A method and apparatus for generating and recognizing human phenotypic ontology terminology models

Abstract

This invention discloses a method and apparatus for generating and recognizing human phenotype ontology terminology, applied in the field of data processing technology. The method includes: acquiring a business case database; performing word segmentation on each case in the business case database to obtain an initial phenotype vocabulary; matching the initial phenotype vocabulary with terms in a medical terminology database to obtain a phenotype lexicon; and constructing a human phenotype ontology terminology recognition model based on the phenotype lexicon and the human phenotype ontology. The human phenotype ontology terminology recognition model is a model for recognizing the correlation between words in the phenotype lexicon and the human phenotype ontology. Compared with the technical solution of manually selecting words from cases and then matching them with the human phenotype ontology, this invention achieves intelligent word segmentation processing of cases and intelligent matching processing of case terms with the human phenotype ontology, improving the accuracy and efficiency of the matching.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and apparatus for generating and recognizing human phenotypic ontology terminology recognition models. Background Technology

[0002] Human Phenotype Ontology (HPO) is a standardized vocabulary and classification system for describing human phenotypes. Currently, the manual process for associating Chinese clinical cases with HPOs generally involves: reading clinical cases, summarizing phenotype and symptom keywords, retrieving associated HPOs from the HPO database based on these keywords, and selecting the HPO with the highest correlation. This entire process is time-consuming. Therefore, the current method of manually reading clinical cases, constructing a Chinese case phenotype keyword database, and manually annotating the associations between key phenotype terms and HPOs results in low efficiency and accuracy in constructing these associations. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method and apparatus for generating and recognizing human phenotype ontology terminology recognition models, which solves the technical problem of low efficiency in determining the human phenotype ontology corresponding to a case in the prior art.

[0004] To address the aforementioned technical problems, this invention provides a method for generating a human phenotype ontology terminology recognition model, comprising:

[0005] Obtain the business case database, and perform word segmentation on each case in the business case database to obtain the initial phenotypic vocabulary;

[0006] The initial phenotypic vocabulary is matched with terms in a medical terminology database to obtain a phenotypic vocabulary.

[0007] A human phenotypic ontology terminology recognition model is constructed based on the phenotypic lexicon and the human phenotypic ontology; wherein, the human phenotypic ontology terminology recognition model is a model for recognizing the correlation between words in the phenotypic lexicon and the human phenotypic ontology.

[0008] Optionally, after performing word segmentation on each case in the business case database to obtain the initial phenotypic vocabulary, the method further includes:

[0009] The initial phenotypic vocabulary is expanded by short vocabulary expansion to obtain an expanded phenotypic vocabulary;

[0010] Accordingly, the step of matching the initial phenotypic vocabulary with terms in a medical terminology database to obtain a phenotypic lexicon includes:

[0011] The extended phenotypic vocabulary is matched with the terms in the medical terminology database to obtain the phenotypic vocabulary.

[0012] Optionally, after matching the initial phenotypic vocabulary with terms in a medical terminology database to obtain a phenotypic vocabulary, the method further includes:

[0013] The terminology in the phenotypic lexicon is subjected to character similarity calculation to determine synonym groups;

[0014] Accordingly, the construction of a human phenotype ontology terminology recognition model based on the phenotype lexicon and the human phenotype ontology includes:

[0015] The human phenotype ontology term recognition model is constructed based on the synonym grouping, the phenotypic lexicon, and the human phenotype ontology.

[0016] Optionally, the step of constructing a human phenotype ontology term recognition model based on the phenotype lexicon and the human phenotype ontology includes:

[0017] The probability of occurrence of a phenotypic keyword is determined based on the number of cases in the business case database and the number of times each phenotypic keyword in the phenotypic thesaurus appears in the cases.

[0018] Based on the business case database, determine the number of times each phenotypic keyword and each human phenotypic ontology entry in the human phenotypic ontology appear simultaneously, and determine the probability of simultaneous occurrence.

[0019] The human phenotype ontology terminology recognition model is constructed based on Bayes' theorem, the occurrence probability of the phenotypic keywords, and the co-occurrence probability.

[0020] Optionally, the step of performing word segmentation on each case in the business case database to obtain an initial phenotypic vocabulary includes:

[0021] The numbers, punctuation marks, and special symbols of each case in the business case database are first filtered to obtain a filtered business case database.

[0022] The word segmentation process is performed on each case in the filtered business case database to obtain the initial phenotypic vocabulary.

[0023] Optionally, the step of performing word segmentation on each case in the business case database to obtain an initial phenotypic vocabulary includes:

[0024] The word segmentation process is performed on each case in the business case database to obtain initial word segmentation data;

[0025] The initial word segmentation data is subjected to a second filtering process to obtain the initial phenotypic vocabulary; wherein, the second filtering process includes at least one of character set filtering, part-of-speech filtering and preset vocabulary filtering.

[0026] This invention also provides a method for human phenotypic ontology term recognition, comprising:

[0027] Obtain electronic medical records;

[0028] The electronic medical record is subjected to keyword filtering to obtain medical record terminology;

[0029] All human phenotype ontology corresponding to the case terms are determined using a human phenotype ontology terminology recognition model; wherein, the human phenotype ontology terminology recognition model is a model obtained based on the aforementioned human phenotype ontology terminology recognition model generation method.

[0030] The present invention also provides a human phenotype ontology terminology recognition model generation apparatus, comprising:

[0031] The word segmentation module is used to obtain a business case database, perform word segmentation on each case in the business case database, and obtain an initial phenotypic vocabulary.

[0032] The standard terminology matching module is used to match the initial phenotypic vocabulary with terms in the medical terminology database to obtain a phenotypic vocabulary.

[0033] A human phenotype ontology terminology recognition model construction module is used to construct a human phenotype ontology terminology recognition model based on the phenotype lexicon and the human phenotype ontology; wherein, the human phenotype ontology terminology recognition model is a model for recognizing the correlation between words in the phenotype lexicon and the human phenotype ontology.

[0034] The present invention also provides an electronic device, comprising:

[0035] Memory, used to store computer programs;

[0036] A processor is configured to implement the above-described human phenotype ontology terminology recognition model generation method and the steps of the above-described human phenotype ontology terminology recognition method when executing the computer program.

[0037] The present invention also provides a computer-readable storage medium on which a computer program is stored, wherein when the computer program is executed by a processor, it implements the steps of the above-described human phenotype ontology terminology recognition model generation method and the above-described human phenotype ontology terminology recognition method.

[0038] As can be seen, this invention obtains an initial phenotypic vocabulary by acquiring a business case database, performing word segmentation on each case in the business case database, and matching the initial phenotypic vocabulary with a medical terminology database to obtain a phenotypic lexicon. Based on the phenotypic lexicon and the human phenotypic ontology, a human phenotypic ontology terminology recognition model is constructed. The human phenotypic ontology terminology recognition model is a model for recognizing the correlation between words in the phenotypic lexicon and the human phenotypic ontology, realizing intelligent word segmentation processing of cases and intelligent matching processing with the human phenotypic ontology, thereby improving matching accuracy and matching efficiency.

[0039] In addition, the present invention also provides a method and apparatus for human phenotype ontology term recognition, which also have the above-mentioned beneficial effects. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0041] Figure 1 A flowchart illustrating a method for generating a human phenotype ontology terminology recognition model, provided in an embodiment of the present invention;

[0042] Figure 2 This is a flowchart illustrating a word segmentation process provided in an embodiment of the present invention.

[0043] Figure 3 A flowchart illustrating a synonym grouping method provided in this embodiment of the invention;

[0044] Figure 4 A flowchart illustrating a method for generating a human phenotype ontology terminology recognition model, provided in an embodiment of the present invention;

[0045] Figure 5 A flowchart illustrating the generation process of a human phenotype ontology terminology recognition model, provided for an embodiment of the present invention;

[0046] Figure 6 A flowchart of a method for recognizing human phenotype ontology terms provided in an embodiment of the present invention;

[0047] Figure 7 This is a schematic diagram of the structure of a human phenotype ontology terminology recognition model generation device provided in an embodiment of the present invention;

[0048] Figure 8 This is a schematic diagram of the structure of a human phenotype ontology terminology recognition device provided in an embodiment of the present invention;

[0049] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please refer to Figure 1 , Figure 1 A flowchart illustrating a method for generating a human phenotypic ontology terminology recognition model, provided in an embodiment of the present invention. The method may include:

[0052] S100: Obtain the business case database, perform word segmentation on each case in the business case database, and obtain the initial phenotypic vocabulary.

[0053] This embodiment does not limit the specific process of segmenting each case in the business case database to obtain the initial phenotypic vocabulary. For example, the cases in the business case database can be filtered first, and then each case in the business case database can be segmented to obtain the initial phenotypic vocabulary. This embodiment does not limit the specific method of segmentation. For example, a conditional random field segmentation model can be used to segment the case text data in the business case database; or, natural language processing can be used to segment each case in the business case database to obtain the initial phenotypic vocabulary.

[0054] It should be noted that after segmenting each case in the business case database to obtain the initial phenotypic vocabulary, the process may further include: expanding the initial phenotypic vocabulary with shorter terms to obtain an expanded phenotypic vocabulary; correspondingly, matching the initial phenotypic vocabulary with terms in a medical terminology database to obtain a phenotypic lexicon, which may include: matching the expanded phenotypic vocabulary with terms in a medical terminology database to obtain a phenotypic lexicon. This embodiment does not limit the specific method of short term expansion. For example, it can expand forward; or expand backward; or expand forward and backward simultaneously. For example, "congenital dysplasia" may be split into "congenital" and "dysplasia"; or multiple cases such as "congenital," "development," and "dysplasia," etc. This implementation will merge overly finely segmented phenotypic vocabulary. It is understood that this embodiment performs short term expansion to ensure the accuracy of terminology, improve the accuracy of subsequent matching, and thus improve the accuracy of model construction. For easier understanding, please refer to [reference needed]. Figure 2 , Figure 2This is a flowchart illustrating a word segmentation process provided in an embodiment of the present invention. The purpose of word segmentation is to extract key medical phenotypic terms from historical business case text data. The construction process involves using a conditional random field (CRF) word segmentation model to quickly obtain a coarse set of segmented terms from the case text. This set may contain a large number of terms unrelated to medical phenotypic meanings, therefore a filtering rule needs to be designed. The processing steps are as follows:

[0055] 1) Filter based on character set value range to extract Chinese and English characters, and filter out all numbers, punctuation marks and special characters.

[0056] In this embodiment, the extraction of Chinese and English characters can be performed using the regular expression: [\u4e00-\u9fa5a-zA-Z]+.

[0057] 2) Use a conditional random field model for word segmentation, retaining nouns, verbs, and adjectives while filtering out other words.

[0058] 3) Preset filter word library to filter out words that are meaningless for phenotypic judgment and frequently appear (such as admission, examination, patient, etc.).

[0059] 4) Short vocabulary expansion.

[0060] For example, "congenital malformation" might be broken down into "congenital" and "malformation"; or into multiple categories such as "congenital", "development", and "malformation". Then, the terms are filtered and merged, and overly detailed phenotypic terms are selected.

[0061] 5) Scan the publicly available medical terminology database BIOS (Medical Knowledge Graph Dataset) to match the standard descriptions of medical terms, ensuring the professionalism of the phenotypic thesaurus.

[0062] 6) Filter and retain high-frequency words. Each phenotypic word is counted only once in a case, and a phenotypic vocabulary is generated.

[0063] It should be noted that the above-mentioned word segmentation process for each case in the business case database to obtain the initial phenotypic vocabulary may include: performing a first filtering process on the numbers, punctuation, and special symbols of each case in the business case database to obtain a filtered business case database; and performing word segmentation on each case in the filtered business case database to obtain the initial phenotypic vocabulary. This embodiment improves the efficiency of subsequent word segmentation by filtering numbers, punctuation, and special symbols before word segmentation.

[0064] It should be noted that the above-mentioned word segmentation processing of each case in the business case database to obtain the initial phenotypic vocabulary may include: performing word segmentation processing on each case in the business case database to obtain initial segmented word data; and performing a second filtering process on the initial segmented word data to obtain the initial phenotypic vocabulary; wherein, the second filtering process includes at least one of character set filtering, part-of-speech filtering, and preset vocabulary filtering. This embodiment performs a second filtering process on the initial segmented word data to filter out unimportant words, reducing the impact on model construction and thus improving the accuracy of model construction.

[0065] S101, Match the initial phenotypic vocabulary with the terms in the medical terminology database to obtain the phenotypic vocabulary.

[0066] The Biomedical Informatics Ontology System (BIOS) in this embodiment is a high-quality, comprehensive medical knowledge graph built upon large-scale, multi-type authoritative medical text data. This knowledge base offers an open-source medical terminology lexicon containing over 18 million Chinese medical terms. This invention uses a columnar ClickHouse (column-based storage database) to store the terminology file. After generating the terminology dataset, it can quickly scan a standard terminology database and include matched terms in the final lexicon. This embodiment matches the initial phenotypic vocabulary with terms in the medical terminology database to match standard descriptive forms of medical terms, ensuring the professionalism of the phenotypic lexicon.

[0067] It should be noted that after matching the initial phenotypic vocabulary with terms in the medical terminology database to obtain the phenotypic lexicon, the process may further include: calculating character similarity for the terms in the phenotypic lexicon to determine synonym groups; correspondingly, constructing a human phenotypic ontology terminology recognition model based on the phenotypic lexicon and the human phenotypic ontology may include: constructing a human phenotypic ontology terminology recognition model based on synonym groups, the phenotypic lexicon, and the human phenotypic ontology. This embodiment does not limit the specific synonym groups. For example, synonym groups could be "muscle tone" corresponding to "dystonia" and "decreased muscle tone in the limbs"; or synonym groups could be "atrial septal defect" corresponding to "atrial septal defect" and "ventricular septal defect". For ease of understanding, please refer to [reference needed]. Figure 3 , Figure 3 This is a flowchart illustrating a synonym grouping method provided in an embodiment of the present invention. The synonym grouping method may specifically include the following steps:

[0068] 1) Calculate the string similarity between each word and then summarize and sort them;

[0069] 2) For words with high character similarity, filter out words with high semantic similarity and group them;

[0070] 3) For words that are not grouped, determine whether they belong to a group and decide whether they belong to a separate group;

[0071] 4) Summarize and merge to form the final synonym group.

[0072] String similarity calculation:

[0073] Using the difflib (text comparison) module of Python (a scripting language), the matching degree index between words is calculated. The calculation formula is as follows: Where T is the total number of elements in the two strings, and M is the number of matched elements. The value is 1.0 if the two strings are identical, and 0.0 if they are completely different. Empirically, a ratio (matching degree) greater than 0.6 means the two sequences are approximately matched.

[0074] S102, Construct a human phenotype ontology terminology recognition model based on a phenotype lexicon and a human phenotype ontology; wherein, the human phenotype ontology terminology recognition model is a model for recognizing the correlation between words in the phenotype lexicon and the human phenotype ontology.

[0075] This embodiment does not limit the specific method of constructing a human phenotype ontology terminology recognition model based on a phenotype lexicon and a human phenotype ontology, as long as an association can be established between the phenotype lexicon and the human phenotype ontology. For example, the human phenotype ontology terminology recognition model can be constructed based on the similarity between words in the phenotype lexicon and words in the human phenotype ontology. Alternatively, the human phenotype ontology terminology recognition model can be determined based on the number of associations between words in the phenotype lexicon and words in the human phenotype ontology. This embodiment does not limit the specific process of subsequently using the human phenotype ontology terminology recognition model. For example, the electronic medical record can be directly compared with the human phenotype ontology terminology recognition model to determine the human phenotype ontology; or the electronic medical record can be filtered for keywords before being compared with the human phenotype ontology terminology recognition model to determine the human phenotype ontology; or the electronic medical record can be filtered and subject to keyword selection before being compared with the human phenotype ontology terminology recognition model to determine the human phenotype ontology.

[0076] It should be noted that the above-mentioned construction of a human phenotype ontology terminology recognition model based on a phenotype lexicon and a human phenotype ontology may include: determining the probability of occurrence of phenotype keywords based on the number of cases in the business case database and the number of times each phenotype keyword in the phenotype lexicon appears in the cases; determining the probability of simultaneous occurrence based on the number of times each phenotype keyword and each human phenotype ontology term appear simultaneously in the business case database; and constructing the human phenotype ontology terminology recognition model based on Bayes' theorem, the probability of occurrence of phenotype keywords, and the probability of simultaneous occurrence. For ease of understanding, this invention provides a flowchart example of the construction of a human phenotype ontology terminology recognition model, and the processing steps may include:

[0077] 1) Load the phenotypic thesaurus and HPO thesaurus, scan all case texts, assuming the number of cases is N, and count the number of occurrences W for each phenotypic keyword. n (Each word is counted only once per case), then the probability of the phenotypic keyword appearing is P. w =W n / N;

[0078] 2. Count the number of times each phenotypic keyword W and each HPO term H appear simultaneously. n The probability P of the simultaneous occurrence of the phenotypic keyword and the HPO is then... wh =WH n / N;

[0079] 3. According to Bayes' theorem, given the presence of phenotypic keyword W, what is the probability of the HPO term H appearing? This probability is the association probability between the phenotypic keyword W and the HPO term H, forming a probability lexicon of association between key medical phenotypes and HPOs, in the following form:

[0080] 4. The probability of HPO association is shared within the synonym groups in the thesaurus.

[0081] The method for generating a human phenotype ontology terminology recognition model provided in this invention includes: acquiring a business case database; performing word segmentation on each case in the business case database to obtain an initial phenotype vocabulary; matching the initial phenotype vocabulary with terms in a medical terminology database to obtain a phenotype lexicon; and constructing a human phenotype ontology terminology recognition model based on the phenotype lexicon and the human phenotype ontology; wherein, the human phenotype ontology terminology recognition model is a model for recognizing the correlation between words in the phenotype lexicon and the human phenotype ontology. Compared with the technical solution of manually selecting criteria from the words in the cases to match the human phenotype ontology, this invention achieves intelligent word segmentation processing of the cases and performs matching processing with the human phenotype ontology, improving matching accuracy and efficiency. Furthermore, it expands short vocabulary to improve the accuracy of phenotype lexicon construction; and it performs similarity calculation to determine synonym groups. Since words in synonym groups can share the same association, it improves the efficiency of subsequent matching. Furthermore, the association model is determined based on the frequency of related occurrences in the phenotypic lexicon and the human phenotypic ontology, thereby improving the model generation efficiency. Additionally, the numbers, punctuation marks, and special symbols in each case in the business case database are first filtered to improve the accuracy of word segmentation. Furthermore, the initial word segmentation data is second filtered to improve the accuracy of keyword extraction.

[0082] For a clearer understanding of this invention, please refer to the following details. Figure 4 , Figure 4A flowchart illustrating a method for generating a human phenotype ontology terminology recognition model, provided in an embodiment of the present invention, may specifically include:

[0083] S400, retrieve the business case database.

[0084] Please refer to the framework diagram corresponding to this embodiment. Figure 5 , Figure 5 This invention provides a flowchart of the generation process of a human phenotypic ontology terminology recognition model. The overall technical solution process is summarized as follows: 1) Word segmentation: segmenting the case text, filtering vocabulary (based on vocabulary length, part of speech, etc.), selecting and screening high-frequency words to form a phenotypic lexicon; 2) Lexicon analysis: grouping synonyms, associating phenotypic terms with HPOs based on statistical methods, and verifying the correctness of the association; 3) Construction of a dictionary of associations between phenotypic terms and HPOs: sharing associated HPOs within each group of synonyms, generating a dictionary of phenotypic terms and HPOs, which includes the association probability of phenotypic terms and HPO entries; 4) Application design: extracting phenotypic keywords from the input text, identifying key phenotypic terms, and recommending HPO entries with a high probability of associating with those phenotypic terms.

[0085] S401 filters the numbers, punctuation, and special symbols in each case in the business case database to obtain the case text data.

[0086] S402, the conditional random field segmentation model is used to segment the case text data to obtain the initial phenotypic segmentation data.

[0087] S403, perform character set filtering, part-of-speech filtering and non-important word filtering on the initial phenotypic word segmentation data to obtain the initial filtered phenotypic word data.

[0088] S404, the initial filtered phenotypic vocabulary data is expanded with short vocabulary to obtain the data to be compared.

[0089] In this embodiment, short word expansion involves concatenating words. A suitable splitting granularity needs to be selected. This method concatenates words shorter than a preset length to form a pending word; this length can be 4.

[0090] S405, scan the publicly available medical terminology database Medical Knowledge Graph dataset, match the standard description format to the data to be compared, and obtain standard case data.

[0091] S406, high-frequency words are filtered from standard case data to generate a phenotypic thesaurus; each word is counted only once in a case.

[0092] S407 calculates character similarity between words in the phenotypic lexicon and groups synonyms based on character similarity.

[0093] S408. Based on statistical methods, calculate the degree of correlation between phenotypic words in the phenotypic lexicon and entries in the human phenotypic ontology, and construct a term recognition model for the human phenotypic ontology.

[0094] For a clearer understanding of this invention, please refer to the following details. Figure 6 , Figure 6 A flowchart of a method for recognizing human phenotypic ontology terms provided in this embodiment of the invention may specifically include:

[0095] S600, obtain electronic medical records.

[0096] S601, perform keyword filtering on electronic medical records to obtain medical record terminology.

[0097] This embodiment does not limit the specific method of keyword screening. For example, keywords can be filtered out; or case terminology can be determined based on keywords.

[0098] S602, using a human phenotype ontology terminology recognition model to determine all human phenotype ontology corresponding to case terms; wherein, the human phenotype ontology terminology recognition model is a model obtained based on the human phenotype ontology terminology recognition model generation method.

[0099] This embodiment does not limit the specific process of determining all human phenotype ontologies corresponding to case terms using a human phenotype ontology terminology recognition model. For example, the phenotype vocabulary corresponding to the case term can be determined, and the corresponding human phenotype ontology term can be determined based on the phenotype vocabulary; or the similarity between the case term and the phenotype vocabulary can be determined, and the phenotype vocabulary with the highest similarity can be used as the target vocabulary, thereby determining the human phenotype ontology corresponding to the target vocabulary; or when it is determined that the case term corresponds to multiple human phenotype ontologies, the human phenotype ontology with a correlation exceeding a preset threshold can be used as the target human phenotype ontology.

[0100] The human phenotype ontology term recognition method provided in this invention, compared with most existing solutions that require manual human phenotype ontology term recognition, can directly use the human phenotype ontology term recognition model to determine all human phenotype ontology corresponding to case terms, thereby improving the accuracy and intelligence of human phenotype ontology term recognition.

[0101] The following describes a human phenotype ontology terminology recognition model generation device provided by an embodiment of the present invention. The human phenotype ontology terminology recognition model generation device described below and the human phenotype ontology terminology recognition model generation method described above can be referred to in correspondence with each other.

[0102] Please refer to the details. Figure 7 , Figure 7A schematic diagram of a human phenotype ontology terminology recognition model generation device provided in an embodiment of the present invention may include:

[0103] The word segmentation processing module 100 is used to obtain a business case database, perform word segmentation processing on each case in the business case database, and obtain an initial phenotypic vocabulary.

[0104] The standard terminology matching module 200 is used to match the initial phenotypic vocabulary with terms in the medical terminology database to obtain a phenotypic vocabulary.

[0105] The human phenotype ontology terminology recognition model construction module 300 is used to construct a human phenotype ontology terminology recognition model based on the phenotype lexicon and the human phenotype ontology; wherein, the human phenotype ontology terminology recognition model is a model for recognizing the correlation between words in the phenotype lexicon and the human phenotype ontology.

[0106] Furthermore, based on the above embodiments, the above-mentioned human phenotype ontology terminology recognition model generation apparatus may further include:

[0107] The short vocabulary expansion module is used to expand the initial phenotypic vocabulary into an expanded phenotypic vocabulary.

[0108] Accordingly, the standard terminology matching module 200 may include:

[0109] A standard terminology matching unit is used to match the extended phenotypic vocabulary with terms in the medical terminology database to obtain the phenotypic vocabulary.

[0110] Furthermore, based on any of the above embodiments, the above-described human phenotype ontology terminology recognition model generation apparatus may further include:

[0111] The character similarity calculation module is used to calculate the character similarity of terms in the phenotypic lexicon and determine the grouping of synonyms;

[0112] Accordingly, the human phenotype ontology terminology recognition model construction module 300 may include:

[0113] The first human phenotype ontology terminology recognition model construction unit is used to construct the human phenotype ontology terminology recognition model based on the synonym grouping, the phenotype lexicon, and the human phenotype ontology.

[0114] Furthermore, based on any of the above embodiments, the above-mentioned human phenotype ontology terminology recognition model construction module 300 may include:

[0115] The keyword occurrence probability determination unit is used to determine the occurrence probability of a phenotypic keyword based on the number of cases in the business case database and the number of times each phenotypic keyword in the phenotypic thesaurus appears in the cases;

[0116] Simultaneous occurrence probability determination unit is used to determine the number of times each phenotype keyword and each human phenotype ontology entry in the human phenotype ontology appear simultaneously based on the business case database, and to determine the simultaneous occurrence probability.

[0117] The second human phenotype ontology terminology recognition model construction unit is used to construct the human phenotype ontology terminology recognition model based on Bayes' theorem, the occurrence probability of the phenotype keywords, and the simultaneous occurrence probability.

[0118] Furthermore, based on any of the above embodiments, the word segmentation processing module 100 may include:

[0119] The first filtering unit is used to perform a first filtering process on the numbers, punctuation marks and special symbols of each case in the business case database to obtain a filtered business case database.

[0120] The first word segmentation processing unit is used to perform word segmentation processing on each case in the filtered business case database to obtain the initial phenotypic vocabulary.

[0121] Furthermore, based on any of the above embodiments, the word segmentation processing module 100 may include:

[0122] The second word segmentation processing unit is used to perform word segmentation processing on each case in the business case database to obtain initial word segmentation data;

[0123] The second filtering unit is used to perform a second filtering process on the initial word segmentation data to obtain the initial phenotypic vocabulary; wherein the second filtering process includes at least one of character set filtering, part-of-speech filtering and preset vocabulary filtering.

[0124] It should be noted that the order of the modules and units in the above-mentioned human phenotype ontology terminology recognition model generation device can be changed without affecting the logic.

[0125] The human phenotype ontology terminology recognition model generation provided in this embodiment of the invention may include: a word segmentation processing module 100, used to acquire a business case database and perform word segmentation processing on each case in the business case database to obtain an initial phenotype vocabulary; a standard terminology matching module 200, used to match the initial phenotype vocabulary with terms in a medical terminology database to obtain a phenotype vocabulary; and a human phenotype ontology terminology recognition model construction module 300, used to construct a human phenotype ontology terminology recognition model based on the phenotype vocabulary and the human phenotype ontology; wherein, the human phenotype ontology terminology recognition model is a model for recognizing the correlation between words in the phenotype vocabulary and the human phenotype ontology. Compared with the technical solution of manually selecting standard matching criteria for words in cases to the human phenotype ontology, this invention achieves intelligent word segmentation processing of cases and performs matching processing with the human phenotype ontology, improving matching accuracy and efficiency. Furthermore, it expands short vocabulary to improve the accuracy of phenotype vocabulary construction; and it performs similarity calculation to determine synonym groups. Since words in synonym groups can share the same association, it improves the efficiency of subsequent matching. Furthermore, the association model is determined based on the frequency of related occurrences in the phenotypic lexicon and the human phenotypic ontology, thereby improving the model generation efficiency. Additionally, the numbers, punctuation marks, and special symbols in each case in the business case database are first filtered to improve the accuracy of word segmentation. Furthermore, the initial word segmentation data is second filtered to improve the accuracy of keyword extraction.

[0126] The following describes a human phenotype ontology terminology recognition device provided by an embodiment of the present invention. The human phenotype ontology terminology recognition device described below can be referred to in correspondence with the human phenotype ontology terminology recognition method described above.

[0127] Please refer to the details. Figure 8 , Figure 8 A schematic diagram of a human phenotype ontology terminology recognition device provided in an embodiment of the present invention may include:

[0128] Electronic medical record acquisition module 400 is used to acquire electronic medical records.

[0129] The filtering module 500 is used to filter electronic medical records by keywords to obtain medical record terminology.

[0130] The human phenotype ontology terminology recognition module 600 is used to determine all human phenotype ontology corresponding to case terms using a human phenotype ontology terminology recognition model; wherein, the human phenotype ontology terminology recognition model is a model obtained based on the human phenotype ontology terminology recognition model generation method.

[0131] It should be noted that the order of the modules and units in the aforementioned human phenotype ontology terminology recognition device can be changed without affecting the logic.

[0132] Compared with most existing solutions that require manual human phenotype ontology terminology recognition, the human phenotype ontology terminology recognition device provided in this invention can directly determine all human phenotype ontology corresponding to case terms using a human phenotype ontology terminology recognition model, thereby improving the accuracy and intelligence of human phenotype ontology terminology recognition.

[0133] The following describes an electronic device provided by an embodiment of the present invention. The electronic device described below can be referred to in correspondence with the human phenotype ontology terminology recognition model generation method and human phenotype ontology terminology recognition method described above.

[0134] Please refer to Figure 9 , Figure 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention may include:

[0135] Memory 10 is used to store computer programs;

[0136] The processor 20 is used to execute computer programs to implement the above-described human phenotype ontology terminology recognition model generation method and human phenotype ontology terminology recognition method.

[0137] The memory 10, processor 20, and communication interface 30 all communicate with each other through the communication bus 40.

[0138] In this embodiment of the invention, the memory 10 is used to store one or more programs. The programs may include program code, which includes computer operation instructions. In this embodiment of the invention, the memory 10 may store programs for implementing the following functions:

[0139] Obtain the business case database, perform word segmentation on each case in the business case database, and obtain the initial phenotypic vocabulary.

[0140] The initial phenotypic vocabulary is matched with terms in a medical terminology database to obtain a phenotypic vocabulary.

[0141] A human phenotypic ontology terminology recognition model is constructed based on a phenotypic lexicon and a human phenotypic ontology; the human phenotypic ontology terminology recognition model is a model for recognizing the correlation between words in the phenotypic lexicon and the human phenotypic ontology.

[0142] Alternatively, memory 10 may store a program for implementing the following functions:

[0143] Obtain electronic medical records.

[0144] Keyword filtering was performed on the electronic medical records to obtain medical terminology.

[0145] The human phenotype ontology terminology recognition model is used to determine all human phenotype ontology corresponding to case terms; wherein, the human phenotype ontology terminology recognition model is a model obtained based on the human phenotype ontology terminology recognition model generation method.

[0146] In one possible implementation, the memory 10 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; and the data storage area may store data created during use.

[0147] Furthermore, memory 10 may include read-only memory and random access memory, providing instructions and data to the processor. A portion of the memory may also include NVRAM. The memory stores operating systems and operating instructions, executable modules, or data structures, or subsets thereof, or extended sets thereof, wherein the operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and handling hardware-based tasks.

[0148] Processor 20 can be a central processing unit (CPU), an application-specific integrated circuit, a digital signal processor, a field-programmable gate array, or other programmable logic device. Processor 20 can be a microprocessor or any conventional processor. Processor 20 can call programs stored in memory 10.

[0149] The communication interface 30 can be an interface for the communication module, used to connect with other devices or systems.

[0150] Of course, it should be noted that, Figure 9 The structure shown does not constitute a limitation on the electronic device in the embodiments of the present invention. In practical applications, the electronic device may include more than Figure 9 More or fewer components as shown, or combinations of certain components.

[0151] The following describes the computer-readable storage medium provided in the embodiments of the present invention. The computer-readable storage medium described below can be referred to in correspondence with the human phenotype ontology terminology recognition model generation method and the human phenotype ontology terminology recognition method described above.

[0152] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for generating a human phenotype ontology terminology recognition model and the method for recognizing human phenotype ontology terms.

[0153] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0154] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0155] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0156] Finally, it should be noted that in this document, relationships such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0157] The foregoing has provided a detailed description of the human phenotype ontology terminology recognition model generation and recognition method and apparatus provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A human phenotype ontology term identification model generation method, characterized by, The method comprises the following steps: obtaining a business case library, performing word segmentation processing on each case in the business case library by using a conditional random field word segmentation model to obtain initial word segmentation data, performing second filtering processing on the initial word segmentation data to obtain initial phenotype vocabulary, wherein the second filtering processing comprises at least one of character set filtering, part-of-speech filtering and preset vocabulary filtering; performing forward expansion and backward expansion on the initial phenotype vocabulary to obtain expanded phenotype vocabulary; matching the expanded phenotype vocabulary with terms in a medical terminology database to obtain a phenotype vocabulary library; constructing a human phenotype ontology term recognition model based on the phenotype vocabulary library and a human phenotype ontology, wherein the human phenotype ontology term recognition model is a model for recognizing the relevance of vocabulary in the phenotype vocabulary library to the human phenotype ontology; wherein constructing the human phenotype ontology term recognition model based on the phenotype vocabulary library and the human phenotype ontology comprises: based on the number of cases in the service case library and each phenotype keyword in the phenotype keyword library the number of occurrences in the cases, to determine the appearance probability of the phenotype keyword ; determining a number of simultaneous occurrences of each human phenotype ontology term in the human phenotype ontology and each phenotype keyword in the business case library , determining a simultaneous occurrence probability ; constructing the human phenotype ontology term recognition model based on the Bayes formula, the appearance probability of the phenotype keyword, and the simultaneous appearance probability; wherein the human phenotype ontology term recognition model is . 2.The human phenotype ontology term identification model generation method of claim 1, wherein, after matching the expanded phenotype vocabulary with the terms in the medical terminology database to obtain the phenotype vocabulary library, further comprising: performing character similarity calculation on the terms in the phenotype vocabulary library to determine synonym groups; correspondingly, the method of constructing the human phenotype ontology term recognition model based on the phenotype vocabulary library and the human phenotype ontology comprises: constructing the human phenotype ontology term recognition model based on the synonym groups, the phenotype vocabulary library and the human phenotype ontology. 3.The human phenotype ontology term identification model generation method of claim 1, wherein, The method comprises the following steps: obtaining a business case library, performing word segmentation processing on each case in the business case library by using a conditional random field word segmentation model to obtain initial word segmentation data, performing second filtering processing on the initial word segmentation data to obtain initial phenotype vocabulary, wherein the second filtering processing comprises at least one of character set filtering, part-of-speech filtering and preset vocabulary filtering; performing first filtering processing on numbers, punctuation marks and special symbols of each case in the business case library to obtain a filtered business case library; 4. A human phenotype ontology term identification method, characterized by, performing the word segmentation processing on each case in the filtered business case library to obtain the initial word segmentation data. The method comprises the following steps: obtaining an electronic case; performing keyword screening on the electronic case to obtain case terms; 5.A human phenotype ontology term identification model generation device characterized by comprising: determining all human phenotype ontologies corresponding to the case terms by using a human phenotype ontology term recognition model, wherein the human phenotype ontology term recognition model is a model obtained by the method for generating a human phenotype ontology term recognition model according to any one of claims 1 to 3. The method comprises the following steps: a word segmentation processing module is configured to obtain a business case library, perform word segmentation processing on each case in the business case library by using a conditional random field word segmentation model to obtain initial word segmentation data, perform second filtering processing on the initial word segmentation data to obtain initial phenotype vocabulary, wherein the second filtering processing comprises at least one of character set filtering, part-of-speech filtering and preset vocabulary filtering; a short vocabulary expansion module is configured to perform forward expansion and backward expansion on the initial phenotype vocabulary to obtain expanded phenotype vocabulary; a standard term matching module is configured to match the expanded phenotype vocabulary with terms in a medical terminology database to obtain a phenotype vocabulary library; The human phenotype ontology term recognition model construction module is configured to construct a human phenotype ontology term recognition model based on the phenotype vocabulary and the human phenotype ontology, wherein the human phenotype ontology term recognition model is a model for recognizing the correlation degree between the vocabulary in the phenotype vocabulary and the human phenotype ontology. The human phenotype ontology term recognition model construction module comprises: a keyword occurrence probability determining unit configured to determine a keyword occurrence probability based on a case share in the service case library and each phenotype keyword in the phenotype keyword library a number of occurrences in the case, and determine the phenotype keyword occurrence probability ; A simultaneous occurrence probability determining unit is configured to determine, according to the service case library, a number of simultaneous occurrences of each human phenotype ontology term in the human phenotype ontology and the each phenotype keyword , determine the simultaneous occurrence probability ; The human phenotype ontology term recognition model construction unit is configured to construct the human phenotype ontology term recognition model based on a Bayesian formula, the appearance probability of the phenotype keyword, and the simultaneous appearance probability. .

6. An electronic device, comprising: The human phenotype ontology term recognition model construction module comprises: The memory is configured to store a computer program. The processor is configured to execute the computer program to implement the steps of the human phenotype ontology term recognition model generation method according to any one of claims 1 to 3 and the human phenotype ontology term recognition method according to claim 4.

7. A computer-readable storage medium, characterized in that, The computer program is stored on the computer readable storage medium and is configured to be executed by the processor to implement the steps of the human phenotype ontology term recognition model generation method according to any one of claims 1 to 3 and the human phenotype ontology term recognition method according to claim 4.

Images