CN116226297B - Visual search method, system, equipment and storage medium for data model - Google Patents
Visual search method, system, equipment and storage medium for data model Download PDFInfo
- Publication number
- CN116226297B CN116226297B CN202310497197.3A CN202310497197A CN116226297B CN 116226297 B CN116226297 B CN 116226297B CN 202310497197 A CN202310497197 A CN 202310497197A CN 116226297 B CN116226297 B CN 116226297B
- Authority
- CN
- China
- Prior art keywords
- data model
- preset
- similarity
- relation
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a visual search method, a visual search system, visual search equipment and a visual search storage medium for a data model. The data model searching method comprises the following steps: acquiring the relation characteristics of a target data model; acquiring search features; and returning search results according to the target data model when the search features are matched with the relationship features of the target data model. The invention is beneficial to improving the searching speed and the searching accuracy.
Description
Technical Field
The present invention relates to the field of data model management, and in particular, to a method, a system, an apparatus, and a storage medium for visual searching of a data model.
Background
A Data Model (Data Model), which may also be referred to as a Data resource Model, is a description of objective things and their logical relationships, and refers to a Data structure representing entities and databases of relationships between entities, and is a set of strictly defined concepts that can express interrelationships between Data, operations on Data, and definition constraint rules related to semantics. A data model is typically made up of three elements, data structure, data manipulation, and integrity constraints.
The data model describes and expresses the attribute and the relation of the data source entity from multiple angles and multiple levels, and establishes the corresponding relation between the entity and the data model. The good data model can improve the structuring level of the data resource information, support reasonable classification and efficient access of the data resource, realize optimal allocation of the resource and smooth communication and exchange of the information.
In the prior art, in the process of searching a data model, after a user proposes a group of search keywords, a search system traverses a data model base according to the search keywords and returns one or more data models matched with the search keywords as query results. The larger the database model, the more time it takes to traverse, which will take more search time. Moreover, when the number of search keywords is large, the situation that the data model cannot be matched easily occurs; when the number of search keywords is small, a plurality of data models may be obtained by matching, and since it is difficult to determine the data model closest to the user's search intention, a data model not closest to the user's search intention may be returned, resulting in a decrease in search accuracy.
Disclosure of Invention
The invention provides a visual search method, a visual search system, visual search equipment and a visual search storage medium for a data model, which are beneficial to improving search speed and search accuracy.
In a first aspect, a data model searching method is provided, including:
acquiring the relation characteristics of a target data model;
acquiring search features;
and returning search results according to the target data model when the search features are matched with the relation features of the target data model.
In some embodiments, the data model searching method, wherein the obtaining the relationship feature of the target data model includes:
extracting a plurality of features in the target data model;
screening the plurality of features to obtain key features;
according to the key characteristics, determining the similarity relation between the target data model and each data model in a data model library;
and determining the relation characteristic of the target data model according to the similarity relation between the target data model and each data model in the data model library, wherein the relation characteristic is a one-dimensional characteristic vector or a matrix characteristic vector.
In some embodiments, the data model searching method, wherein the screening the plurality of features to obtain the key features includes:
extracting text content in a target data model, and performing word segmentation, part-of-speech tagging and stop word filtering on the text content to obtain a plurality of candidate words;
Respectively calculating first weights of all candidate words, wherein the first weights are in direct proportion to word frequencies and unique values, the word frequencies are the ratio of the number of current candidate words to the total number of candidate words, and the unique values are the ratio of the total number of data models in a data model library to the number of data models containing the current candidate words;
respectively calculating second weights of all candidate words, wherein the second weights are in direct proportion to the number of adjacent data models containing the current candidate word, and the adjacent data models are data models which are directly connected with the target data model in a data model database;
respectively calculating comprehensive weights of all candidate words, wherein the comprehensive weights are in direct proportion to the first weights and the second weights;
and sequencing the candidate words according to the sequence from large to small of the comprehensive weight, and selecting a first preset number of candidate words from front to back as key features according to the sequencing.
In some embodiments, the data model searching method, wherein the determining, according to the key feature, a similarity relationship between the target data model and each data model in the data model library includes:
obtaining a similarity relation conversion table which comprises a plurality of preset data models and relation features corresponding to the preset data models;
Extracting preset characteristics in a preset data model;
calculating the similarity between the key features and preset features to obtain a preset data model with the maximum similarity with the target data model;
and determining the similarity relation between the target data model and each data model in a data model library according to a preset data model with the maximum similarity with the target data model.
In some embodiments, the data model searching method, wherein the constructing of the similarity relation conversion table includes:
acquiring a second preset number of data models from the data model library to obtain a plurality of initial data models;
constructing a relationship feature of an initial data model, and adjusting the relationship feature of the initial data model for a plurality of times until the relationship feature of the initial data model meets a target state or reaches a preset adjustment frequency, wherein in one adjustment process, when the similarity between a first initial data model and a second initial data model is larger than that between the first initial data model and a third initial data model, and the relationship feature distance between the first initial data model and the second initial data model is larger than that between the first initial data model and the third initial data model, the relationship feature distance between the first initial data model and the third initial data model is increased, or the relationship feature distance between the first initial data model and the second initial data model is reduced, wherein the relationship feature distance is the Euclidean distance of the relationship feature in the plurality of initial data models;
Selecting initial data models according to a preset selection proportion or a third preset quantity according to the relation characteristic distribution condition of the initial data models to obtain a plurality of preset data models;
and establishing a similar relation conversion table according to the preset data model and the corresponding relation characteristics.
In some embodiments, the data model searching method, wherein the determining the relationship feature of the target data model according to the similarity relationship between the target data model and each data model in the data model library includes:
acquiring a relationship characteristic corresponding to a preset data model with the maximum similarity with the target data model, and obtaining a first relationship characteristic;
determining a preset data model meeting a first preset requirement according to the similarity between the key feature and each preset feature, and obtaining corresponding relation features of the preset data model meeting the first preset requirement to obtain one or more second relation features, wherein the first preset requirement is that the similarity between the key feature and the preset feature is larger than the first preset similarity, and/or the first preset requirement is that the similarity between the key feature and the preset feature is smaller than the second preset similarity, and the second preset similarity is smaller than the first preset similarity;
And modifying the first relation feature according to the second relation feature, and taking the modified first relation feature as the relation feature of the target data model.
In some embodiments, the data model searching method, wherein when the search feature matches the relationship feature of the target data model, returns a search result according to the target data model, includes:
obtaining a similarity relation conversion table which comprises a plurality of preset data models and relation features corresponding to the preset data models;
extracting preset characteristics in a preset data model;
calculating the similarity between the search feature and a preset feature;
determining a preset data model meeting a second preset requirement according to the similarity between the search feature and each preset feature, and obtaining corresponding relation features of the preset data model meeting the second preset requirement to obtain one or more third relation features, wherein the second preset requirement is that the similarity between the search feature and the preset feature is larger than a third preset similarity, and/or the second preset requirement is that the similarity between the search feature and the preset feature is smaller than a fourth preset similarity, and the third preset similarity is larger than the fourth preset similarity;
Determining the relation features of the search features according to the third relation features;
calculating the similarity between the relation features of the search features and the corresponding relation features of each data model in the data model library;
and when the similarity of the relation features of the target data model and the relation features of the search features is the maximum similarity and is greater than a similarity threshold, matching the search features with the relation features of the target data model, and returning search results according to the target data model.
In a second aspect, a data model retrieval system is provided, including:
the acquisition module is used for acquiring the relation characteristics of the target data model and acquiring the search characteristics;
and the matching module is used for returning search results according to the target data model when the search features are matched with the relation features of the target data model.
In a third aspect, an electronic device is provided comprising a memory and a processor, the memory storing a computer program executable on the processor, wherein the steps of the data model search method as described above are implemented when the processor executes the program.
In a fourth aspect, a storage medium is provided, on which a computer program is stored, wherein the computer program, when being executed by a processor, implements the steps of the data model search method as described above.
The data model searching method can compare the relation features of the data model with the searching features, and when the searching features are matched with the relation features of the target data model, the target data model is obtained through searching. According to the invention, the data model is searched according to the relation characteristics of the data model, and the searching speed and the searching accuracy are facilitated.
Drawings
Various other advantages and benefits of the present invention will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. It is evident that the figures described below are only some embodiments of the invention, from which other figures can be obtained without inventive effort for a person skilled in the art.
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a data model traceability analysis method according to one embodiment of the invention;
FIG. 2 is a schematic diagram of a neighboring data model as a data model according to one embodiment of the invention;
FIG. 3 is a diagram of a distribution of relational features of an initial data model according to one embodiment of the invention;
FIG. 4 is a schematic diagram of a data model search system according to one embodiment of the invention;
FIG. 5 is a diagram of an interactive interface of a data model search system in accordance with one embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Description of the embodiments
It should be noted that, without conflict, the embodiments and features of the embodiments in the present application may be combined with each other, and the present invention will be further described in detail with reference to the drawings and the specific embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used in the description of the applications herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having" and any variations thereof in the description and claims of the present application and in the description of the figures above are intended to cover non-exclusive inclusions. The terms first, second and the like in the description and in the claims or in the above-described figures, are used for distinguishing between different objects and not necessarily for describing a sequential or chronological order.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
In order to better understand the technical solutions of the present application, the following description will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings.
In an application scene of constructing a digital quality management system or a data integration system, reasonable classification, efficient access and optimal configuration of data resources are realized by constructing and managing a data model. In order to realize a rapid searching function of a data model, the embodiment provides a data model traceability analysis method.
Fig. 1 is a flow chart of a data model tracing analysis method in the present embodiment. Referring to fig. 1, the present invention provides a data model searching method, which includes: step 10, step 20, step 30.
And step 10, acquiring the relation characteristics of the target data model.
The target data model is a data model in a database of data models.
The relational features are features that are generated based on relationships between the target data model and other data models in the data model library. For example, the relational features are features that are generated based on similarity relationships between the target data model and other data models.
And step 20, acquiring search characteristics.
The search features are features extracted according to the user input information, and can reflect the search requirements and search intentions of the user.
And step 30, returning a search result according to the target data model when the search feature is matched with the relation feature of the target data model.
And judging whether to search for the needed data model or not according to the matching degree of the relation features of the search features and the target data model. For example, when the similarity between the search feature and the relationship feature is higher and is greater than a preset similarity threshold, the search feature is matched with the relationship feature of the target data model, and the data model corresponding to the relationship feature is returned, namely, the target data model is returned.
The data model searching method can compare the relation features of the data model with the searching features, and when the searching features are matched with the relation features of the target data model, the target data model is obtained through searching. According to the invention, the data model is searched according to the relation characteristics of the data model, and the searching speed and the searching accuracy are facilitated.
In some embodiments, a data model searching method, wherein obtaining a relational feature of a target data model, comprises: step 11, step 12, step 13, step 14.
And step 11, extracting a plurality of characteristics in the target data model.
For example, for text content in the target data model, analyzing the text content, extracting important information in the text content, and obtaining a plurality of features in the target data model.
In addition, the plurality of features in the target data model may also be attribute features of the target data model.
And step 12, screening the plurality of features to obtain key features.
The multiple features are screened, some key features are selected, the operation amount of the subsequent steps can be reduced, and the searching speed is improved.
And 13, determining the similarity relation between the target data model and each data model in the data model library according to the key characteristics.
The similarity relationship is a similarity relationship between the data models. For example, the closer the relationship between two data models, the more similar the two data models.
Optionally, the similarity between the data models is calculated by key features of the data models.
And step 14, determining the relation characteristics of the target data model according to the similarity relation between the target data model and each data model in the data model library.
The relationship features in the target data model represent the degree of tightness of the relationship of the target data model with other data model models, and can be calculated through the similarity. For example, the relational features are constructed based on the similarity between the target data model and each of the data models in the database of data models.
In some application scenarios, the relational feature is a one-dimensional feature vector or a matrix feature vector. In other words, the relational features of the data model are represented by one-dimensional feature vectors or matrix feature vectors. For example, the relationship feature of the target data model is a one-dimensional feature vector, where the one-dimensional feature vector includes a plurality of numbers, and each number represents a similarity between the target data model and each of the other data models in the data model library.
In some embodiments, a data model searching method, wherein screening a plurality of features to obtain key features includes: step 121, step 122, step 123, step 124, step 125.
And 121, extracting text content in the target data model, and performing word segmentation, part-of-speech tagging and filtering stop words on the text content to obtain a plurality of candidate words.
Candidate words are generally words with actual meanings in the text content, and can embody the meaning of the text to a certain extent. Specifically, word segmentation and part-of-speech tagging can be performed on text content, and extraction of candidate words is completed.
In the process of acquiring candidate words, because words with little effect, such as stop words, exist in text content, the words need to be deleted, so that the searching speed is improved. In particular, natural language processing systems may be employed.
Step 122, respectively calculating first weights of the candidate words, wherein the first weights are in direct proportion to word frequencies, the word frequencies are the ratio of the number of the current candidate words to the total number of the candidate words, and the unique values are the ratio of the total number of data models in the data model library to the number of the data models containing the current candidate words.
The candidate word may be a word or a word.
The current candidate word is a candidate word selected from the plurality of candidate words at the time of the current calculation.
The first weight may evaluate the importance of the candidate word to a data model. In general, the importance of a candidate word is proportional to the number of times it appears in the text content, while the importance of a candidate word is inversely proportional to the frequency with which it appears in other data models.
The larger the first weight, the more important the representation of the candidate word is. And selecting important words and representative words as keywords of the data model through word frequency and unique values.
Alternatively, W 1 =f 1 ×f 2
Wherein W is 1 For the first weight, f 1 Is word frequency, f 2 Is a single value.
Step 123, respectively calculating second weights of the candidate words, wherein the second weights are in direct proportion to the number of adjacent data models containing the current candidate word, and the adjacent data models are directly connected with the target data model in the data model database.
The larger the second weight, the more relevant candidate words tend to be obtained with the target data model, and the similarity between directly connected data models is increased.
The adjacent data models have preset relations, and the preset relations can be determined according to requirements. Optionally, the preset relationship is a relationship with a direct relationship, specifically, a direct relationship between the data models means that a direct logical relationship exists between the two data models. Thus, adjacent data models may also be referred to as connected data models, where adjacent data models are closely related.
For example, referring to fig. 2, the first data model 401, the second data model 402, the third data model 403, and the fourth data model 404 are data models with direct connection, and the four data models are adjacent data models. Specifically, the relationship between the parent data model and the child data model may be considered as a direct relationship between the two, for example, if the automobile data model is a parent data model, then the truck data model, the sedan data model, and the bus data model are child data models. It can be seen that there is a direct link between these data models, and that the relationship between them is a kind of adjacency.
In the process of system construction, the logical relations among the data models are stored in the data model library, and the logical relations are directly read.
Alternatively, W 2 =n÷N
Wherein W is 2 For the second weight, N is the number of neighboring data models that contain the current candidate word, and N is the total number of data models.
Step 124, calculating the comprehensive weight of each candidate word, wherein the comprehensive weight is proportional to the first weight and the second weight.
Alternatively, w=αw 1 +βW 2
Wherein W is comprehensive weight, W 1 Is of a first weight, W 2 For the second weight, α is the first coefficient and β is the second coefficient. The first coefficient and the second coefficient are determined according to the actual application scene and are used for adjusting the proportion between the first weight and the second weight.
In other embodiments, the key feature may be selected from the candidate word directly by virtue of the first weight without calculating the second weight.
The above calculation formula is an example, and can be adjusted and modified as required in practical application.
And 125, sequencing the candidate words according to the sequence from large to small of the comprehensive weight, and selecting a first preset number of candidate words from front to back as key features according to the sequencing.
That is, the key features include one or more candidate words.
For example, after the comprehensive weight of each candidate word is obtained, the candidate words are ordered according to the order from high to low of the comprehensive weight, and the first 5 candidate words with larger comprehensive weight are selected to obtain the key features.
In some embodiments, the data model searching method, according to the key features, determines a similarity relationship between the target data model and each data model in the data model library, including: step 131, step 132, step 133, step 134.
Step 131, obtaining a similarity transformation table, wherein the similarity transformation table comprises a plurality of preset data models and relationship features corresponding to the preset data models.
The similarity relation conversion table is used for determining similarity relations between the target data model and other data models, and further, relation features (relation feature vectors) can be generated in a subsequent step.
For example, the similarity relationship conversion table has a plurality of preset data models and a plurality of relationship features, and each preset data model corresponds to a unique relationship feature.
The preset data model is a data model selected from a data model library, and typically, the preset data model is a representative data model in the data model library. Optionally, the data model may be selected from the data model library by a random selection manner or a classification selection manner, so as to obtain a preset data model.
Generally, the number of data models in the data model library is larger, and the number of data models in the similarity relation conversion table is smaller than that of the data models in the data model library, so that the search speed is improved.
And 132, extracting preset characteristics in the preset data model.
The preset features in the preset data model are actually equivalent to the key features, that is, the preset features correspond to the key features of the preset data model. The method for acquiring the preset features of the preset data model can refer to a key feature acquisition method of the target data model.
Optionally, the similarity relation conversion table includes preset features of a preset data model, which can be directly read and used for calculation, thereby being beneficial to improving the searching speed.
And 133, calculating the similarity between the key features and the preset features to obtain a preset data model with the maximum similarity with the target data model.
For example, the similarity between the key features of the target data model and the preset features of each preset data model is sequentially calculated, and the preset data model with the maximum similarity with the target data model is selected.
Optionally, calculating the similarity of the key feature and the preset feature includes: the key features and the preset features are vectorized, and then cosine similarity between the key features and the preset features is calculated. Specifically, word2vec may be employed to vectorize key features and preset features.
And 134, determining the similarity relation between the target data model and each data model in the data model library according to the preset data model with the maximum similarity with the target data model.
It will be appreciated that the similarity transformation table corresponds to a condensed database of data models, and thus, determining the relationship between the target data model and the data models in the similarity transformation table corresponds to determining the similarity between the target data model and each data model in the database of data models, or calculating the similarity between the target data model and each data model in the database of data models.
For example, after the preset data model with the maximum similarity to the target data model (hereinafter referred to as the preset data model with the maximum similarity) is obtained, the preset data model with the maximum similarity can be found in the data model library, and then the similarity relationship between the target data model and each data model in the data model library is determined according to the preset data model with the maximum similarity, for example, the similarity relationship between the preset data model with the maximum similarity and each data model in the data model library is the similarity relationship between the target data model and each data model in the data model library.
In other embodiments, the similarity relationship between the target data model and each data model in the data model library may be determined without using a similarity relationship conversion table. For example, the similarity between the target data model and all the data models in the data model library may be directly calculated, so as to determine the similarity relationship between the target data model and each data model in the data model library.
In some embodiments, the data model searching method, wherein the constructing of the similarity relation conversion table includes: step 151, step 152, step 153, step 154.
Step 151, obtaining a second preset number of data models from the data model library to obtain a plurality of initial data models, and obtaining a plurality of initial data models.
The second preset number is smaller than or equal to the number of data models in the data model library. It is generally determined according to the number of data models in the data model library, and is not suitable for too much or too little.
And 152, constructing a relationship feature of the initial data model, and adjusting the relationship feature of the initial data model for a plurality of times until the relationship feature of the initial data model meets a target state or reaches a preset adjustment frequency, wherein in one adjustment process, when the similarity between the first initial data model and the second initial data model is larger than that between the first initial data model and the third initial data model, and the relationship feature distance between the first initial data model and the second initial data model is larger than that between the first initial data model and the third initial data model, the relationship feature distance between the first initial data model and the third initial data model is increased, or the relationship feature distance between the first initial data model and the second initial data model is reduced, wherein the relationship feature distance is the Euclidean distance of the relationship feature.
The larger the relationship feature distance is, the lower the similarity between the corresponding data models is.
By adjusting the relationship features of the initial data models for a plurality of times, the relationship feature distance between the initial data models with higher similarity can be reduced, and the relationship feature distance between the initial data models with lower similarity can be increased. It can be seen that by making multiple adjustments to the relational features of the initial data models, higher initial data models can be brought together based on the relational features.
The relationship state between the relationship features of the data model can be evaluated by establishing an objective function. Optionally, the smaller the function value of the objective function, the smaller the relationship feature distance between the initial data models indicating higher similarity, and the larger the relationship feature distance between the initial data models indicating lower similarity. The target state is that the relation characteristic distance between the initial data models with higher similarity is smaller than the expected distance, and the relation characteristic distance between the initial data models with lower similarity is larger than the expected distance, and the relation characteristic distance can be determined specifically through whether the target function value is larger than a preset function threshold value.
In some embodiments, in increasing the relational feature distance between the first initial data model and the third initial data model, or decreasing the relational feature distance between the first initial data model and the second initial data model, comprises: when the relation characteristic distance between the first initial data model and the third initial data model is larger than or equal to a first distance threshold value and the relation characteristic distance between the first initial data model and the second initial data model is larger than or equal to a second distance threshold value, reducing the relation distance between the first initial data model and the second initial data model, and keeping the relation characteristic distance between the first initial data model and the third initial data model; and when the relation characteristic distance between the first initial data model and the third initial data model is smaller than or equal to a first distance threshold value and the relation characteristic distance between the first initial data model and the second initial data model is smaller than or equal to a second distance threshold value, maintaining the relation distance between the first initial data model and the second initial data model, and increasing the relation characteristic distance between the first initial data model and the third initial data model. The distance of the initial data model is controlled, so that the situation that the distance of the initial data model is too small is avoided, the initial data models are uniformly distributed, and the searching accuracy is improved.
Step 153, selecting the initial data models according to the relationship feature distribution condition of the initial data models and the preset selection proportion or the third preset quantity, so as to obtain a plurality of preset data models.
The distribution of the relational features of the initial data model is a distribution formed from the relational features. It can be understood that in the adjustment process, the relationship feature distance between the initial data models with higher similarity is continuously reduced, and the relationship feature distance between the initial data models with lower similarity is increased, that is, the relationship feature distribution condition of the initial data models is adjusted.
Optionally, according to the relationship features of the initial data model, the initial data model is displayed in a two-dimensional coordinate system or other multidimensional coordinate systems, and the relationship feature distribution situation is visually displayed, as shown in fig. 3.
Referring to fig. 3, since the relationship features of the initial data models 501 are related to the initial data models 501, in the relationship feature distribution situation of the initial data models 501, the euclidean distance between the initial data models 501 with higher similarity is smaller, the euclidean distance between the initial data models 501 with lower similarity is larger, as shown in fig. 3, the initial data models 501 with different theme types are represented by using graphs with different shapes or filling patterns, the initial data models 501 with the same theme type are clustered together, and the distance between the initial data models 501 with the same theme type is smaller.
The number of the preset data models is smaller than that of the initial data models, so that the calculation amount of the subsequent steps is reduced.
The relation characteristic distribution condition of the preset data model is the same as or similar to that of the initial data model. Specifically, referring to fig. 3, in the relationship feature distribution diagram of the initial data model, an initial data model may be selected at intervals of a preset distance as the preset data model.
In other embodiments, step 153 is not performed, and the initial data model is directly used as the preset data model.
And 154, establishing a similar relation conversion table according to the preset data model and the corresponding relation characteristics.
Optionally, a relationship feature distribution diagram of the preset data model is generated according to the preset data model and the corresponding relationship features, similar to fig. 3. From this point of view, the similarity relationship conversion table corresponds to a relationship feature distribution pattern of the preset data model.
In some embodiments, the data model searching method, wherein determining the relationship feature of the target data model according to the similarity relationship between the target data model and each data model in the data model library, comprises: step 141, step 142, step 144.
And 141, acquiring the relation characteristic corresponding to the preset data model with the maximum similarity with the target data model to obtain a first relation characteristic.
And directly reading the relation characteristic corresponding to the preset data model with the maximum similarity from the similarity relation conversion table as a first relation characteristic.
Step 142, determining a preset data model meeting a first preset requirement according to the similarity between the key feature and each preset feature, and obtaining the corresponding relation feature of the preset data model meeting the first preset requirement to obtain one or more second relation features, wherein the first preset requirement is that the similarity between the key feature and the preset feature is greater than the first preset similarity, and/or the first preset requirement is that the similarity between the key feature and the preset feature is smaller than the second preset similarity, and the second preset similarity is smaller than the first preset similarity.
The preset data model is screened according to the first preset requirement, so that the preset data model meeting the first preset requirement is obtained, the operand of the subsequent step can be reduced, and the searching accuracy is guaranteed.
Specifically, according to the similarity between the target data model and the preset data model, a preset data model with the similarity larger than the first preset similarity and a preset data model with the similarity smaller than the second preset similarity are selected.
And according to the screened preset data models, the relation features corresponding to the preset data models meeting the first preset requirements are read from the similarity relation conversion table, so that a plurality of second relation features are obtained.
And 144, modifying the first relation feature according to the second relation feature, and taking the modified first relation feature as the relation feature of the target data model.
And modifying the second relation feature on the basis of the first relation feature, so that the relation feature with the similarity being close to that of the relation feature being larger than the first preset similarity and/or the relation feature with the similarity being far from that of the relation feature being smaller than the second preset similarity is obtained after modification. That is, the distance between the modified obtained relationship feature and the relationship feature with the similarity larger than the first preset relationship feature is smaller than the distance between the first relationship feature before the modification and the relationship feature with the similarity larger than the first preset relationship feature, and the distance between the modified obtained relationship feature with the similarity smaller than the second preset relationship feature is larger than the distance between the first relationship feature before the modification and the relationship feature with the similarity smaller than the second preset relationship feature.
Optionally, modifying the first relationship feature according to the similarity between the second relationship feature and the key feature to obtain the relationship feature of the target data model, wherein the similarity between the second relationship feature and the key feature is used as the modified distance weight value. The higher the similarity between the second relational feature and the key feature, the smaller the distance between the relational feature of the target data model and the second relational feature.
Alternatively, d=d×r -a
Wherein D is the distance between the relation feature and the second relation feature of the target data model, D is the distance between the first relation feature and the second relation feature, R is the similarity between the key feature of the target data model and the preset feature of the corresponding preset data model, a is the preset coefficient, and a is equal to or greater than 1.
In some embodiments, a data model search method, wherein returning search results from a target data model when the search features match relationship features of the target data model, comprises: step 301, step 302, step 303, step 304, step 305, step 306, step 307.
Step 301, obtaining a similarity transformation table, wherein the similarity transformation table comprises a plurality of preset data models and relationship features corresponding to the preset data models.
The similarity relation conversion table is stored in the search system in advance, and can be directly called when searching is performed. The similarity relation conversion table used in the matching process and the similarity relation conversion table used when the search feature of the target data model is acquired may be the same similarity relation conversion table.
Step 302, extracting preset features in a preset data model.
The preset features can directly exist in the similarity relation conversion table and correspond to the preset data models one by one. And when the method is executed, the preset characteristics in the preset data model are directly extracted from the similarity relation conversion table.
Step 303, calculating the similarity between the search feature and the preset feature.
Optionally, the calculated similarity method includes obtaining a search feature vector and a preset feature vector by vectorizing the search feature and the preset feature; and then, calculating cosine similarity between the search feature vector and the preset feature vector to obtain similarity between the search feature and the preset feature.
Step 304, determining a preset data model meeting a second preset requirement according to the similarity between the search feature and each preset feature, and obtaining the corresponding relation feature of the preset data model meeting the second preset requirement to obtain one or more third relation features, wherein the second preset requirement is that the similarity between the search feature and the preset feature is greater than the third preset similarity, and/or the second preset requirement is that the similarity between the search feature and the preset feature is less than the fourth preset similarity, and the third preset similarity is greater than the fourth preset similarity.
The second preset requirement may be the same as or different from the first preset requirement. By screening out some representative preset data models, the operation speed can be improved on the premise of ensuring the searching accuracy.
Step 305, determining the relation features of the search feature according to the third relation feature.
For example, the similarity between the third relationship feature and the search feature is used as a distance weight value, and a new relationship feature is synthesized to obtain the relationship feature of the search feature.
For example, g= (Σ (-1) k ×r i ×G i )/n
G is the relational feature of the search feature, Σ (-1) k ×r i ×G i Representation pair (-1) k ×r i ×G i Summing, r i G for searching similarity of the feature and the ith third relation feature i For the ith third relation feature, n is the total number of the third relation features, k is 1 or 2, 1 is taken when the similarity between the search feature and the preset feature is smaller than the fourth preset similarity k, and 2 is taken when the similarity between the search feature and the preset feature is larger than the third preset similarity k.
The relationship features that determine the search features are not limited to the above-described method. For example, the relationship feature having the greatest similarity in the similar relationship conversion table may be used as the relationship feature of the search feature, or the relationship feature of the search feature may be determined by a method similar to that in steps 141, 142, and 143.
And 306, calculating the similarity between the relation features of the search features and the relation features corresponding to each data model in the data model library.
And searching a data model closest to the relation features of the search features from the data model library through relation feature similarity calculation. Since the relational features are usually in the form of vectors, similarity is easy to calculate, which is beneficial to realizing quick search.
In some embodiments, the relational features may represent the positions of the data models, so that the data model closest to the relational features of the search features may be searched in a manner of gradually narrowing the search range based on the position features of the relational features, and all the relational features of the data models are not required to participate in similarity calculation.
Step 307, when the similarity between the relationship feature of the target data model and the relationship feature of the search feature is the maximum similarity and is greater than the similarity threshold, the relationship feature of the search feature and the target data model is matched, and the search result is returned according to the target data model.
The search features can reflect the intention of the user, vectorize the search features, and search the data model with the largest relation feature with the search features according to the similarity. And when the data model with the largest relation characteristic with the search characteristic is the target data model, the relation characteristic of the search characteristic and the target data model is matched, and the target data model is used as a search result and returned to the user. The similarity threshold can filter the data model below the threshold, and error search results are prevented from being returned.
The data model searching method can realize a visual searching process and a visual result. For example, according to the search information, the relationship features are generated, and the relationship features are visually displayed in the distribution diagram, so that a user can conveniently determine the difference between the search intention and the search result through the distance between the data models.
Fig. 4 is a schematic structural diagram of a data model search system 60 according to the present embodiment, and as shown in fig. 4, the data model search system 60 includes: an acquisition module 601 and a matching module 602.
An obtaining module 601, configured to obtain a relationship feature of the target data model, and obtain a search feature;
and a matching module 602, configured to return a search result according to the target data model when the search feature matches the relationship feature of the target data model.
The data model search system 60 of the present invention is capable of comparing the relationship features of the data model with the search features, and searching for a target data model when the search features match the relationship features of the target data model. According to the invention, the data model is searched according to the relation characteristics of the data model, and the searching speed and the searching accuracy are facilitated.
In some implementations, the acquisition module 601 is further configured to extract a plurality of features in the target data model; screening the multiple features to obtain key features; according to the key characteristics, determining the similarity relation between the target data model and each data model in the data model library; and determining the relation characteristic of the target data model according to the similar relation between the target data model and each data model in the data model library, wherein the relation characteristic is a one-dimensional characteristic vector or a matrix characteristic vector.
In some embodiments, the obtaining module 601 is further configured to extract text content in the target data model, perform word segmentation, part-of-speech tagging, and filter stop words on the text content, to obtain a plurality of candidate words; respectively calculating first weights of all candidate words, wherein the first weights are in direct proportion to word frequencies and unique values, the word frequencies are the ratio of the number of the current candidate words to the total number of the candidate words, and the unique values are the ratio of the total number of data models in a data model library to the number of data models containing the current candidate words; respectively calculating second weights of the candidate words, wherein the second weights are in direct proportion to the number of adjacent data models containing the current candidate word, and the adjacent data models are data models which are directly connected with the target data model in the data model base; respectively calculating the comprehensive weight of each candidate word, wherein the comprehensive weight is in direct proportion to the first weight and the second weight; and sequencing the candidate words according to the sequence from large to small of the comprehensive weight, and selecting a first preset number of candidate words from front to back according to the sequencing as key features.
In some embodiments, the obtaining module 601 is further configured to obtain a similarity transformation table, where the similarity transformation table includes a plurality of preset data models and relationship features corresponding to the preset data models; extracting preset characteristics in a preset data model; calculating the similarity between the key features and preset features to obtain a preset data model with the maximum similarity with the target data model; and determining the similarity relation between the target data model and each data model in the data model library according to the preset data model with the maximum similarity with the target data model.
In some embodiments, the obtaining module 601 is further configured to obtain a second preset number of data models from the data model library, so as to obtain a plurality of initial data models; constructing a relationship feature of the initial data model, and adjusting the relationship feature of the initial data model for a plurality of times until the relationship feature of the initial data model meets a target state or reaches a preset adjustment frequency, wherein in one adjustment process, when the similarity between the first initial data model and the second initial data model is larger than the similarity between the first initial data model and the third initial data model, and the relationship feature distance between the first initial data model and the second initial data model is larger than the relationship feature distance between the first initial data model and the third initial data model, the relationship feature distance between the first initial data model and the third initial data model is increased, or the relationship feature distance between the first initial data model and the second initial data model is reduced, wherein the relationship feature distance is the Euclidean distance of the relationship feature in the plurality of initial data models; selecting initial data models according to a preset selection proportion or a third preset quantity according to the relation characteristic distribution condition of the initial data models to obtain a plurality of preset data models; and establishing a similar relation conversion table according to the preset data model and the corresponding relation characteristics.
In some embodiments, the obtaining module 601 is further configured to obtain a relationship feature corresponding to a preset data model with a maximum similarity to the target data model, so as to obtain a first relationship feature; determining a preset data model meeting a first preset requirement according to the similarity between the key feature and each preset feature, and obtaining corresponding relation features of the preset data model meeting the first preset requirement to obtain one or more second relation features, wherein the first preset requirement is that the similarity between the key feature and the preset feature is greater than the first preset similarity, and/or the first preset requirement is that the similarity between the key feature and the preset feature is smaller than the second preset similarity, and the second preset similarity is smaller than the first preset similarity; and modifying the first relation feature according to the second relation feature, and taking the modified first relation feature as the relation feature of the target data model.
In some embodiments, the matching module 602 is further configured to obtain a similarity transformation table, where the similarity transformation table includes a plurality of preset data models and relationship features corresponding to the preset data models; extracting preset characteristics in a preset data model; calculating the similarity between the search feature and the preset feature; determining a preset data model meeting second preset requirements according to the similarity between the search features and each preset feature, and obtaining corresponding relation features of the preset data model meeting the second preset requirements to obtain one or more third relation features, wherein the second preset requirements are that the similarity between the search features and the preset features is greater than the third preset similarity, and/or the second preset requirements are that the similarity between the search features and the preset features is smaller than the fourth preset similarity, and the third preset similarity is greater than the fourth preset similarity; determining the relation features of the search features according to the third relation features; calculating the similarity between the relation features of the search features and the corresponding relation features of each data model in the data model library; and when the similarity of the relation features of the target data model and the relation features of the search features is the maximum similarity and is larger than a similarity threshold, matching the relation features of the search features and the target data model, and returning a search result according to the target data model.
FIG. 5 is a diagram of an interactive interface of a data model search system. In fig. 5, the data model is also referred to as a data resource model. The user can input corresponding search information in the interactive interface, the system processes the search information to form search features, and then the search of the data model is completed, wherein the model codes are the relationship features.
The data model searching system can realize visual searching process and result display. And issuing query conditions in the system, generating relationship features, and visually displaying the relationship features in the distribution diagram, so that a user can conveniently determine the difference between the search intention and the search result through the distance between the data models.
The operation principle of the data model search system of this embodiment may refer to the above data model search method, and will not be described herein.
Fig. 6 is a schematic structural diagram of an electronic device according to the present invention. The electronic device comprises a memory 701 and a processor 702, the memory 701 storing a computer program executable on the processor 702, wherein the steps of the data model search method as described above are implemented when the processor 702 executes the program.
The electronic device comprises a memory 701, a processor 702, which are communicatively connected to each other via a system bus 703. It should be noted that only electronic devices having components 701-703 are shown in the figures, but it is understood that not all of the illustrated components are required to be implemented and that more or fewer components may be implemented instead. It will be appreciated by those skilled in the art that the electronic device herein is an electronic device capable of automatically performing numerical calculations and/or information processing according to predetermined or stored instructions, and the hardware thereof includes, but is not limited to, microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASICs), programmable gate arrays (fields-Programmable Gate Array, FPGAs), digital processors (Digital Signal Processor, DSPs), embedded devices, etc.
The electronic device may be a computing device such as a desktop computer, a notebook computer, a palm computer, a cloud server, and the like. The device can perform man-machine interaction with a user through a keyboard, a mouse, a remote controller, a touch pad or a voice control device.
The memory 701 includes at least one type of readable storage medium including flash memory, hard disk, multimedia card, card memory (e.g., SD or DX memory, etc.), random Access Memory (RAM), static Random Access Memory (SRAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), programmable Read Only Memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 701 may be an internal storage module of the device, such as a hard disk or memory of the device. In other embodiments, the memory 701 may also be an external storage device of the device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash Card (Flash Card), etc. that are provided on the device. Of course, the memory 701 may also include both the internal memory module of the device and its external memory device. In this embodiment, the memory 701 is typically used to store an operating system and various types of application software installed on the device. In addition, the memory 701 can also be used to temporarily store various types of data that have been output or are to be output.
The processor may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor is typically used to control the overall operation of the device. In this embodiment, the processor is configured to execute computer readable instructions or process data stored in the memory.
The present invention provides a storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the data model search method as described above.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk), including several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method of the embodiments of the present application.
The data integration system is developed according to the data model searching method, the system is tested, the response time can reach millisecond level, the average response time is 100ms, and the requirement that the response time proposed by a customer is not more than 3s is met.
It is apparent that the embodiments described above are only some embodiments of the present application, but not all embodiments, the preferred embodiments of the present application are given in the drawings, but not limiting the patent scope of the present application. This application may be embodied in many different forms, but rather, embodiments are provided in order to provide a more thorough understanding of the present disclosure. Although the present application has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications may be made to the embodiments described in the foregoing, or equivalents may be substituted for elements thereof. All equivalent structures made by the specification and the drawings of the application are directly or indirectly applied to other related technical fields, and are also within the protection scope of the application.
Claims (7)
1. A data model search method, comprising:
Acquiring the relation characteristics of a target data model;
acquiring search features;
returning a search result according to the target data model when the search feature is matched with the relation feature of the target data model;
the obtaining the relation characteristic of the target data model comprises the following steps:
extracting a plurality of features in the target data model;
screening the plurality of features to obtain key features;
according to the key characteristics, determining the similarity relation between the target data model and each data model in a data model library;
determining the relation characteristic of the target data model according to the similarity relation between the target data model and each data model in a data model library, wherein the relation characteristic is a one-dimensional characteristic vector or a matrix characteristic vector;
and determining the similarity relationship between the target data model and each data model in the data model library according to the key characteristics, wherein the method comprises the following steps of:
obtaining a similarity relation conversion table, wherein the similarity relation conversion table comprises a plurality of preset data models and relation features corresponding to the preset data models;
extracting preset features in the preset data model;
Calculating the similarity between the key features and preset features to obtain a preset data model with the maximum similarity with the target data model;
according to a preset data model with the maximum similarity with the target data model, determining the similarity relationship between the target data model and each data model in a data model library;
the construction of the similarity relation conversion table comprises the following steps:
acquiring a second preset number of data models from the data model library to obtain a plurality of initial data models;
constructing a relationship feature of the initial data model, and adjusting the relationship feature of the initial data model for multiple times until the relationship feature of the initial data model meets a target state or reaches preset adjustment times, wherein in the one-time adjustment process, when the similarity between a first initial data model and a second initial data model is larger than that between the first initial data model and a third initial data model, and the relationship feature distance between the first initial data model and the second initial data model is larger than that between the first initial data model and the third initial data model, the relationship feature distance between the first initial data model and the third initial data model is increased, or the relationship feature distance between the first initial data model and the second initial data model is reduced, wherein the relationship feature distance is the Euclidean distance of the relationship feature;
Selecting initial data models according to the relation characteristic distribution condition of the initial data models and the preset selection proportion or the third preset quantity to obtain a plurality of preset data models;
and establishing a similar relation conversion table according to the preset data model and the corresponding relation characteristics.
2. The method of claim 1, wherein the screening the plurality of features to obtain key features comprises:
extracting text content in the target data model, and performing word segmentation, part-of-speech tagging and filtering stop words on the text content to obtain a plurality of candidate words;
respectively calculating first weights of the candidate words, wherein the first weights are in direct proportion to word frequencies and unique values, the word frequencies are the ratio of the number of the current candidate words to the total number of the candidate words, and the unique values are the ratio of the total number of data models in a data model library to the number of data models containing the current candidate words;
respectively calculating second weights of the candidate words, wherein the second weights are in direct proportion to the number of adjacent data models containing the current candidate word, and the adjacent data models are data models which are directly connected with the target data model in a data model library;
Respectively calculating comprehensive weights of the candidate words, wherein the comprehensive weights are proportional to the first weights and the second weights;
and sequencing the candidate words according to the sequence from large to small of the comprehensive weight, and selecting a first preset number of candidate words from front to back as key features according to the sequencing.
3. The data model searching method according to claim 1, wherein the determining the relationship feature of the target data model according to the similarity relationship between the target data model and each data model in the data model library comprises:
acquiring a relationship characteristic corresponding to a preset data model with the maximum similarity with the target data model, and obtaining a first relationship characteristic;
determining the preset data model meeting a first preset requirement according to the similarity between the key feature and each preset feature, and obtaining the relation feature corresponding to the preset data model meeting the first preset requirement to obtain one or more second relation features, wherein the first preset requirement is that the similarity between the key feature and the preset feature is larger than a first preset similarity, and/or the first preset requirement is that the similarity between the key feature and the preset feature is smaller than a second preset similarity, and the second preset similarity is smaller than the first preset similarity;
And modifying the first relation feature according to the second relation feature, and taking the modified first relation feature as the relation feature of the target data model.
4. The data model search method of claim 1, wherein the returning search results from the target data model when the search features match the relationship features of the target data model comprises:
obtaining a similarity relation conversion table which comprises a plurality of preset data models and relation features corresponding to the preset data models;
extracting preset features in the preset data model;
calculating the similarity between the search feature and the preset feature;
determining the preset data model meeting a second preset requirement according to the similarity between the search feature and each preset feature, and obtaining the relation feature corresponding to the preset data model meeting the second preset requirement to obtain one or more third relation features, wherein the second preset requirement is that the similarity between the search feature and the preset feature is larger than a third preset similarity, and/or the second preset requirement is that the similarity between the search feature and the preset feature is smaller than a fourth preset similarity, and the third preset similarity is larger than the fourth preset similarity;
Determining the relation features of the search features according to the third relation features;
calculating the similarity between the relation features of the search features and the corresponding relation features of each data model in the data model library;
and when the similarity of the relation features of the target data model and the relation features of the search features is the maximum similarity and is greater than a similarity threshold, matching the search features with the relation features of the target data model, and returning search results according to the target data model.
5. A data model retrieval system, comprising:
the acquisition module is used for acquiring the relation characteristics of the target data model and acquiring the search characteristics;
the matching module is used for returning a search result according to the target data model when the search feature is matched with the relation feature of the target data model;
the obtaining the relation characteristic of the target data model comprises the following steps:
extracting a plurality of features in the target data model;
screening the plurality of features to obtain key features;
according to the key characteristics, determining the similarity relation between the target data model and each data model in a data model library;
Determining the relation characteristic of the target data model according to the similarity relation between the target data model and each data model in a data model library, wherein the relation characteristic is a one-dimensional characteristic vector or a matrix characteristic vector;
and determining the similarity relationship between the target data model and each data model in the data model library according to the key characteristics, wherein the method comprises the following steps of:
obtaining a similarity relation conversion table, wherein the similarity relation conversion table comprises a plurality of preset data models and relation features corresponding to the preset data models;
extracting preset features in the preset data model;
calculating the similarity between the key features and preset features to obtain a preset data model with the maximum similarity with the target data model;
according to a preset data model with the maximum similarity with the target data model, determining the similarity relationship between the target data model and each data model in a data model library;
the construction of the similarity relation conversion table comprises the following steps:
acquiring a second preset number of data models from the data model library to obtain a plurality of initial data models;
constructing a relationship feature of the initial data model, and adjusting the relationship feature of the initial data model for multiple times until the relationship feature of the initial data model meets a target state or reaches preset adjustment times, wherein in the one-time adjustment process, when the similarity between a first initial data model and a second initial data model is larger than that between the first initial data model and a third initial data model, and the relationship feature distance between the first initial data model and the second initial data model is larger than that between the first initial data model and the third initial data model, the relationship feature distance between the first initial data model and the third initial data model is increased, or the relationship feature distance between the first initial data model and the second initial data model is reduced, wherein the relationship feature distance is the Euclidean distance of the relationship feature;
Selecting initial data models according to the relation characteristic distribution condition of the initial data models and the preset selection proportion or the third preset quantity to obtain a plurality of preset data models;
and establishing a similar relation conversion table according to the preset data model and the corresponding relation characteristics.
6. An electronic device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that the steps of the data model search method according to any one of claims 1 to 4 are implemented when the processor executes the program.
7. A storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the data model search method according to any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310497197.3A CN116226297B (en) | 2023-05-05 | 2023-05-05 | Visual search method, system, equipment and storage medium for data model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310497197.3A CN116226297B (en) | 2023-05-05 | 2023-05-05 | Visual search method, system, equipment and storage medium for data model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116226297A CN116226297A (en) | 2023-06-06 |
| CN116226297B true CN116226297B (en) | 2023-07-25 |
Family
ID=86569768
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310497197.3A Active CN116226297B (en) | 2023-05-05 | 2023-05-05 | Visual search method, system, equipment and storage medium for data model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116226297B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110990685A (en) * | 2019-10-12 | 2020-04-10 | 中国平安财产保险股份有限公司 | Voice search method, voice search device, voice search storage medium and voice search device based on voiceprint |
| CN114330703A (en) * | 2021-11-15 | 2022-04-12 | 腾讯科技（深圳）有限公司 | Method, device and equipment for updating search model and computer-readable storage medium |
| CN114911917A (en) * | 2022-07-13 | 2022-08-16 | 树根互联股份有限公司 | Asset meta-information searching method and device, computer equipment and readable storage medium |
| CN115563394A (en) * | 2022-11-24 | 2023-01-03 | 腾讯科技（深圳）有限公司 | Search recall method, recall model training method, device and computer equipment |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109189824B (en) * | 2018-08-10 | 2022-04-26 | 创新先进技术有限公司 | Method and device for retrieving similar articles |
| CN110502548A (en) * | 2019-07-26 | 2019-11-26 | 视联动力信息技术股份有限公司 | A kind of search result recommended method, device and computer readable storage medium |
| WO2021074990A1 (en) * | 2019-10-16 | 2021-04-22 | 三菱電機株式会社 | Search device, search method, search program, and learning model search system |
| CN111177317A (en) * | 2019-12-20 | 2020-05-19 | 吕梁学院 | Literature theory rapid retrieval query system and method |
| CN111782898B (en) * | 2020-07-07 | 2024-05-24 | 华青融天(北京)软件股份有限公司 | Data source searching method and device and electronic equipment |
| CN112183484B (en) * | 2020-10-30 | 2023-08-29 | 北京百度网讯科技有限公司 | Image processing method, device, equipment and storage medium |
| CN112269792B (en) * | 2020-12-11 | 2021-07-02 | 腾讯科技（深圳）有限公司 | Data query method, device, equipment and computer readable storage medium |
| CN112308055B (en) * | 2020-12-30 | 2021-11-02 | 北京沃东天骏信息技术有限公司 | Evaluation method and device of face retrieval system, electronic equipment and storage medium |
| CN112632395A (en) * | 2020-12-31 | 2021-04-09 | 深圳追一科技有限公司 | Search recommendation method and device, server and computer-readable storage medium |
| CN113282832A (en) * | 2021-06-10 | 2021-08-20 | 北京爱奇艺科技有限公司 | Search information recommendation method and device, electronic equipment and storage medium |
| CN115496865B (en) * | 2022-11-18 | 2023-03-14 | 南京智程信息科技有限公司 | Similar model searching method, device and storage medium |
-
2023
- 2023-05-05 CN CN202310497197.3A patent/CN116226297B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110990685A (en) * | 2019-10-12 | 2020-04-10 | 中国平安财产保险股份有限公司 | Voice search method, voice search device, voice search storage medium and voice search device based on voiceprint |
| CN114330703A (en) * | 2021-11-15 | 2022-04-12 | 腾讯科技（深圳）有限公司 | Method, device and equipment for updating search model and computer-readable storage medium |
| CN114911917A (en) * | 2022-07-13 | 2022-08-16 | 树根互联股份有限公司 | Asset meta-information searching method and device, computer equipment and readable storage medium |
| CN115563394A (en) * | 2022-11-24 | 2023-01-03 | 腾讯科技（深圳）有限公司 | Search recall method, recall model training method, device and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116226297A (en) | 2023-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109190023B (en) | Collaborative recommendation method and device and terminal equipment | |
| CN110866181B (en) | Resource recommendation method, device and storage medium | |
| JP7403605B2 (en) | Multi-target image text matching model training method, image text search method and device | |
| CN110728313B (en) | Classification model training method and device for intention classification recognition | |
| CN107818491A (en) | Electronic installation, Products Show method and storage medium based on user's Internet data | |
| CN111553556A (en) | Business data analysis method and device, computer equipment and storage medium | |
| CN108304381B (en) | Entity edge establishing method, device and equipment based on artificial intelligence and storage medium | |
| CN111428486B (en) | Article information data processing method, device, medium and electronic equipment | |
| CN114706985A (en) | Text classification method and device, electronic equipment and storage medium | |
| CN116933130A (en) | Enterprise industry classification method, system, equipment and medium based on big data | |
| CN116226297B (en) | Visual search method, system, equipment and storage medium for data model | |
| CN111639161A (en) | System information processing method, apparatus, computer system and medium | |
| CN113988878B (en) | Graph database technology-based anti-fraud method and system | |
| CN113761379B (en) | Commodity recommendation method and device, electronic equipment and medium | |
| CN115544214A (en) | Event processing method and device and computer readable storage medium | |
| CN115631370A (en) | Identification method and device of MRI (magnetic resonance imaging) sequence category based on convolutional neural network | |
| CN115358397A (en) | Parallel graph rule mining method and device based on data sampling | |
| CN114398877A (en) | Theme extraction method and device based on artificial intelligence, electronic equipment and medium | |
| CN109787784B (en) | Group recommendation method and device, storage medium and computer equipment | |
| US11366833B2 (en) | Augmenting project data with searchable metadata for facilitating project queries | |
| CN116501841B (en) | Fuzzy query method, system and storage medium for data model | |
| Li et al. | Hierarchical representation and similarity assessment for reuse of sheet metal parts | |
| CN114238583B (en) | Natural language processing method, device, computer equipment and storage medium | |
| CN111259209A (en) | User intention prediction method based on artificial intelligence, electronic device and storage medium | |
| CN114840686B (en) | Knowledge graph construction method, device, equipment and storage medium based on metadata |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |