A data-idiom-based internet search ranking method and system

Abstract

The application discloses a data-communicative-based cyber search sorting method and system, which comprises the following steps: determining a search result data set matched with a search text used for describing a target digital object to be searched from a registration table of a cyber network distributed in multiple scattered data spaces; determining the query similarity of the metadata of each digital object in the search result data set to the search text, and determining the pragmatic weight value of each digital object in the search result data set; determining the search result evaluation value of each digital object in the search result data set according to the query similarity and the pragmatic weight value of each digital object in the search result data set, so as to obtain the sorting result of the digital objects in the search result data set. The application is based on the premise that the data in the data space is not out of the domain, so that the data sorting in the cyber search process is matched with the actual data use relationship, and the precision rate of the cyber search and the user satisfaction are improved.

Description

Technical Field

[0001] This application belongs to the field of data network search, specifically relating to a data network search ranking method, system, device, and storage medium based on data usage. Background Technology

[0002] Various industries and organizations store data locally, forming data spaces. Data within a data space can be shared and circulated, but data between data spaces must adhere to a security requirement of not leaving the domain. Based on the data network, data can be encapsulated into Digital Objects (DOs), supporting data retrieval and access through DO identifiers and metadata. Querying DOs relies on the interaction between the DO Registry and the DORP (Digital Object Repository), with data users completing search requests through the DOIP (Digital Object Interface Protocol). The security requirement of data remaining within the domain renders traditional search engine frameworks based on data aggregation, indexing, and retrieval processes unsuitable.

[0003] Current data network search systems primarily rely on inverted indexes and query similarity calculations to filter search results.

[0004] However, under the current security requirement of not leaving the domain, the process of retrieving data objects suffers from insufficient retrieval accuracy, which affects the user experience. Summary of the Invention

[0005] This application aims to provide a data-based network search and sorting method, system, device, and storage medium that, while meeting the requirement that data in each data space does not leave its domain, solves the problems of insufficient accuracy and poor search experience in data object retrieval.

[0006] In a first aspect, embodiments of this application disclose a data network search and ranking method based on data pragmatics, including:

[0007] From the registry of a data network distributed across multiple dispersed data spaces, a dataset of search results matching the search text used to describe the target digital object to be searched is determined; the dataset of search results includes a first digital object matching the search text and a second digital object pragmatically associated with the first digital object;

[0008] Determine the query similarity of the metadata of each digital object in the retrieval result dataset to the retrieval text, and determine the pragmatic weight value of each digital object in the retrieval result dataset;

[0009] Based on the query similarity and pragmatic weight values ​​of each digital object in the retrieval result dataset, a retrieval result evaluation value is determined for each digital object in the retrieval result dataset, so as to obtain a ranking result for the digital objects in the retrieval result dataset.

[0010] Secondly, embodiments of this application also disclose a data network search and ranking system based on data pragmatics, comprising:

[0011] The dataset module is used to determine a dataset of search results that matches the search text used to describe the target digital object to be searched, from a registry of a data network distributed in multiple dispersed data spaces; the dataset of search results includes a first digital object that matches the search text, and a second digital object that is pragmatically associated with the first digital object;

[0012] The parameter calculation module is used to determine the query similarity of the metadata of each digital object in the retrieval result dataset to the retrieval text, and to determine the pragmatic weight value of each digital object in the retrieval result dataset.

[0013] The evaluation and ranking module is used to determine the evaluation value of each digital object in the search results dataset based on the query similarity and pragmatic weight value of each digital object in the search results dataset, so as to obtain the ranking result of the digital objects in the search results dataset.

[0014] Thirdly, embodiments of this application also disclose an electronic device, including a processor and a memory, wherein the memory stores a program or instructions that can run on the processor, and the program or instructions, when executed by the processor, implement the steps of the method described in the first aspect.

[0015] Fourthly, embodiments of this application also disclose a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the method described in the first aspect.

[0016] In summary, by extending the retrieval result dataset, composed of digital objects obtained from different data spaces, to include not only the first digital object directly matching the retrieval text but also the second digital object pragmatically related to the first digital object, the search results are no longer limited to text matching. Instead, based on the pragmatic relationships between digital objects stored in different data spaces, a more comprehensive search coverage is achieved without compromising the domain of data within each data space, thus improving the contextual matching degree of the data. Furthermore, by calculating the query similarity of each digital object to evaluate its direct relevance to the retrieval text, the retrieval results are ensured to prioritize returning pairs with a high degree of matching with the query text. This approach avoids situations where search results deviate from users' actual needs. Furthermore, it introduces a pragmatic weighting evaluation mechanism, calculating weights based on the pragmatic relevance between digital objects. This breaks through the limitations of existing search ranking algorithms that rely solely on text matching or webpage link analysis, enabling search results to reflect the usability of digital objects in the data space. This not only optimizes the search ranking logic but also makes the ranking results more aligned with user needs, improving search accuracy and information retrieval efficiency. Ultimately, the ranking of search results is calculated by comprehensively utilizing query similarity and pragmatic weight values, ensuring that the final ranking reflects the relevance of data objects and the importance of application scenarios, thus improving the precision and user satisfaction of the data network search system. Therefore, the method based on this application's embodiments ensures that data ranking matches actual application relationships, thereby improving the user's search experience in a distributed data network storage environment. By optimizing search accuracy, it not only improves the precision of the search system but also effectively reduces search errors caused by unreasonable ranking, enhancing the convenience for users to obtain key information. Attached Figure Description

[0017] In the attached diagram:

[0018] Figure 1 This is a flowchart illustrating the steps of a data network search and ranking method based on data usage, as provided in an embodiment of this application.

[0019] Figure 2 This is a flowchart illustrating the steps of another data-based network search and ranking method provided in this application embodiment;

[0020] Figure 3 This refers to the process of generating a search result dataset under the embodiments of this application;

[0021] Figure 4 This is a process for calculating a pragmatic weight value in the embodiments of this application;

[0022] Figure 5 This is a complete sorting process under the embodiments of this application;

[0023] Figure 6 This is a block diagram of a data network search and sorting device based on data usage, provided in an embodiment of this application.

[0024] Figure 7 This is a block diagram of an electronic device provided in one embodiment of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, without limiting the number of objects; for example, a first object can be one or more. Furthermore, in this application, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects have an "or" relationship.

[0027] In this application, considering that data is not only an isolated unit of information storage, but also an interconnected digital object with data pragmatics, therefore, Figure 1 The image shows a data network search and sorting method based on data usage provided in an embodiment of this application.

[0028] The method may include the following steps:

[0029] Step 101: Determine the dataset of search results that matches the search text used to describe the target digital object to be searched from the registry of the data network, which is distributed in multiple dispersed data spaces.

[0030] The retrieval results dataset contains a first digital object that matches the retrieval text, and a second digital object that is pragmatically associated with the first digital object.

[0031] In some embodiments of this application, since relying solely on the search text makes it difficult to directly match digital objects stored in a dispersed data space to meet query requirements in a data network environment, the search system needs to expand the scope of search results to enhance data relevance and search coverage. Therefore, the search system determines the dataset of search results that match the user's input search text. Specifically, the system can first parse the user's search text and query the digital object registry to obtain the first matching digital object. Subsequently, based on the pragmatic relations of the first digital object, the system further associates and expands to second digital objects with pragmatic relations to form a complete dataset of search results. Pragmatic relations refer to the usage relationships of digital objects in different application scenarios, which determine the strength and applicability of the association between data objects. After performing this step, the search results are no longer limited to text matching, but can generate a more comprehensive dataset of search results based on the application relationships between data. This reduces the search bias caused by the current reliance on text matching in data network search systems and improves the contextual matching degree of search results.

[0032] In a specific example, a user enters the keyword "distributed storage architecture" into a search system, aiming to obtain data resources related to this technology. During the process, the search system retrieves the first digital object matching the keyword from various digital object registry entries, such as a technical document on "design specifications for distributed storage architecture." Next, the system analyzes the metadata of this first digital object, identifies its pragmatic relationships with other data objects, and expands the search results dataset by including related second digital objects, such as research literature on "distributed storage optimization algorithms." Thus, the final search results obtained by the user not only include technical documents directly matching the search text but also encompass a range of related materials expanded based on pragmatic relationships. This makes the search results more accurately meet the user's research needs, improving search efficiency and the comprehensiveness of information acquisition.

[0033] Step 102: Determine the query similarity of the metadata of each digital object in the retrieval results dataset to the retrieval text, and determine the pragmatic weight value of each digital object in the retrieval results dataset.

[0034] In some embodiments of this application, considering that text matching alone cannot fully reflect the actual relevance of data objects, it is necessary to determine the pragmatic weight value by combining data pragmatic relations while calculating the query similarity of each digital object. Therefore, the search system needs to evaluate the digital objects in the retrieval result dataset to ensure that the ranking results meet the user's query needs. During execution, the system first parses the retrieval text and extracts query keywords for subsequent similarity calculation. Next, for each digital object in the retrieval result dataset, the system uses the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm to calculate the query similarity of the object, measuring the degree of matching between its text content and the query text. Subsequently, the system constructs a link structure between digital objects based on data pragmatic relations and uses the Digital Object Rank (DO-Rank) algorithm to calculate the pragmatic weight value, measuring the importance of the digital object in the data space. The pragmatic weight value reflects the degree of influence of the digital object in a specific application scenario. In this way, the search system can rank digital objects based on comprehensive evaluation results, improving the accuracy of the retrieval results.

[0035] In a specific example, a user enters the keyword "distributed storage architecture" into a search system, hoping to obtain relevant technical information. During the process, the system first extracts the keyword "distributed storage" for query similarity calculation and uses the TF-IDF algorithm to calculate the similarity values ​​of multiple technical documents matching this keyword. Subsequently, the system further analyzes the metadata of each document in the search results dataset, identifies its pragmatic relationships with other digital objects, and calculates the pragmatic weight value of each document based on the pragmatic relationship structure using the DO-Rank algorithm. For example, a document describing "distributed storage optimization strategies" may receive a higher weight due to its strong pragmatic relationship, while a document only concerning "storage device hardware performance evaluation" may receive a lower weight. After completing this process, the user's final search results not only consider text matching but also the influence of pragmatic relationships, making the search results more relevant to the user's actual needs and improving search accuracy and information value.

[0036] Step 103: Determine the retrieval result evaluation value for each digital object in the retrieval result dataset based on the query similarity and pragmatic weight value of each digital object in the retrieval result dataset, so as to obtain the ranking result of the digital objects in the retrieval result dataset.

[0037] In some embodiments of this application, considering that relying solely on the matching degree of the query text cannot fully reflect the actual value of the data object, while combining pragmatic relation evaluation can improve the rationality of the ranking and make the ranking results more consistent with the real application scenario of the data, the search system needs to determine the retrieval result evaluation value based on the query similarity and pragmatic weight value of each digital object in the retrieval result dataset, in order to finally generate the ranking results. The system can first calculate the query similarity of each digital object in the retrieval result dataset, which is used to measure the degree of matching between the text content of the digital object and the query text. Subsequently, the system combines the weighted calculation of data pragmatic relations to determine the pragmatic weight value of each digital object, which reflects the relevance and application value of the digital object in the data network environment. After the calculation of query similarity and pragmatic weight value is completed, the system uses linear weighting or other optimization methods to combine the two to generate the retrieval result evaluation value, and ranks all digital objects according to the evaluation value. In this way, the search system can adjust the ranking of search results according to the pragmatic attributes of the data, so that the ranking results can reflect both the text matching degree of the data object and comprehensively consider its pragmatic weight, thereby improving the accuracy of retrieval and the effectiveness of data utilization.

[0038] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During the process, the system first calculates the query similarity of each document in the search results dataset. For instance, a document about "distributed storage optimization algorithms" might have a high similarity, while another document discussing "cloud storage hardware devices" might have a low similarity. Next, the system analyzes the pragmatic relationships between the documents, calculating pragmatic weight values ​​based on the citation frequency, application scenario relevance, and weight distribution of the data objects. For example, if a document is cited in multiple storage system studies, it might have a high pragmatic weight. Finally, the system combines query similarity and pragmatic weight values ​​to calculate a search result evaluation score and ranks the data objects according to their evaluation scores. The final ranking results obtained by the user are not only based on the query text matching degree but also comprehensively consider the actual application value of the data objects, making the search results more closely aligned with the user's needs and improving retrieval efficiency and information usability.

[0039] In summary, by extending the retrieval result dataset, composed of digital objects obtained from different data spaces, to include not only the first digital object directly matching the retrieval text but also the second digital object pragmatically related to the first digital object, the search results are no longer limited to text matching. Instead, based on the pragmatic relationships between digital objects stored in different data spaces, a more comprehensive search coverage is achieved without compromising the domain of data within each data space, thus improving the contextual matching degree of the data. Furthermore, by calculating the query similarity of each digital object to evaluate its direct relevance to the retrieval text, the retrieval results are ensured to prioritize returning pairs with a high degree of matching with the query text. This approach avoids situations where search results deviate from users' actual needs. Furthermore, it introduces a pragmatic weighting evaluation mechanism, calculating weights based on the pragmatic relevance between digital objects. This breaks through the limitations of existing search ranking algorithms that rely solely on text matching or webpage link analysis, enabling search results to reflect the usability of digital objects in the data space. This not only optimizes the search ranking logic but also makes the ranking results more aligned with user needs, improving search accuracy and information retrieval efficiency. Ultimately, the ranking of search results is calculated by comprehensively utilizing query similarity and pragmatic weight values, ensuring that the final ranking reflects the relevance of data objects and the importance of application scenarios, thus improving the precision and user satisfaction of the data network search system. Therefore, the method based on this application's embodiments ensures that data ranking matches actual application relationships, thereby improving the user's search experience in a distributed data network storage environment. By optimizing search accuracy, it not only improves the precision of the search system but also effectively reduces search errors caused by unreasonable ranking, enhancing the convenience for users to obtain key information.

[0040] Figure 2 This is another data-based network search and sorting method provided in the embodiments of this application.

[0041] The method may include the following steps:

[0042] Step 201: Determine the dataset of search results that matches the search text used to describe the target digital object to be searched from the registry of the data network, which is distributed in multiple dispersed data spaces.

[0043] The retrieval results dataset contains a first digital object that matches the retrieval text, and a second digital object that is pragmatically associated with the first digital object.

[0044] The method shown in this step has been explained in step 101 and will not be repeated here.

[0045] Optionally, step 201 includes the following sub-steps:

[0046] Sub-step 2011: Extract matching text for the target numeric object from the retrieved text.

[0047] In some embodiments of this application, the search system needs to extract text matching the target digital object from the search text to ensure that the subsequent retrieval process can correctly locate the data object. This step is performed because the user's search text typically contains multiple words or phrases, some of which may be directly related to the target digital object, while others may be irrelevant information or descriptive content. Therefore, text parsing and filtering are necessary to improve the accuracy of the search. During execution, the system first preprocesses the input search text, including word segmentation, stop word removal, punctuation filtering, and stemming, to ensure that the extracted text has a standardized format. Subsequently, the system analyzes the degree of matching between the search text and the metadata of the digital object in the existing data space based on an inverted index or semantic vector model, and filters out the matching text that best matches the characteristics of the target digital object. In this way, the search system can improve retrieval accuracy based on the refined matching text, reduce interference from invalid data, and enhance the relevance of the search results.

[0048] In a specific example, a user enters "efficient distributed data storage architecture" into a search system to retrieve relevant technical information. During the process, the search system first segments the search text, obtaining keywords such as "efficient," "distributed," "data storage," and "architecture." Then, the system filters out the modifier "efficient" and analyzes the matching degree between the remaining keywords and the metadata of the digital objects based on an inverted index or semantic vector model. Finally, the system extracts "distributed data storage architecture" as the matching text for subsequent retrieval. After following this process, the user's query can accurately locate technical documents related to distributed storage architecture, without being affected by irrelevant modifiers, making the search results more relevant to the user's needs and improving the efficiency of information retrieval.

[0049] Sub-step 2012 involves matching the matching text against the metadata of the digital objects stored in each data space, as recorded in the registry of each data space, to determine a first digital object from at least one data space based on the matched target metadata, and to determine a second digital object based on the pragmatic association between the digital objects stored in the data space corresponding to the determined first digital object, as recorded in the registry.

[0050] In some embodiments of this application, to expand the scope of search results and enhance data relevance and search coverage, the matching text is matched with the metadata of digital objects stored in each data space, recorded in the registry of each data space. Based on the matched target metadata, a first digital object is determined from at least one data space as a Basic Digital Object (BDO). Then, based on the pragmatic relationships between digital objects stored in the data space corresponding to the determined first digital object, a second digital object is determined as an Extended Digital Object (EDO) corresponding to the BDO. Specifically, the system can first parse the user's search text and query the digital object registry to obtain a matching first digital object. Subsequently, based on the pragmatic relationships of the first digital object, the system further associates and extends it to second digital objects with pragmatic relationships to form a complete search result dataset. After performing this step, the search results are no longer limited to text matching but can generate a more comprehensive search result dataset based on the application relationships between data. This reduces the search bias caused by traditional search systems relying solely on text matching and improves the contextual matching degree of search results.

[0051] In a specific example, a user enters the keyword "distributed storage architecture" into a search system, aiming to obtain data resources related to this technology. During the process, the search system retrieves the first digital object matching the keyword from the digital object registry, such as a technical document on "design specifications for distributed storage architecture." Next, the system analyzes the metadata of this first digital object, identifies its pragmatic relationships with other data objects, and expands the search results dataset by including related second digital objects, such as research literature on "distributed storage optimization algorithms." Thus, the final search results obtained by the user not only include technical documents directly matching the search text but also encompass a range of related materials expanded based on pragmatic relationships, making the search results more accurately meet the user's research needs and improving search efficiency and the comprehensiveness of information acquisition.

[0052] Optionally, sub-step 2012 includes the following sub-steps:

[0053] Sub-step 20121: Obtain the registry deployed in each data space, and match the matching text with the metadata of the digital objects in the data space recorded in each registry, so as to determine the first digital object from the data space based on the matched target metadata.

[0054] In some embodiments of this application, considering that the search text is usually the query content entered by the user, while the digital objects in the data space store structured information, the matching relationship between the two needs to be determined through calculation and analysis to ensure the accuracy of the search results. The search system needs to match the matching text with the metadata stored in the digital object registry of the data space to determine the first digital object most relevant to the search text. Specifically, the system first loads the metadata of the digital objects in the data space and performs keyword extraction and text normalization processing on the matching text. Subsequently, the system uses an inverted index or a text similarity calculation method based on the TF-IDF algorithm to match the matching text with the metadata of the digital objects to filter out the target metadata that best matches the query content. In this way, the search system can accurately identify the first digital object in the data space based on the matching text, ensuring that the search results are highly consistent with the user's needs, improving the accuracy of the search and the quality of the query matching.

[0055] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During the process, the system first loads all storage metadata from the digital object registry in the data space and performs word segmentation on the input search text, extracting core terms such as "distributed storage" and "architecture." Subsequently, the system queries the data space based on an inverted index structure and uses the TF-IDF algorithm to calculate the similarity value between the metadata of each digital object and the matching text. For example, a document about "distributed storage optimization strategies" might achieve a high match, while a document discussing "storage device maintenance techniques" might have a lower match. Finally, the system selects the digital object with the highest similarity as the first digital object and uses it for subsequent search ranking. After following this process, the user's final search results accurately reflect their query needs, enabling the search system to provide more precise and tailored data resources, improving search quality and the search experience.

[0056] Sub-step 20122: Determine the second digital object corresponding to the determined first digital object from each registry according to the preset pragmatic depth association relationship.

[0057] like Figure 3 As shown, in some embodiments of this application, considering that digital objects based solely on text matching cannot fully reflect the application relationships between data in the Internet of Things, extending pragmatic association objects can optimize search coverage and improve the scenario matching degree of retrieval. Therefore, the search system needs to identify and extend the first digital object, i.e., BDO (… Figure 3 B) in the text has a deep pragmatic relevance ( Figure 3 EDO (with arrow in the middle) Figure 3The system first analyzes the metadata of the first digital object to identify its pragmatic relationships with other digital objects. Then, based on a pragmatic weighting evaluation mechanism, the system filters out digital objects that meet the preset pragmatic depth association and identifies them as the second digital object. Pragmatic depth association refers to the pragmatic hierarchical relationship of digital objects in a specific application scenario; it determines the pragmatic connectivity and data transferability of search results, such as... Figure 3 The second digital object in the search results includes second digital objects whose pragmatic depth of association with the first digital object is either 1 (directly associated with the first digital object) or 2 (indirectly associated with the first digital object through another second digital object). This allows the search system to reasonably expand the search scope based on the actual application scenarios of the first digital object, enabling the ranking results to cover digital objects with greater pragmatic value and improving the application adaptability of the search results.

[0058] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical information. During the process, the system first matches the text against metadata in the digital object registry, identifying the first digital object, such as a technical document on "distributed storage architecture design principles." Subsequently, the system analyzes the pragmatic relationships of this first digital object, identifying its associations with other digital objects in the data space, and filters out digital objects that meet a preset pragmatic depth, such as a research paper on "distributed storage optimization algorithms." After this process is completed, the user's final search results not only include directly matched technical documents but also relevant research materials expanded based on pragmatic associations, making the search results more aligned with the user's research needs and improving the search system's accuracy and the comprehensive value of the information.

[0059] Sub-step 2013 generates a search result dataset based on the matched first and second numeric objects.

[0060] In some embodiments of this application, the search system needs to generate a retrieval result dataset based on the matched first and second digital objects to ensure that the final search results comprehensively reflect the text matching degree and pragmatic relationship between data objects. This step is performed because simply returning digital objects that directly match the search text may not fully reflect the data's usage scenario. Expanding the search by combining pragmatically related objects helps improve the completeness and usability of the query results. During execution, the system first loads the first digital object and its metadata into the retrieval result dataset and further analyzes the pragmatic relationship of the object, filtering out second digital objects that meet the preset pragmatic depth of relationship. Subsequently, the system merges the first and second digital objects and stores the retrieval result dataset according to an optimized data structure for subsequent sorting and display. The retrieval result dataset refers to the information collection containing all relevant digital objects ultimately formed during the search process, and its structure supports subsequent search optimization and query expansion. After performing this step, the search system can ensure that the search results are not only based on query text matching but also reflect the application relationship between data objects, improving the accuracy of the search and the comprehensiveness of information acquisition.

[0061] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to obtain relevant technical information. During the process, the system first searches the digital object registry to retrieve the first digital object directly matching the search text, such as a technical document on "distributed storage architecture design principles." Then, the system analyzes the pragmatic relationships of this first digital object, identifies its associated second digital object, such as a research paper on "distributed storage optimization algorithms," and adds it to the search results dataset. Finally, the system generates the dataset, ensuring that the user's search results include both technical documents directly matching the query text and related extended information, thus enhancing the richness and practical application value of the search results.

[0062] Step 202: Determine the query similarity of the metadata of each digital object in the retrieval results dataset to the retrieval text, and determine the pragmatic weight value of each digital object in the retrieval results dataset.

[0063] The method shown in this step has been explained in step 102 and will not be repeated here.

[0064] Optionally, to determine the query similarity of the metadata of each numeric object in the retrieval results dataset to the retrieval text, step 202 includes the following sub-steps:

[0065] Sub-step 2021: Extract the search segmentation set of the search text from the search text, and extract the metadata segmentation set of each digital object from the metadata of each search result dataset.

[0066] In some embodiments of this application, since the search text typically contains multiple words, some of which may be the core concepts of the query, while others may be descriptive expressions or stop words, the text needs to be parsed and filtered to extract keywords that can effectively represent the query content. Therefore, the search system needs to extract a search word segmentation set from the search text to ensure the accuracy of subsequent query similarity calculations. During execution, the system first performs word segmentation processing on the input search text, including word segmentation recognition using natural language processing technology and removal of stop words, to ensure that the extracted word segmentation set accurately reflects the query intent. Subsequently, the system performs the same word segmentation processing on the metadata of digital objects in each search result dataset to obtain the metadata word segmentation set for each digital object, ensuring that the vocabulary expression format of all data objects is consistent and suitable for subsequent similarity calculations. After performing this step, the search system can unify the text processing method, extract core keywords, improve the accuracy of search matching, and reduce the interference of irrelevant words on the search results.

[0067] In a specific example, a user enters the query text "distributed storage architecture optimization technology" into the search system, aiming to retrieve relevant technical documents. During execution, the system first performs word segmentation analysis on the query text, extracting "distributed storage," "architecture," "optimization," and "technology" as the retrieval word set. Subsequently, the system performs the same word segmentation process on the metadata of digital objects in the data space, extracting the metadata word set for each document. For example, the metadata word set for a technical document might contain keywords such as "distributed storage," "data management," and "performance optimization." After completing this execution process, the search system successfully constructs the word sets for both the query text and metadata, enabling subsequent similarity calculations to be based on a unified lexical structure, improving search accuracy, and ensuring that the search results are highly consistent with the user's query topic.

[0068] Sub-step 2022: Determine the metadata term frequency of each search term in the search term set that appears in each metadata term set, and determine the inverse document frequency of each search term in the search term set that appears in all metadata term sets.

[0069] In some embodiments of this application, considering that the metadata of different digital objects may contain the same search terms but with different frequencies of occurrence, it is necessary to measure the importance of the term in a single digital object and combine it with the distribution in the entire search result dataset to optimize text matching. The search system needs to calculate the term frequency of each search term in the metadata of the digital object, as well as the inverse document frequency of the search term in all metadata term sets, to ensure the accuracy of query similarity calculation. During execution, the system first scans the metadata term set of digital objects in each search result dataset, counts the number of times each term in the search term set appears in these metadata, and calculates the metadata term frequency (TF). Subsequently, the system searches for the occurrence of each search term in all metadata term sets and calculates the inverse document frequency (IDF). This value reflects the scarcity of the search term in the entire data space; a lower frequency indicates that the term has high discriminative power, while a higher frequency indicates that the term is more common. After performing this step, the search system can adjust the query similarity calculation based on the term frequency and inverse document frequency of the retrieved words, so that words with higher importance have greater weight in the final score, thereby improving the accuracy of search results.

[0070] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During execution, the system first traverses the metadata of all digital objects in the search results dataset and counts the frequency of "distributed storage" and "architecture" in each document within the search term set. For example, a document about "distributed storage optimization strategies" might contain "distributed storage" multiple times in its metadata, thus having a high metadata term frequency for that word. Subsequently, the system counts the occurrence of this search term across all metadata term sets to calculate inverse document frequency (IVF). For instance, if "architecture" appears in multiple documents, its IVF is low, while if "distributed storage" appears only in a few documents, its IVF is high. After completing this execution process, the search system can appropriately adjust the weights of term frequency and IVF, ensuring that the final search similarity calculation better reflects the discriminative power of the search terms and improves the relevance and matching accuracy of the search results.

[0071] Sub-step 2023: The weighted sum of the metadata term frequency and inverse document frequency of each search word in the search word segment set for each digital object in the search result dataset is determined as the query similarity of the metadata of each digital object in the search result dataset to the search text.

[0072] In some embodiments of this application, since using term frequency or inverse document frequency alone may lead to search results biased towards high-frequency or rare words, and weighted summation can balance the influence of both, making query similarity accurately reflect the degree of text matching, the search system needs to calculate the query similarity of each digital object to the searched text based on the weighted sum of the metadata term frequency and inverse document frequency of each search term in the search term set, in order to optimize the search matching degree. During execution, the system first extracts the term set of each digital object from the metadata of each digital object in the search results dataset, and calculates the metadata term frequency and inverse document frequency for the search terms within it. Subsequently, the system performs a weighted summation of the metadata term frequency and inverse document frequency according to a preset weight factor to calculate the query similarity of each digital object. Query similarity is a numerical indicator that measures the degree of content matching between a digital object and the searched text, affecting the final ranking results. After performing this step, the search system can optimize the ranking logic based on the weighted query similarity calculation, making the search results more in line with user query needs, improving the accuracy of the search and the effectiveness of the information.

[0073] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During execution, the system first iterates through the metadata of all digital objects in the search results dataset and calculates the metadata frequency of "distributed storage" and "architecture" in the search terminology set. For example, a technical document about "distributed storage optimization strategies" might contain "distributed storage" multiple times, thus its metadata frequency is high. Next, the system calculates the inverse document frequency (IVF) of the search term across all metadata terminology sets. For instance, if "architecture" appears in multiple documents, its IVF is low, while "distributed storage" appears in only a few documents, resulting in a higher IVF. Finally, the system performs a weighted sum of term frequencies and IVF based on preset weighting factors to calculate the query similarity for each digital object, which is then used for subsequent ranking. After this execution process, the user's final search results are not only ranked according to the search text matching degree but also incorporate weighted calculations, making the search results more accurate and improving the accuracy and search experience.

[0074] Specifically, the digital object do k For the search text Query similarity It can be done through formula Solve for, where This indicates word segmentation in the query text Q. word frequency, express Inverse document frequency.

[0075] Optionally, to determine the pragmatic weight value for each numeric object in the retrieval results dataset, step 202 includes the following sub-steps:

[0076] Sub-step 2024: Establish a Markov chain model to characterize the pragmatic relationships between digital objects in the retrieval results dataset.

[0077] The established Markov chain model defines the pragmatic weight value between digital objects with direct relationships as the association probability value between digital objects with direct relationships in the Markov chain model, and defines the weight numerical relationship between the pragmatic weight values ​​between digital objects without direct relationships as the probability numerical relationship between the association probability values ​​between digital objects without direct relationships in the Markov chain model.

[0078] In some embodiments of this application, considering that the pragmatic relationships between digital objects in the extended search results dataset not only affect the ranking of search results but also determine the application value and contextual matching degree of the data, it is necessary to quantify the degree of association between different objects through a probabilistic model. The search system needs to establish a Markov Chain Model (MCM) to characterize the pragmatic relationships between digital objects in the search results dataset. Specifically, the system first constructs the state space of the Markov Chain Model and maps all digital objects in the search results dataset to state nodes in the model. Subsequently, the system defines state transition probabilities based on the pragmatic weight values ​​of the digital objects. For digital objects with direct relationships, their pragmatic weight values ​​are directly defined as the association probability values ​​in the Markov chain. For digital objects without direct relationships, the relationship between their pragmatic weight values ​​is constrained by the numerical relationship between indirect association probability values. This ensures that the association structure of all data objects can be reasonably represented in the model. Such a search system can accurately characterize the pragmatic relationships of digital objects using Markov chain models, so that the ranking logic not only considers the degree of text matching, but also reflects the actual application value of data objects in the Internet of Things environment, thereby improving the accuracy and rationality of search results.

[0079] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During execution, the system first identifies digital objects in the search results dataset and constructs a Markov chain model, where each digital object is defined as a state node. Subsequently, the system calculates the pragmatic weights between objects and converts the weight values ​​between directly related digital objects into state transition probabilities. For example, a technical document about "distributed storage optimization strategies" might have a high pragmatic weight and form a high-probability association with a document about "distributed storage architecture performance evaluation." Simultaneously, based on the weight values ​​between indirectly related data objects, the system derives the probability values ​​of indirect associations and establishes appropriate numerical relationships in the model. After completing this execution process, the search system can more accurately represent the pragmatic connections between data objects, making the ranking results more aligned with the query requirements, improving the quality of information retrieval and the usability of the search system.

[0080] Sub-step 2025: Solve the Markov chain model to obtain the pragmatic weight value for each digital object.

[0081] In some embodiments of this application, since the pragmatic relationships between digital objects are not only reflected in static association structures but also involve dynamic probability transitions, it is necessary to quantify the pragmatic influence of data objects by solving Markov chains. Therefore, the search system needs to solve the Markov chain model to calculate the pragmatic weight value of each digital object. During execution, the system first initializes the transition matrix of the Markov chain, which defines the pragmatic association relationships of all digital objects in the retrieval result dataset. Subsequently, the system uses iterative calculation methods, such as power iteration or the Gauss-Seidel method, to solve for the weight distribution of each digital object in a stable state, and finally obtains the pragmatic weight value of each digital object. The Markov chain model is used to describe the pragmatic influence of data objects, reflecting the relevance and application value of data through probability transition relationships. In this way, the search system can optimize the ranking logic based on the solution results of the Markov chain model, giving higher weights to data objects with stronger pragmatic influence, thereby improving the rationality and pragmatic matching degree of the search results.

[0082] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During execution, the system first establishes a Markov chain model, where state nodes correspond to digital objects in the search results dataset, and initializes the transition probability matrix. For instance, a technical document on "distributed storage optimization strategies" might have a higher transition probability due to being cited in multiple studies. Subsequently, the system uses a power-law iterative calculation to determine the pragmatic weights of each data object in a stable state, and finally uses the result for ranking optimization. After this execution process, the user's final search results are not only ranked based on query text matching but also incorporate pragmatic relation calculations, making the search results more aligned with the user's research needs and improving retrieval accuracy and the application value of the information.

[0083] Optionally, sub-step 2025 includes the following sub-steps:

[0084] Sub-step 20251 initializes the pragmatic weight value of each digital object in the retrieval result dataset to a preset initial pragmatic weight value.

[0085] In some embodiments of this application, considering that pragmatic weight values ​​are key parameters affecting search ranking, improper initialization may lead to deviations or instability in the calculation process, thereby affecting the rationality of the final retrieval results. Therefore, the search system needs to initialize the pragmatic weight value of each digital object in the retrieval result dataset to ensure that subsequent iterative calculations can be performed based on stable initial values. During execution, the system first defines a preset initial pragmatic weight value, such as 1, and assigns this value to all digital objects in the retrieval result dataset, giving them a uniform pragmatic weight at the start of iterative calculations. The preset initial pragmatic weight value is usually uniformly distributed or set based on prior knowledge to provide stable input for the Markov chain model. In this way, the search system can ensure that the pragmatic weight calculation of all digital objects starts from a unified initial condition, improving the convergence of the calculation and providing a reasonable basis for subsequent pragmatic weight optimization.

[0086] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During execution, the system first identifies all digital objects in the search results dataset and initializes their pragmatic weights, for example, setting an initial weight of 1 for all documents to ensure computational stability. Subsequently, the system updates the pragmatic weights of each digital object based on the state transition matrix in a Markov chain model to optimize search ranking. After this process is completed, the final search results obtained by the user are not only ranked based on the query text matching degree but also incorporate the calculated initialized pragmatic weights, making the search results more aligned with the user's query needs and improving search accuracy and the rationality of information retrieval.

[0087] Sub-step 20252 updates the pragmatic weight value of each fourth digital object by taking the weighted sum of the ratio of the first pragmatic weight value to the second pragmatic weight value of each third digital object in the retrieval result dataset, the product of the product of the first pragmatic weight value and the pragmatic weight value of the third digital object, the product of the damping factor coefficient of the retrieval result dataset, and the sum of the pragmatic weight evaluation correction value.

[0088] Here, the third digital object is the digital object on the next higher node of the fourth digital object in each Markov chain of the Markov chain model; the first pragmatic weight value is the pragmatic weight value from the third digital object to the fourth digital object on the Markov chain where the third digital object is located; the second pragmatic weight value is the sum of the pragmatic weight values ​​of all digital objects except the third digital object to the fourth digital object; and the pragmatic weight evaluation correction value is the probability value of the fourth digital object forming a pragmatic association with any digital object in the retrieval result dataset.

[0089] In some embodiments of this application, since the pragmatic relationships of data objects in a data network environment are dynamically changing, relying solely on static weight calculations is insufficient to reflect the actual influence between objects. Therefore, iterative updates of weight values ​​are necessary to ensure that the final ranking better reflects the application scenario of the data. In this case, the search system needs to update the pragmatic weight value of each fourth digital object to ensure that the Markov chain model accurately represents the pragmatic relationships between digital objects. Specifically, considering the following mathematical model: for a digital object numbered k (do... k ), then do k pragmatic weight value W prag (do) k The calculation method for () is shown in the formula:

[0090] ,

[0091] Where DF represents the damping factor, used to control the probability of jumping between DOs based on pragmatic relations; N represents the total number of DOs in the current directed graph (i.e., the data size of the retrieval result dataset); M (do k ) indicates the direct pointer to do k A collection of digital objects; prag mk Represents the numeric object do m and do k Pragmatic weight between; P(do) m ) represents do m The sum of pragmatic weights pointing to other digital objects. This search system optimizes pragmatic weight calculation based on the Markov chain model, ensuring that the ranking logic not only reflects text matching but also comprehensively considers the application relationships between digital objects, thus improving the rationality and matching accuracy of search results.

[0092] like Figure 4 As shown, in the initial state, the initial pragmatic weight values ​​of the numeric objects do1, do2, do3, and do4 are all 1, and the pragmatic weights of do1 and do2 are prag. 12 =0.75, pragmatic weight of do2 and do4 24 =0.5, pragmatic weights of do2 and do3 32 =0.25, damping factor DF=0.85, then if do2 is considered as the fourth digital object, then the pragmatic weight value W of do2 after one iteration is... prag (do2) Calculate according to the following formula:

[0093] .

[0094] In this way, the search results that users ultimately obtain are not only ranked based on text matching, but also take into account pragmatic weight optimization, making the search results more in line with the query needs and improving the accuracy of retrieval and the effectiveness of information acquisition.

[0095] Step 203: The weighted sum of the query similarity and pragmatic weight values ​​of each digital object in the retrieval result dataset is determined as the retrieval result evaluation value of each digital object in the retrieval result dataset.

[0096] In some embodiments of this application, since using query similarity or pragmatic weight values ​​alone may lead to ranking bias, a weighted sum method can combine the effects of both to improve the rationality of the retrieval results. Therefore, the search system needs to calculate the retrieval result evaluation value for each digital object based on the weighted sum of query similarity and pragmatic weight values ​​to ensure that the final ranking accurately reflects the relevance and application value of the data. Specifically, the system can first extract the query similarity of each digital object from the retrieval result dataset. This value is used to measure the degree of matching between the text content of the digital object and the query text. Subsequently, the system can calculate the pragmatic weight value of the digital object based on the DO-rank algorithm to reflect its pragmatic relevance strength in the data network environment. After the calculation is completed, the system uses a predefined weighting factor to perform a weighted sum of the query similarity and pragmatic weight values, and uses the resulting weighted sum as the retrieval result evaluation value of the digital object. After performing this step, the search system can rank the digital objects in the retrieval result dataset according to the comprehensive evaluation value, so that the final ranking can reflect both the text matching degree and comprehensively consider the pragmatic relationship of the data objects, thereby improving retrieval accuracy and information utilization value.

[0097] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to retrieve relevant technical documents. During the process, the system first calculates the query similarity of each document in the search results dataset. For example, a document about "distributed storage optimization strategies" might have a high query similarity, while another document only covering "general storage hardware analysis" might have a low similarity. Next, the system calculates the pragmatic weight values ​​of these documents. For instance, a document cited in multiple storage system studies might receive a higher pragmatic weight. Finally, the system uses preset weighting factors to weight and sum the query similarity and pragmatic weight values ​​(e.g., a query similarity weight of 0.7 and a pragmatic weight of 0.3) to determine the evaluation value of each document's search results and rank the results accordingly. After this process, the user's final search results are not only ranked based on query text matching but also incorporate pragmatic weight evaluation, making the search results more relevant to the user's actual needs and improving the accuracy and information value of the search.

[0098] Specifically, the digital object do k Evaluation value of search results for the search text Q Specifically, it can be done through the formula. Records, including Represents the numeric object do k Metadata (do k The query similarity between Q and the retrieved text Q. Indicates do k The pragmatic weight values ​​are α and β, which are preset weights.

[0099] Step 204: Arrange the digital objects according to the size of the retrieval result evaluation value of each digital object in the retrieval result dataset to obtain the sorting result of the digital objects in the retrieval result dataset.

[0100] In some embodiments of this application, since simply calculating the evaluation value of search results cannot directly improve the user's search experience, only through reasonable sorting can high-value data objects be prioritized in the search results, thereby improving the usability of the search. Therefore, the search system needs to sort the digital objects based on the evaluation value of each digital object in the search result dataset to ultimately determine the order of the search results. Specifically, the system can first collect the evaluation values ​​of all digital objects and arrange them according to a preset sorting rule, such as descending order, so that digital objects with higher evaluation values ​​are displayed first. The sorting rule can be dynamically adjusted according to the query scenario to optimize the arrangement of data objects. In this way, the search system can provide sorting results based on a comprehensive evaluation of query matching degree and pragmatic weight, thereby ensuring that users can access the most relevant digital objects first and improving the efficiency of information retrieval.

[0101] In a specific example, a user enters the keyword "distributed storage architecture" into a search system to obtain relevant technical information. During the process, the system first calculates the evaluation score of each document's search results. For instance, a document about "distributed storage optimization strategies" might receive a higher evaluation score, while another document discussing "storage device maintenance techniques" might receive a lower score. The system then sorts these documents in descending order, ensuring that the document with the highest evaluation score appears first in the search results. After this process is completed, the final search results seen by the user are sorted according to a comprehensive evaluation of the data object, prioritizing high-quality data and improving the accuracy of the search and the convenience of data retrieval.

[0102] like Figure 5 The diagram shows a complete sorting process according to an embodiment of this application:

[0103] Step S1: The system receives the search text Q input by the user;

[0104] Step S2: Query the digital object registry in multiple data spaces to determine the first matching digital object;

[0105] Step S3: Based on the pragmatic relationships between digital objects, expand the search scope in multiple data spaces, identify second digital objects that have a pragmatic relationship with the first digital object, and then query them:

[0106] Step S3.1: Based on the pragmatic relationship between digital objects, expand the search scope to obtain the second digital object, and then obtain the search result dataset;

[0107] Step S3.2: Calculate the query similarity of each numeric object in the retrieval results dataset;

[0108] Step S3.3: Calculate the pragmatic weight value of each numeric object in the retrieval results dataset;

[0109] Step S3.4: Based on query similarity and pragmatic weight values, the system performs a final ranking of all digital objects and generates an optimized search results list. This ranking logic prioritizes data objects with high matching scores and high pragmatic value, enabling users to obtain content that meets their needs more quickly and accurately, thus improving the efficiency of information retrieval and the performance of the search system.

[0110] In summary, by extending the retrieval result dataset, composed of digital objects obtained from different data spaces, to include not only the first digital object directly matching the retrieval text but also the second digital object pragmatically related to the first digital object, the search results are no longer limited to text matching. Instead, based on the pragmatic relationships between digital objects stored in different data spaces, a more comprehensive search coverage is achieved without compromising the domain of data within each data space, thus improving the contextual matching degree of the data. Furthermore, by calculating the query similarity of each digital object to evaluate its direct relevance to the retrieval text, the retrieval results are ensured to prioritize returning pairs with a high degree of matching with the query text. This approach avoids situations where search results deviate from users' actual needs. Furthermore, it introduces a pragmatic weighting evaluation mechanism, calculating weights based on the pragmatic relevance between digital objects. This breaks through the limitations of existing search ranking algorithms that rely solely on text matching or webpage link analysis, enabling search results to reflect the usability of digital objects in the data space. This not only optimizes the search ranking logic but also makes the ranking results more aligned with user needs, improving search accuracy and information retrieval efficiency. Ultimately, the ranking of search results is calculated by comprehensively utilizing query similarity and pragmatic weight values, ensuring that the final ranking reflects the relevance of data objects and the importance of application scenarios, thus improving the precision and user satisfaction of the data network search system. Therefore, the method based on this application's embodiments ensures that data ranking matches actual application relationships, thereby improving the user's search experience in a distributed data network storage environment. By optimizing search accuracy, it not only improves the precision of the search system but also effectively reduces search errors caused by unreasonable ranking, enhancing the convenience for users to obtain key information.

[0111] refer to Figure 6 It illustrates a data-driven network search and ranking system 30 based on data pragmatics provided in an embodiment of this application, comprising:

[0112] Data set module 301 is used to determine a search result dataset that matches the search text used to describe the target digital object to be searched from the registry of a data network distributed in multiple dispersed data spaces; the search result dataset contains a first digital object that matches the search text and a second digital object that is pragmatically associated with the first digital object;

[0113] The parameter calculation module 302 is used to determine the query similarity of the metadata of each digital object in the retrieval result dataset to the retrieval text, and to determine the pragmatic weight value of each digital object in the retrieval result dataset.

[0114] The evaluation and ranking module 303 is used to determine the evaluation value of each digital object in the search results dataset based on the query similarity and pragmatic weight value of each digital object in the search results dataset, so as to obtain the ranking result of the digital objects in the search results dataset.

[0115] Optionally, dataset module 301 includes:

[0116] The retrieval and extraction submodule is used to extract matching text for the target numeric object from the retrieved text;

[0117] The digital object submodule is used to match the matching text with the metadata of the digital objects stored in each data space, which are recorded in the registry of each data space, so as to determine the first digital object from at least one data space based on the matched target metadata, and to determine the second digital object based on the pragmatic association relationship between the digital objects stored in the data space recorded in the registry corresponding to the determined first digital object.

[0118] The dataset submodule is used to generate a dataset of search results based on the matched first and second numeric objects.

[0119] Optionally, the digital object submodule includes:

[0120] The first digital object unit is used to obtain the registry deployed in each data space and match the matching text with the metadata of the digital objects in the data space recorded in each registry, so as to determine the first digital object from the data space based on the matched target metadata.

[0121] The second digital object unit is used to determine, according to a preset pragmatic depth association relationship, the second digital object corresponding to the determined first digital object from each registry.

[0122] Optionally, the parameter calculation module 302 includes:

[0123] The retrieval and word segmentation submodule is used to extract the retrieval word segmentation set from the retrieval text and extract the respective metadata word segmentation set from the metadata of the digital objects in each retrieval result dataset;

[0124] The metadata segmentation submodule is used to determine the metadata term frequency of each search term in the search segmentation set that appears in each metadata segmentation set, and to determine the inverse document frequency of each search term in the search segmentation set that appears in all metadata segmentation sets.

[0125] The similarity submodule is used to determine the query similarity of the metadata of each digital object in the search results dataset to the search text by weighting the sum of the metadata term frequency and inverse document frequency of each search term in the search term set.

[0126] Optionally, the parameter calculation module 302 includes:

[0127] The modeling submodule is used to build a Markov chain model to characterize the pragmatic relationships between digital objects in the retrieval results dataset. The built Markov chain model is used to define the pragmatic weight values ​​between digital objects with direct relationships as the association probability values ​​between digital objects with direct relationships in the Markov chain model, and to define the weight numerical relationship between the pragmatic weight values ​​between digital objects without direct relationships as the probability numerical relationship between the association probability values ​​between digital objects without direct relationships in the Markov chain model.

[0128] The modeling and solving submodule is used to solve the Markov chain model to obtain the pragmatic weight value for each digital object.

[0129] Optionally, the modeling and solving submodule includes:

[0130] An initialization unit is used to initialize the pragmatic weight value of each numeric object in the retrieval result dataset to a preset initial pragmatic weight value;

[0131] The iterative solution unit updates the pragmatic weight value of each fourth digital object by summing the weighted sum of the ratio of the first pragmatic weight value to the second pragmatic weight value of each third digital object in the retrieval result dataset, the product of the first pragmatic weight value to the second pragmatic weight value of each third digital object, the product of the first pragmatic weight value to the second pragmatic weight value of each third digital object, the product of the damping factor coefficient of the retrieval result dataset, and the pragmatic weight evaluation correction value. The third digital object is the digital object on each Markov chain in the Markov chain model that is the next-level node above the fourth digital object. The first pragmatic weight value is the pragmatic weight value from the third digital object to the fourth digital object on the Markov chain containing the third digital object. The second pragmatic weight value is the sum of the pragmatic weight values ​​of all digital objects from the third digital object to the fourth digital object. The pragmatic weight evaluation correction value is the probability value that the fourth digital object forms a pragmatic association with any digital object in the retrieval result dataset.

[0132] Optionally, the evaluation and ranking module 303 includes:

[0133] The evaluation value submodule is used to determine the evaluation value of each digital object in the retrieval result dataset by weighting the query similarity and pragmatic weight values ​​of each digital object.

[0134] The sorting submodule is used to rank the digital objects in the search results dataset according to the size of the search result evaluation value of each digital object, so as to obtain the sorting result of the digital objects in the search results dataset.

[0135] In summary, by extending the retrieval result dataset, composed of digital objects obtained from different data spaces, to include not only the first digital object directly matching the retrieval text but also the second digital object pragmatically related to the first digital object, the search results are no longer limited to text matching. Instead, based on the pragmatic relationships between digital objects stored in different data spaces, a more comprehensive search coverage is achieved without compromising the domain of data within each data space, thus improving the contextual matching degree of the data. Furthermore, by calculating the query similarity of each digital object to evaluate its direct relevance to the retrieval text, the retrieval results are ensured to prioritize returning pairs with a high degree of matching with the query text. This approach avoids situations where search results deviate from users' actual needs. Furthermore, it introduces a pragmatic weighting evaluation mechanism, calculating weights based on the pragmatic relevance between digital objects. This breaks through the limitations of existing search ranking algorithms that rely solely on text matching or webpage link analysis, enabling search results to reflect the usability of digital objects in the data space. This not only optimizes the search ranking logic but also makes the ranking results more aligned with user needs, improving search accuracy and information retrieval efficiency. Ultimately, the ranking of search results is calculated by comprehensively utilizing query similarity and pragmatic weight values, ensuring that the final ranking reflects the relevance of data objects and the importance of application scenarios, thus improving the precision and user satisfaction of the data network search system. Therefore, the method based on this application's embodiments ensures that data ranking matches actual application relationships, thereby improving the user's search experience in a distributed data network storage environment. By optimizing search accuracy, it not only improves the precision of the search system but also effectively reduces search errors caused by unreasonable ranking, enhancing the convenience for users to obtain key information.

[0136] Reference Figure 7 This is a block diagram of an electronic device 500 according to another embodiment of the present invention. For example, the electronic device 500 may be provided as a server. The electronic device 500 may include one or more of the following components: a processing component 502, a memory 504, a power supply component 506, a multimedia component 508, an audio component 510, an input / output (I / O) interface 512, a sensor component 514, and a communication component 516.

[0137] The electronic device 500 includes a processing component 502, which further includes one or more processors, and memory resources represented by memory 504 for storing instructions, such as application programs, that can be executed by the processing component 502. The application programs stored in memory 504 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 502 is configured to execute instructions to perform the methods provided in the embodiments of this application.

[0138] Processing component 502 typically controls the overall operation of electronic device 500, such as operations associated with display, telephone calls, data communication, camera operation, and recording operations. Processing component 502 may include one or more processors 520 to execute instructions to perform all or part of the steps of the methods described above. Furthermore, processing component 502 may include one or more modules to facilitate interaction between processing component 502 and other components.

[0139] Memory 504 is used to store various types of data to support the operation of electronic device 500. Memory 504 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0140] Power supply component 506 provides power to various components of electronic device 500. Power supply component 506 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 500.

[0141] Multimedia component 508 includes an interface that provides an output interface between electronic device 500 and user. In some embodiments, the interface may include a liquid crystal display (LCD) and a touch panel (TP). In some embodiments, multimedia component 508 includes a front-facing camera and / or a rear-facing camera.

[0142] Audio component 510 is used to output and / or input audio signals. The received audio signals can be further stored in memory 504 or transmitted via communication component 516.

[0143] Input / output I / O interface 512 provides an interface between processing component 502 and peripheral interface modules.

[0144] Sensor assembly 514 includes one or more sensors for providing state assessment of various aspects of electronic device 500. For example, sensor assembly 514 can detect the on / off state of electronic device 500, or the relative positioning of the components. Sensor assembly 514 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 514 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications.

[0145] The communication component 516 facilitates wired or wireless communication between the electronic device 500 and other devices. The electronic device 500 can access wireless networks based on communication standards, such as WiFi, carrier networks (such as 2G, 3G, 4G, or 5G), or combinations thereof.

[0146] In an exemplary embodiment, the electronic device 500 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to implement the methods provided in the embodiments of this application.

[0147] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 504 including instructions, which can be executed by a processor 520 of an electronic device 500 to perform the above-described method.

[0148] Electronic device 500 can also operate on an operating system stored in memory 504, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0149] It should be noted that, for the sake of simplicity, the method embodiments of this application are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily essential to the embodiments of this application.

[0150] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the application disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the claims below.

Claims

1. A data network search and ranking method based on data pragmatics, characterized in that, include: From the registry of a data network distributed across multiple dispersed data spaces, a dataset of search results matching the search text used to describe the target digital object to be searched is determined; the dataset of search results includes a first digital object matching the search text and a second digital object pragmatically associated with the first digital object; Determine the query similarity of the metadata of each digital object in the retrieval result dataset to the retrieval text, and determine the pragmatic weight value of each digital object in the retrieval result dataset; The retrieval result evaluation value of each digital object in the retrieval result dataset is determined based on the query similarity and pragmatic weight value of each digital object in the retrieval result dataset, so as to obtain the ranking result of the digital objects in the retrieval result dataset; Determining the pragmatic weight value of each digital object in the retrieval result dataset includes: A Markov chain model is established to characterize the pragmatic association relationships between digital objects in the retrieval result dataset. The established Markov chain model defines the pragmatic weight values ​​between digital objects with direct association relationships as the association probability values ​​between the digital objects with direct association relationships in the Markov chain model, and defines the weight numerical relationship between the pragmatic weight values ​​between digital objects without direct association relationships as the probability numerical relationship between the association probability values ​​between the digital objects without direct association relationships in the Markov chain model. Solve the Markov chain model to obtain the pragmatic weight value for each of the digital objects.

2. The data network search and ranking method based on data usage as described in claim 1, characterized in that, The step of determining the search result dataset matching the search text used to describe the target digital object to be searched from the registry of a data network distributed in multiple dispersed data spaces includes: Extract matching text for the target digital object from the retrieved text; The matching text is matched with the metadata of the digital objects stored in each of the data spaces, which are recorded in the registry of each data space, so as to determine the first digital object from at least one of the data spaces based on the matched target metadata, and to determine the second digital object based on the pragmatic association between the digital objects stored in the data spaces recorded in the registry corresponding to the determined first digital object. The search result dataset is generated based on the matched first and second digital objects.

3. The data network search and ranking method based on data pragmatics as described in claim 2, characterized in that, The step of matching the matched text with the metadata of the digital objects stored in each of the data spaces, as recorded in the registry of each data space, to determine the first digital object from at least one of the data spaces based on the matched target metadata, and to determine the second digital object based on the pragmatic association between the digital objects stored in the data spaces corresponding to the determined first digital object, as recorded in the registry, includes: Obtain the registry deployed in each of the data spaces, and match the matching text with the metadata of the digital objects in the data spaces recorded in each of the registry, so as to determine the first digital object from the data space based on the matched target metadata; According to the preset pragmatic depth association relationship, the second digital object corresponding to the determined first digital object is determined from each of the registry entries.

4. The data network search and ranking method based on data pragmatics as described in claim 1, characterized in that, Determining the query similarity of the metadata of each digital object in the retrieval result dataset to the retrieval text includes: Extract the search word set from the search text, and extract the metadata word set from the metadata of each digital object in the search result dataset; Determine the metadata term frequency of each search term in the search term set appearing in each of the metadata term sets, and determine the inverse document frequency of each search term in the search term set appearing in all of the metadata term sets. The weighted sum of the metadata term frequency and inverse document frequency of each digital object in the search result dataset for each search term in the search term set is determined as the query similarity of the metadata of each digital object in the search result dataset to the search text.

5. The data network search and ranking method based on data usage as described in claim 1, characterized in that, Solving the Markov chain model to obtain the pragmatic weight value for each digital object includes: The pragmatic weight value of each digital object in the retrieval result dataset is initialized to a preset initial pragmatic weight value; The pragmatic weight value of each fourth digital object is updated by summing the weighted sum of the ratio of the first pragmatic weight value to the second pragmatic weight value of each third digital object in the retrieval result dataset, the product of the first pragmatic weight value to the second pragmatic weight value of the third digital object, the product of the damping factor coefficient of the retrieval result dataset, and the pragmatic weight evaluation correction value. The third digital object is the digital object on the previous node of the fourth digital object in each Markov chain of the Markov chain model. The first pragmatic weight value is the pragmatic weight value from the third digital object to the fourth digital object on the Markov chain containing the third digital object. The second pragmatic weight value is the sum of the pragmatic weight values ​​of all digital objects other than the fourth digital object from the third digital object. The pragmatic weight evaluation correction value is the probability value that the fourth digital object forms a pragmatic association with any digital object in the retrieval result dataset.

6. The data network search and ranking method based on data pragmatics as described in claim 1, characterized in that, The step of determining the retrieval result evaluation value for each digital object in the retrieval result dataset based on its query similarity and pragmatic weight value, in order to obtain a ranking result for the digital objects in the retrieval result dataset, includes: The weighted sum of the query similarity and pragmatic weight values ​​of each digital object in the retrieval result dataset is determined as the retrieval result evaluation value of each digital object in the retrieval result dataset; The digital objects in the retrieval result dataset are arranged according to the magnitude of the retrieval result evaluation value of each digital object, so as to obtain the sorting result of the digital objects in the retrieval result dataset.

7. A data network search and ranking system based on data pragmatics, characterized in that, include: The dataset module is used to determine a dataset of search results that matches the search text used to describe the target digital object to be searched, from a registry of a data network distributed in multiple dispersed data spaces; the dataset of search results includes a first digital object that matches the search text, and a second digital object that is pragmatically associated with the first digital object; The parameter calculation module is used to determine the query similarity of the metadata of each digital object in the retrieval result dataset to the retrieval text, and to determine the pragmatic weight value of each digital object in the retrieval result dataset. The evaluation and ranking module is used to determine the evaluation value of each digital object in the search result dataset based on the query similarity and pragmatic weight value of each digital object in the search result dataset, so as to obtain the ranking result of the digital objects in the search result dataset; The parameter calculation module includes: The modeling submodule is used to establish a Markov chain model to characterize the pragmatic relationships between digital objects in the retrieval result dataset. The established Markov chain model defines the pragmatic weight values ​​between digital objects with direct relationships as the association probability values ​​between the digital objects with direct relationships in the Markov chain model, and defines the weight numerical relationship between the pragmatic weight values ​​between digital objects without direct relationships as the probability numerical relationship between the association probability values ​​between the digital objects without direct relationships in the Markov chain model. The modeling and solving submodule is used to solve the Markov chain model to obtain the pragmatic weight value of each of the digital objects.

8. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the data usage-based network search and ranking method as described in any one of claims 1 to 6.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the data-based network search and ranking method as described in any one of claims 1 to 6.

Images