a variable at another point in time. Represents the standard deviation determined based on the observed data of the i-th variable.

8. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor is used to implement the method for generating an epidemiological investigation report based on semantic recognition as described in any one of claims 1 to 6 when executing the program stored in the memory.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for generating an epidemiological investigation report based on semantic recognition as described in any one of claims 1 to 6 are implemented.

Images