Metadata-based functional prototype modeling and naming method and system

Abstract

The invention discloses a metadata-based functional prototype modeling and naming method and system, and belongs to the technical field of system modeling, and the method specifically comprises the following steps: establishing a metadatabase, defining a keyword rule, intelligently extracting keywords in a user search request by using a natural language processing and deep learning model, and searching a matching record in the metadatabase; if the matching item is not found, triggering an intelligent naming process, and generating a new name according to an automatic naming rule, a machine learning algorithm and diversity evaluation; the system stores new names and related information by adopting a distributed storage mode, and introduces a data deduplication and consistency check mechanism to ensure real-time updating and data accuracy of a metadatabase; and the modeling and naming efficiency and accuracy of the function prototype are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of system modeling, and in particular to a metadata-based functional prototype modeling naming method and system. Background Art

[0002] With the rapid development of computer technology, network technology and information construction, data storage methods are changing with each passing day, and the amount of data is also growing exponentially. In the field of product development and design, virtual prototype technology has been widely used as a comprehensive design method. Virtual prototype technology integrates simulation technology, information technology and various computer-aided technologies, and can simulate the functions and performance of real products, providing strong support for product design and optimization. However, in the modeling process of virtual prototypes, the naming problem has always been a key technical point that needs to be solved. However, traditional naming methods often rely on manual experience and rules, lack unified standards and specifications, resulting in ambiguity and inconsistency in naming results. This not only affects the management and maintenance of the model, but also increases the difficulty of subsequent development work.

[0003] For example, the Chinese patent application with publication number CN111459918A discloses a system modeling naming method based on a metadata database, which includes: according to the characteristics of the unit or industry where the customer is located, constantly accumulating and collecting the Chinese and English standard names of professional vocabulary required by the user, and then in the system modeling process, according to the naming elements entered by the user during naming, real-time retrieval is performed in the metadata database, and a vocabulary list that meets the user input is returned. The user quickly selects the required metadata according to the naming requirements to name the model elements in Chinese and English. This technical solution can form a unified naming specification according to the unit where each user is located, improve the efficiency and quality of user modeling, and help users build high-quality models without ambiguity between departments.

[0004] The above existing technologies all have the following problems: they rely on users to manually search and select appropriate naming terms in the metadata database, and lack intelligent extraction and matching functions; users need to manually browse and select naming terms in the metadata database, which will consume a lot of time and energy; they mainly focus on the retrieval and selection of existing naming terms, and do not provide specific solutions for the generation of new names. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a metadata-based functional prototype modeling naming method and system, establishes a metadata database and defines keyword rules, uses natural language processing and deep learning models to intelligently extract keywords in user search requests, and searches for matching records in the metadata database; if no match is found, the intelligent naming process is triggered to generate a new name based on automatic naming rules, machine learning algorithms and diversity assessments; the system uses a distributed storage method to save new names and related information, and introduces data deduplication and consistency checking mechanisms to ensure real-time updating of the metadata database and data accuracy; the efficiency and accuracy of functional prototype modeling naming are improved.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The naming method of functional prototype modeling based on metadata includes:

[0008] Step S1: Establish a metadata database and define keyword rules;

[0009] Step S2: According to the defined keyword rules, natural language processing methods are used in combination with deep learning models to intelligently extract keywords from the search request input by the user, and matching strategies are used to search for records matching the keywords in the metadata database;

[0010] Step S3: If a matching record is found in the metadata database, it indicates that the keyword already exists, and the system directly calls the name record corresponding to the keyword. If no matching record is found in the metadata database, it indicates that the keyword is a new search, and the intelligent naming process is triggered. According to the automatic naming rules, combined with the machine learning algorithm and the diversity evaluation method, a new name is generated;

[0011] Step S4: Use a distributed storage method to write the newly generated name and its related information into the metadata database for storage, and introduce a data deduplication mechanism and a consistency check mechanism to obtain an updated metadata database.

[0012] Specifically, the specific steps of step S2 include:

[0013] S2.1: receiving a search request input by a user, and determining the form of the search request input by the user;

[0014] If the search request input by the user is in text form, preprocessing the input search request text;

[0015] If the search request input by the user is in the form of voice and picture, the non-text input is converted into text using voice recognition and OCR technology, and the converted text is pre-processed;

[0016] S2.2: Use natural language processing methods to parse the preprocessed search request, and use a deep learning model to intelligently extract keywords based on the defined keyword length and part of speech requirements to form a keyword list;

[0017] S2.3: Develop keyword matching strategies and determine the priorities and weights of keyword matching strategies;

[0018] S2.4: accessing the metadata database, using the keyword matching strategy to search for records matching the keyword list in the metadata database, and sorting the query results according to the priority and weight to generate a query result list;

[0019] S2.5: According to the query result list, the query results are returned to the user and displayed.

[0020] Specifically, the specific steps of S2.2 include:

[0021] S2.21: obtaining the preprocessed search request text, and performing word segmentation processing, splitting the preprocessed search request text into words or phrases, and obtaining a word segmented text;

[0022] S2.22: Load the pre-trained part-of-speech tagging model, and use the pre-trained part-of-speech tagging model to perform part-of-speech tagging on the segmented text, assigning a part-of-speech to each word or phrase;

[0023] S2.23: defining the length of the keyword as L, and based on the defined keyword length, selecting words or phrases that meet the length condition from the text after part-of-speech tagging and segmentation, to obtain preliminary words or phrases;

[0024] S2.24: defining part-of-speech requirements, and further screening out words or phrases meeting the part-of-speech conditions from the preliminary words or phrases according to the defined part-of-speech requirements, to obtain screened words or phrases;

[0025] S2.25: Load the pre-trained deep learning model, use the pre-trained deep learning model to score the screened words or phrases, and select the N words or phrases with the highest scores as keywords according to the scoring results;

[0026] S2.26: Sort the extracted keywords from high to low according to the scores to form a keyword list.

[0027] Specifically, the specific steps of S2.3 include:

[0028] S2.31: Obtain the keyword list in step S2.2, and classify the keyword list to obtain a classified keyword list;

[0029] S2.32: According to the classified keyword list, perform exact matching, fuzzy matching, partial matching or synonym matching on different keywords, and determine the priority of each matching strategy according to business requirements and the importance of keywords. At the same time, assign weights to different keywords and matching strategies, and generate matching strategy documents;

[0030] S2.33: Apply the matching strategy document to actual business, monitor the implementation effect of the strategy, and optimize and adjust the matching strategy based on the monitoring results.

[0031] Specifically, when performing exact matching, fuzzy matching, partial matching or synonym matching on different keywords in S2.32:

[0032] If the keyword is a professional term or a word with a specific meaning, a string matching algorithm is used to compare the search request entered by the user with the words in the classified keyword list;

[0033] If the search request entered by the user contains spelling errors or is incomplete, the fuzzy matching algorithm is used to calculate the similarity S between the search request entered by the user and the keywords in the classified keyword list, and a similarity threshold h is set. When S>h, the match is considered successful;

[0034] If the search request entered by the user is part of the keywords in the classified keyword list, the prefix tree is used to speed up the partial matching process;

[0035] If the search request input by the user uses a synonym or near-synonym of a keyword, a synonym dictionary is established to associate the keyword with its synonym. When the user input matches a synonym of a word in the keyword list, the match is considered successful.

[0036] Specifically, the specific process of the intelligent naming process in S3 includes:

[0037] S3.1: Generate candidate names according to a preset automatic naming rule to obtain a candidate name list;

[0038] S3.2: Based on the candidate name list, select candidate name features, wherein the candidate name features include character type and similarity with existing names;

[0039] S3.3: Load the pre-built support vector machine model, and use the selected candidate name features to train the pre-built support vector machine model to obtain a trained support vector machine model;

[0040] S3.4: Input the candidate names to be evaluated into the trained support vector machine model, the support vector machine model outputs evaluation results according to the features of the candidate names to be evaluated, and sorts the candidate names according to the evaluation results;

[0041] S3.5: Select the candidate name with the highest score as the optimal name according to the sorting result, add the optimal name as the new name to the metadata database, and return the new name to the requester.

[0042] Specifically, in step S1, the metadata database integrates storage, intelligent indexing and rule management functions.

[0043] The metadata-based functional prototype modeling and naming system includes: metadata database management module, matching module, intelligent naming module, and data storage module;

[0044] The metadata database management module is used to establish and maintain a metadata database containing keyword rules and existing name records;

[0045] The matching module is used to extract keywords from the search request input by the user using a natural language processing method and a deep learning model, and search for matching records in the metadata database;

[0046] The intelligent naming module is used to trigger the intelligent naming process and generate a new name when no matching record is found in the metadata database;

[0047] The data storage module is used to write the newly generated name and related information into the metadata database for storage, and to update the metadata database.

[0048] Specifically, the matching module includes a matching strategy unit, and the matching strategy unit is used to search for records matching the extracted keywords in the metadata database according to defined keyword rules.

[0049] Specifically, the intelligent naming module includes: an automatic naming rule unit and a diversity evaluation unit, wherein the automatic naming rule unit is used to define and update rules for generating new names; and the diversity evaluation unit is used to perform diversity evaluation on the newly generated names.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. The present invention proposes a metadata-based functional prototype modeling and naming system, and optimizes and improves the architecture, operating steps and processes. The system has the advantages of simple processes, low investment and operating costs, and low production work costs.

[0052] 2. The present invention proposes a metadata-based functional prototype modeling naming method. By establishing a metadata database and defining keyword rules, the system can efficiently manage and retrieve information. When a user enters a search request, the system can intelligently extract keywords and quickly find matching records in the metadata database, thereby improving the accuracy and efficiency of information retrieval; for existing names, the system can directly call the record, avoiding repeated searches and naming, saving the user's time and energy.

[0053] 3. The present invention proposes a metadata-based functional prototype modeling naming method. For new search requests, the system can trigger an intelligent naming process and generate new names based on automatic naming rules, machine learning algorithms and diversity evaluation methods. This naming method not only improves the degree of automation of naming, but also ensures the diversity and innovation of names. At the same time, a distributed storage method is used to save the newly generated names and their related information, and a data deduplication mechanism and a consistency check mechanism are introduced to further enhance the reliability and stability of the metadata database. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic diagram of a metadata-based functional prototype modeling and naming method of the present invention;

[0055] Figure 2 This is a principle flow chart of the metadata-based functional prototype modeling and naming method of the present invention;

[0056] Figure 3 A flowchart of keyword matching implementation of the metadata-based functional prototype modeling naming method of the present invention;

[0057] Figure 4 A system architecture diagram is named for metadata-based functional prototype modeling of the present invention. DETAILED DESCRIPTION

[0058] Example 1

[0059] See also Figure 1 and Figure 2 , an embodiment of the present invention provides: a metadata-based functional prototype modeling and naming method, comprising the following steps:

[0060] Step S1: Establish a metadata database and define keyword rules;

[0061] Furthermore, the specific steps of step S1 include:

[0062] (1) Determine the purpose and objectives of the metadata repository, identify the types of data that need to be stored and managed, and determine the specific requirements of the keyword rules;

[0063] (2) Design the database table structure, including fields, data types, indexes, and determine the relationships and constraints between data;

[0064] (3) Use the database management system MySQL to create databases and tables, and configure database security and permissions;

[0065] (4) Determine the method for extracting keywords, set the storage format and location of keywords, and write or configure a rule engine to process keywords;

[0066] (5) Write code or script to extract keywords from the data source and store the extracted keywords in the metadata database;

[0067] (6) Test the metadata database and keyword rules to ensure that they work as expected, and adjust and optimize the database structure and rules to improve performance and accuracy;

[0068] (7) Deploy the metadata database and keyword rules to the production environment, and regularly monitor and maintain the database to ensure its normal operation and data accuracy.

[0069] Step S2: According to the defined keyword rules, natural language processing methods are used in combination with deep learning models to intelligently extract keywords from the search request input by the user, and matching strategies are used to search for records matching the keywords in the metadata database;

[0070] Step S3: If a matching record is found in the metadata database, it indicates that the keyword already exists, and the system directly calls the name record corresponding to the keyword. If no matching record is found in the metadata database, it indicates that the keyword is a new search, and the intelligent naming process is triggered. According to the automatic naming rules, combined with the machine learning algorithm and the diversity evaluation method, a new name is generated;

[0071] Step S4: Use a distributed storage method to write the newly generated name and its related information into the metadata database for storage, and introduce a data deduplication mechanism and a consistency check mechanism to obtain an updated metadata database.

[0072] Example 2

[0073] See also Figure 3 In this embodiment, the specific steps of step S2 include:

[0074] S2.1: receiving a search request input by a user, and determining the form of the search request input by the user;

[0075] If the search request input by the user is in text form, preprocessing the input search request text;

[0076] If the search request input by the user is in the form of voice and picture, the non-text input is converted into text using voice recognition and OCR technology, and the converted text is pre-processed;

[0077] Furthermore, the specific steps of the speech recognition part include:

[0078] (1) Audio collection: Acquire the user's voice input through a microphone collection device;

[0079] (2) Preprocessing: Preprocess the collected audio signals, including noise reduction, echo removal, speech enhancement, etc., to improve the accuracy of speech recognition;

[0080] (3) Feature extraction: extracting parameters that can represent speech features, such as Mel-frequency cepstral coefficients, from the preprocessed audio signal. The feature parameter extraction process is a prior art in the art and is not an inventive solution of the present application, and is not described in detail here.

[0081] (4) Acoustic model matching: Match the extracted feature parameters with the models in the acoustic model library. The acoustic model is usually obtained by training a large amount of speech data and can recognize different phonemes or words. The acoustic model in the present invention uses a hidden Markov model to model and match acoustic features. These models are obtained by training a large amount of speech data and can learn the mapping relationship between speech signals and texts. However, the recognition process of the hidden Markov model is the prior art content in this field and is not the inventive solution of this application, so it will not be described in detail here;

[0082] (5) Language model decoding: Based on the matching results of the acoustic model, the language model is combined for decoding to obtain the most likely text sequence. The language model is used to describe the statistical relationship between words and help the recognition system understand the vocabulary and grammatical structure in the speech.

[0083] (6) Post-processing: Post-process the decoded text, such as adding punctuation marks, converting uppercase and lowercase letters, etc., to improve the readability of the text.

[0084] The specific steps of the OCR part include:

[0085] (1) Image acquisition: obtaining images containing text through a camera, scanner or image file;

[0086] (2) Preprocessing: Preprocess the image, including grayscale, binarization, and denoising, to improve the accuracy of OCR recognition;

[0087] (3) Layout analysis: Perform layout analysis on the preprocessed image to determine information such as the position, size, and direction of the text;

[0088] (4) Character recognition: using an OCR recognition algorithm to recognize characters in an image and convert them into editable text, wherein the OCR recognition algorithm in the present invention uses a convolutional neural network model to recognize and classify characters. However, the specific process of using a convolutional neural network model to recognize and classify characters is the prior art content in this field and is not the inventive solution of the present application, and is not described in detail here;

[0089] (5) Post-processing: Post-process the recognized text, such as error correction and formatting, to improve the accuracy and readability of the text.

[0090] S2.2: Use natural language processing methods to parse the preprocessed search request, and use a deep learning model to intelligently extract keywords based on the defined keyword length and part of speech requirements to form a keyword list;

[0091] The preprocessed search request includes the preprocessed search request text and the preprocessed non-text conversion text.

[0092] S2.3: Develop keyword matching strategies and determine the priorities and weights of keyword matching strategies;

[0093] S2.4: accessing the metadata database, using the keyword matching strategy to search for records matching the keyword list in the metadata database, and sorting the query results according to the priority and weight to generate a query result list;

[0094] S2.5: According to the query result list, the query results are returned to the user and displayed.

[0095] The specific steps of S2.2 include:

[0096] S2.21: obtaining the preprocessed search request text, and performing word segmentation processing, splitting the preprocessed search request text into words or phrases, and obtaining a word segmented text;

[0097] S2.22: Load the pre-trained part-of-speech tagging model, and use the pre-trained part-of-speech tagging model to perform part-of-speech tagging on the segmented text, assigning a part-of-speech to each word or phrase;

[0098] Furthermore, the specific steps of S2.22 include:

[0099] (1) Select a pre-trained part-of-speech tagging model based on the Jieba framework and use the corresponding library function to load the pre-trained model;

[0100] (2) Use the Jieba word segmenter to segment the input text into words or phrases;

[0101] (3) Use the pre-trained part-of-speech tagging model based on the Jieba framework to tag the text after word segmentation. The pre-trained part-of-speech tagging model will assign a part-of-speech tag to each word or phrase, such as noun, verb, or adjective.

[0102] (4) Output the annotated text in the format of words / phrases and their corresponding part-of-speech tags.

[0103] S2.23: defining the length of the keyword as L, and based on the defined keyword length, selecting words or phrases that meet the length condition from the text after part-of-speech tagging and segmentation, to obtain preliminary words or phrases;

[0104] S2.24: defining part-of-speech requirements, and further screening out words or phrases meeting the part-of-speech conditions from the preliminary words or phrases according to the defined part-of-speech requirements, to obtain screened words or phrases;

[0105] Among them, the part-of-speech requirement refers to the standard regarding the part-of-speech of words or phrases that needs to be followed during the screening process. The definition of the part-of-speech requirement is usually based on linguistic classification, such as nouns, words that represent people, things, places, concepts, etc.; verbs, words that represent actions or states; in the present invention, the part-of-speech requirement is set to nouns or verbs, which means that during the screening process, only words or phrases marked as nouns or verbs will be retained. At the same time, it also means that during the screening process, a mechanism is needed, such as natural language processing tools or manual marking, to identify the part of speech of each word or phrase, and screen accordingly.

[0106] S2.25: Load the pre-trained deep learning model, use the pre-trained deep learning model to score the screened words or phrases, and select the N words or phrases with the highest scores as keywords according to the scoring results;

[0107] Furthermore, the specific steps of S2.25 include:

[0108] (1) Select the ELMo pre-trained model according to task requirements and use the deep learning framework TensorFlow to load the ELMo pre-trained model;

[0109] (2) Preprocess the filtered words or phrases, such as word segmentation, stop word removal, stemming or lemmatization, and convert the preprocessed text into an input format acceptable to the model;

[0110] (3) Inputting the preprocessed text into the ELMo pre-trained model to obtain the score of each word or phrase of the ELMo pre-trained model, wherein the score may be the probability value output by the model or the similarity of the embedded vector, and the specific calculation and operation process of the ELMo pre-trained model is the prior art content in this field, which is not the inventive solution of the present application and is not described in detail here;

[0111] (4) Based on the scoring results, select the N words or phrases with the highest scores as keywords.

[0112] S2.26: Sort the extracted keywords from high to low according to the scores to form a keyword list.

[0113] The specific steps of S2.3 include:

[0114] S2.31: Obtain the keyword list in step S2.2, and classify the keyword list to obtain a classified keyword list;

[0115] S2.32: Based on the classified keyword list, perform exact matching, fuzzy matching, partial matching or synonym matching on different keywords, and determine the priority of each matching strategy based on business needs and the importance of keywords. At the same time, assign weights to different keywords and matching strategies, and generate a matching strategy document. The weight assignment should be based on factors such as the keyword's search volume, relevance, and conversion rate. The matching strategy document includes the definition, priority, and weight assignment of each matching strategy.

[0116] S2.33: Apply the matching strategy document to actual business, such as search engine optimization and advertising, monitor the implementation effect of the strategy, and optimize and adjust the matching strategy based on the monitoring results. The implementation effect of the monitoring strategy includes indicators such as traffic, conversion rate, and click-through rate. When optimizing and adjusting the matching strategy, if a single strategy is not effective, consider adjusting its priority or weight. If new keywords appear or old keywords lose relevance, update the keyword list and reformulate the matching strategy.

[0117] When performing exact matching, fuzzy matching, partial matching or synonym matching on different keywords in S2.32:

[0118] If the keyword is a professional term or a word with a specific meaning, a string matching algorithm is used to compare the search request entered by the user with the words in the classified keyword list;

[0119] If the search request entered by the user contains spelling errors or is incomplete, the fuzzy matching algorithm is used to calculate the similarity S between the search request entered by the user and the keywords in the classified keyword list, and a similarity threshold h is set. When S>h, the match is considered successful;

[0120] If the search request entered by the user is part of the keywords in the classified keyword list, the prefix tree is used to speed up the partial matching process;

[0121] If the search request input by the user uses a synonym or near-synonym of a keyword, a synonym dictionary is established to associate the keyword with its synonym. When the user input matches a synonym of a word in the keyword list, the match is considered successful.

[0122] The specific process of the intelligent naming process in S3 includes:

[0123] S3.1: Generate candidate names according to a preset automatic naming rule to obtain a candidate name list;

[0124] S3.2: Based on the candidate name list, select candidate name features, wherein the candidate name features include character type and similarity with existing names;

[0125] S3.3: Loading a pre-built support vector machine model, and training the pre-built support vector machine model using the selected candidate name features to obtain a trained support vector machine model, wherein the construction and training process of the support vector machine model is the prior art content in this field, and is not the inventive solution of the present application, and will not be described in detail here;

[0126] S3.4: Input the candidate names to be evaluated into the trained support vector machine model, the support vector machine model outputs evaluation results according to the characteristics of the candidate names to be evaluated, and sorts the candidate names according to the evaluation results, wherein the operation and implementation process of the support vector machine model is the prior art content in this field, and is not the inventive solution of the present application, and will not be described in detail here;

[0127] S3.5: Select the candidate name with the highest score as the optimal name according to the sorting result, add the optimal name as the new name to the metadata database, and return the new name to the requester.

[0128] In step S1, the metadata database integrates storage, intelligent indexing and rule management functions.

[0129] Example 3

[0130] See also Figure 4 Another embodiment provided by the present invention is a metadata-based functional prototype modeling and naming system, comprising:

[0131] Metadatabase management module, matching module, intelligent naming module, data storage module;

[0132] A metadata database management module, used for establishing and maintaining a metadata database including keyword rules and existing name records;

[0133] A matching module, for extracting keywords from a search request input by a user and searching for matching records in a metadata database using a natural language processing method and a deep learning model;

[0134] The intelligent naming module is used to trigger the intelligent naming process and generate a new name when no matching record is found in the metadata database;

[0135] The data storage module is used to write the newly generated name and related information into the metadata database for storage, and update the metadata database.

[0136] The metadata database management module includes: a metadata database establishment unit, a keyword rule definition unit, a data deduplication unit, and a consistency check unit;

[0137] A metadata database establishment unit, used to initialize the metadata database, including defining the database structure and storage format;

[0138] A keyword rule definition unit, used for defining and updating rules for extracting keywords from user search requests;

[0139] A data deduplication unit is used to check and remove duplicate data before writing the data into the metadata database;

[0140] The consistency check unit is used to ensure the consistency and integrity of the data in the metadata database.

[0141] The matching module includes: natural language processing unit, deep learning model unit, and matching strategy unit;

[0142] Natural language processing unit, used to pre-process the text input by the user, such as word segmentation and part-of-speech tagging;

[0143] Deep learning model unit, based on pre-trained deep learning model, intelligently extracts keywords from user input;

[0144] The matching strategy unit is used to search for records matching the extracted keywords in the metadata database according to the defined keyword rules.

[0145] The intelligent naming module includes: automatic naming rule unit, machine learning algorithm unit, and diversity assessment unit;

[0146] The automatic naming rules unit is used to define and update the rules for generating new names;

[0147] A machine learning algorithm unit, used to generate new names that conform to the rules based on a trained machine learning model;

[0148] The diversity assessment unit is used to conduct diversity assessment on newly generated names to ensure the uniqueness and innovation of the names.

[0149] The data storage module includes: a distributed storage unit, a data writing unit, and a metadata database update unit;

[0150] Distributed storage unit, using distributed storage method to improve the reliability and performance of data storage;

[0151] A data writing unit, used for writing the newly generated name and related information into the metadata database;

[0152] The metadata database update unit is used to update the status and index of the metadata database after the data is written to ensure the real-time and accuracy of the data.

[0153] In summary, the metadata-based functional prototype modeling and naming system realizes the whole process of extracting keywords from user search requests, finding matching records in the metadata database, generating new names, and storing and updating the metadata database through the metadata database management module, matching module, intelligent naming module, data storage module and the units they contain. The various modules and units cooperate with each other to complete the naming task of the system.

[0154] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose and scope of protection of the present invention, and all of these are within the protection of the present invention.

[0155] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. A metadata-based functional prototype modeling naming method, characterized in that: include: Step S1: Establish a metadata database and define keyword rules; Step S2: According to the defined keyword rules, natural language processing methods are used in combination with deep learning models to intelligently extract keywords from the search request input by the user, and matching strategies are used to search for records matching the keywords in the metadata database; Step S3: If a matching record is found in the metadata database, it indicates that the keyword already exists, and the system directly calls the name record corresponding to the keyword. If no matching record is found in the metadata database, it indicates that the keyword is a new search, and the intelligent naming process is triggered. According to the automatic naming rules, combined with the machine learning algorithm and the diversity evaluation method, a new name is generated; Step S4: Use a distributed storage method to write the newly generated name and its related information into the metadata database for storage, and introduce a data deduplication mechanism and a consistency check mechanism to obtain an updated metadata database.

2. The metadata-based functional prototype modeling and naming method according to claim 1, characterized in that: The specific steps of step S2 include: S2.1: receiving a search request input by a user, and determining the form of the search request input by the user; If the search request input by the user is in text form, preprocessing the input search request text; If the search request input by the user is in the form of voice and picture, the non-text input is converted into text using voice recognition and OCR technology, and the converted text is pre-processed; S2.2: Use natural language processing methods to parse the preprocessed search request, and use a deep learning model to intelligently extract keywords based on the defined keyword length and part of speech requirements to form a keyword list; S2.3: Develop keyword matching strategies and determine the priorities and weights of keyword matching strategies; S2.4: accessing the metadata database, using the keyword matching strategy to search for records matching the keyword list in the metadata database, and sorting the query results according to the priority and weight to generate a query result list; S2.5: According to the query result list, the query results are returned to the user and displayed.

3. The metadata-based functional prototype modeling and naming method according to claim 2, characterized in that: The specific steps of S2.2 include: S2.21: obtaining the preprocessed search request text, and performing word segmentation processing, splitting the preprocessed search request text into words or phrases, and obtaining a word segmented text; S2.22: Load the pre-trained part-of-speech tagging model, and use the pre-trained part-of-speech tagging model to perform part-of-speech tagging on the segmented text, assigning a part-of-speech to each word or phrase; S2.23: defining the length of the keyword as L, and based on the defined keyword length, selecting words or phrases that meet the length condition from the text after part-of-speech tagging and segmentation, to obtain preliminary words or phrases; S2.24: defining part-of-speech requirements, and further screening out words or phrases meeting the part-of-speech conditions from the preliminary words or phrases according to the defined part-of-speech requirements, to obtain screened words or phrases; S2.25: Load the pre-trained deep learning model, use the pre-trained deep learning model to score the screened words or phrases, and select the N words or phrases with the highest scores as keywords according to the scoring results; S2.26: Sort the extracted keywords from high to low according to the scores to form a keyword list.

4. The metadata-based functional prototype modeling and naming method according to claim 3, characterized in that: The specific steps of S2.3 include: S2.31: Obtain the keyword list in step S2.2, and classify the keyword list to obtain a classified keyword list; S2.32: According to the classified keyword list, perform exact matching, fuzzy matching, partial matching or synonym matching on different keywords, and determine the priority of each matching strategy according to business requirements and the importance of keywords. At the same time, assign weights to different keywords and matching strategies, and generate matching strategy documents; S2.33: Apply the matching strategy document to actual business, monitor the implementation effect of the strategy, and optimize and adjust the matching strategy based on the monitoring results.

5. The metadata-based functional prototype modeling and naming method according to claim 4, characterized in that: When performing exact matching, fuzzy matching, partial matching or synonym matching on different keywords in S2.32: If the keyword is a professional term or a word with a specific meaning, a string matching algorithm is used to compare the search request entered by the user with the words in the classified keyword list; If the search request entered by the user contains spelling errors or is incomplete, the fuzzy matching algorithm is used to calculate the similarity S between the search request entered by the user and the keywords in the classified keyword list, and a similarity threshold h is set. When S>h, the match is considered successful; If the search request entered by the user is part of the keywords in the classified keyword list, the prefix tree is used to speed up the partial matching process; If the search request input by the user uses a synonym or near-synonym of a keyword, a synonym dictionary is established to associate the keyword with its synonym. When the user input matches a synonym of a word in the keyword list, the match is considered successful.

6. The metadata-based functional prototype modeling and naming method according to claim 5, characterized in that: The specific process of the intelligent naming process in S3 includes: S3.1: Generate candidate names according to a preset automatic naming rule to obtain a candidate name list; S3.2: Based on the candidate name list, select candidate name features, wherein the candidate name features include character type and similarity with existing names; S3.3: Load the pre-built support vector machine model, and use the selected candidate name features to train the pre-built support vector machine model to obtain a trained support vector machine model; S3.4: Input the candidate names to be evaluated into the trained support vector machine model, the support vector machine model outputs evaluation results according to the features of the candidate names to be evaluated, and sorts the candidate names according to the evaluation results; S3.5: Select the candidate name with the highest score as the optimal name according to the sorting result, add the optimal name as the new name to the metadata database, and return the new name to the requester.

7. The metadata-based functional prototype modeling and naming method according to claim 6, characterized in that: In step S1, the metadata database integrates storage, intelligent indexing and rule management functions.

8. A metadata-based functional prototype modeling naming system, which is used to implement the metadata-based functional prototype modeling naming method according to any one of claims 1 to 7, characterized in that: include: Metadatabase management module, matching module, intelligent naming module, data storage module; The metadata database management module is used to establish and maintain a metadata database containing keyword rules and existing name records; The matching module is used to extract keywords from the search request input by the user using a natural language processing method and a deep learning model, and search for matching records in the metadata database; The intelligent naming module is used to trigger the intelligent naming process and generate a new name when no matching record is found in the metadata database; The data storage module is used to write the newly generated name and related information into the metadata database for storage, and to update the metadata database.

9. The metadata-based functional prototype modeling and naming system according to claim 8, characterized in that: The matching module comprises a matching strategy unit, and the matching strategy unit is used to search for records matching the extracted keywords in the metadata database according to defined keyword rules.

10. The metadata-based functional prototype modeling and naming system according to claim 9, characterized in that: The intelligent naming module includes: an automatic naming rule unit and a diversity evaluation unit. The automatic naming rule unit is used to define and update rules for generating new names; and the diversity evaluation unit is used to perform diversity evaluation on the newly generated names.

Images