Service data space-time label generation method and system based on space entity

Abstract

The invention relates to the field of data retrieval, in particular to a business data space-time label generation method and system based on a space entity. The method has the beneficial effects that the effective label is set for the business data by considering and identifying the anchoring word in the business data, screening the effective anchoring set, determining the anchoring characteristic, calculating the information description effective parameter for the business data based on the anchoring characteristic; and adaptively generating feedback data for different service data based on the validity labels subsequently. When a large amount of service data is processed, the quality of different service data is defective and irregular, and the accuracy of the feedback data is improved by analyzing the validity of the service data and adaptively selecting different strategies for generating the feedback data, so that manual inquiry is reduced, and the availability of the service data is improved.

Description

Technical Field

[0001] This invention relates to the field of data retrieval, and in particular to a method and system for generating spatiotemporal labels for business data based on spatial entities. Background Technology

[0002] With the rapid development of the internet, various online business systems have sprung up. Taking urban governance as an example, work order texts often mix geographical location, time requirements, and information descriptions. Faced with this situation, traditional manual processing methods, due to their inherent defects such as low efficiency and slow response, can no longer meet the current urgent needs of business for rapid response and efficient processing. In addition, problems such as spatiotemporal mismatch and duplicate annotation can also generate a lot of erroneous feedback. Therefore, some business processing technologies based on large models have emerged.

[0003] For example, Chinese Patent Publication No. CN119990285A discloses a smart city rule engine system and method based on a large model. Applied to the field of IoT technology, it includes: a data acquisition and receiving port, a text and voice analysis module, an intent command recognition module, a rule chain module, and a smart city business driving module. The method includes: performing intent recognition processing on client text and voice data to obtain intent recognition results; parsing the intent recognition results to obtain a command parsing list; combining the intent recognition results and the command parsing list to form a command set, and generating a rule chain that conforms to the logical description of the rule engine; and automatically executing business scenarios based on the rule chain and business data. In this way, this invention automatically generates and optimizes rule chains through large model technology, and automatically processes business scenarios through the rule engine. Compared with traditional rule engine systems, it has stronger intelligence, automation, and scalability.

[0004] However, the following problems still exist in the existing technology.

[0005] 1. In reality, different users, based on their own habits and cognitive levels, will provide business data of varying quality, some of which may contain numerous potential keywords. Current technologies do not consider filtering out interfering and invalid keywords, resulting in high computational demands and poor analysis results.

[0006] 2. In existing technologies, there is a lack of closed-loop optimization mechanisms based on data validity feedback, and some ambiguous business data often require manual inquiry to supplement information; Summary of the Invention

[0007] To address these issues, this invention provides a method for generating spatiotemporal labels for business data based on spatial entities. This method overcomes the problems in existing technologies, such as the existence of numerous potential keywords leading to huge computational demands, poor analysis results, lack of a closed-loop optimization mechanism based on data validity feedback, the need for manual inquiry to supplement information for some ambiguous business data, and the failure to fully explore and utilize the actual information contained in the business data.

[0008] To achieve the above objectives, the present invention provides a method for generating spatiotemporal tags for business data based on spatial entities, comprising:

[0009] Acquire business data uploaded by users and perform spatiotemporal tagging processing on the business data, including setting spatiotemporal semantic tags for the business data;

[0010] Identify anchor words in business data, construct several reference anchor sets based on anchor words, filter effective anchor sets based on the combination relationship of keywords in the reference anchor sets, and determine anchor features, including determining the number of effective anchor sets and the semantic matching degree between effective anchor sets.

[0011] Based on anchored features, calculate effective parameters for information description of business data and set validity labels for business data;

[0012] Based on the non-validity labels, business data is analyzed to generate feedback data, including:

[0013] Extract invalid business data, determine core semantic clusters based on the semantic correlation between anchor words, determine the search range based on the effective anchor sets corresponding to each anchor word in the core semantic clusters, find missing keywords in the historical database based on each keyword in the search range, generate feedback statements based on the missing keywords, and send the feedback statements to the user.

[0014] The system receives response data from the user, tags it, combines it with business data, and stores it in the business database.

[0015] Furthermore, anchor words are identified in the business data, and several reference anchor sets are constructed based on these anchor words, including:

[0016] Perform semantic segmentation and keyword extraction on the text description of business data to identify several keywords;

[0017] Each keyword is compared with the anchor term database to determine the anchor terms;

[0018] Based on anchor words and adjacent keywords, determine the reference anchor set;

[0019] The anchor terminology is pre-defined.

[0020] Furthermore, the process of selecting effective anchor sets based on the combination relationships of keywords within the reference anchor set includes:

[0021] Determine the first tag combination of anchor words and adjacent keywords;

[0022] Determine the combination of anchor words and adjacent keywords in the second tag;

[0023] Valid values ​​are calculated based on the occurrence probabilities of the first and second tag combinations in the historical database;

[0024] Based on the comparison results between the valid values ​​and the valid threshold, it is determined whether the reference anchor set is a valid anchor set.

[0025] Furthermore, the process of calculating the effective parameters for the information description of the business data includes,

[0026] The ratio of the number of anchor sets to the predetermined threshold number of anchor sets is used as a quantitative consideration factor.

[0027] The ratio of the mean semantic matching degree to the predetermined semantic matching degree threshold is calculated as a semantic consideration factor;

[0028] We obtain the effective parameters for information description by weighting and summing the quantitative and semantic factors.

[0029] The semantic matching degree is determined based on the semantic relevance between keywords within the effective anchor set.

[0030] Furthermore, the process of setting validity labels for business data includes,

[0031] If the valid parameters of the information description are greater than the preset validity threshold, then a validity label is set for the business data;

[0032] If the valid parameters of the information description are less than or equal to the preset validity threshold, then a non-valid label is set for the business data.

[0033] Further, the process of extracting invalid business data, determining core semantic clusters based on the semantic relevance between anchor words, and determining the narrowing range based on the effective anchor sets corresponding to each anchor word within the core semantic clusters includes:

[0034] Clustering is performed based on the semantic relevance between anchor words to identify several anchor words that meet the clustering conditions, forming a core semantic cluster composed of each anchor word and the keywords between the anchor words;

[0035] Extract all valid anchor sets from the core semantic clusters, and use each of these valid anchor sets as the search scope;

[0036] The clustering condition is that the semantic correlation between anchor words is greater than a predetermined anchor semantic correlation threshold.

[0037] Furthermore, it also includes analyzing business data based on the aforementioned invalidity tags to generate feedback data, wherein,

[0038] Extract valid business data, traverse each valid anchor set, and use each valid anchor set as the search scope;

[0039] The system searches for missing keywords in the historical database based on the keywords within the search range, generates a feedback statement based on the missing keywords, and sends the feedback statement to the user.

[0040] Furthermore, the process of finding missing keywords in the historical database based on the keywords within the search scope, and generating feedback statements based on the missing keywords, includes:

[0041] In the historical database, keywords that have a preset relationship with all keywords within the search scope are retrieved to form a potential set of related keywords;

[0042] From the set of potentially related keywords, exclude keywords included in the search scope to obtain the set of missing keywords;

[0043] Generate a response statement in the form of a query that includes the missing keywords;

[0044] Specifically, if the semantic relevance between keywords is greater than a predetermined semantic relevance threshold and the keyword is a noun, then a relevance relationship is determined to exist.

[0045] Furthermore, the process of setting spatiotemporal semantic tags for the business data includes,

[0046] Identify the time and location of business data, and define the time and location as spatiotemporal semantic tags.

[0047] On the other hand, a system for generating spatiotemporal labels for business data based on spatial entities is also provided, which includes,

[0048] The data acquisition module is used to acquire business data uploaded by the user terminal and perform spatiotemporal tagging processing on the business data, including setting spatiotemporal semantic tags for the business data.

[0049] The anchoring identification module is used to identify anchor words in business data, construct several reference anchor sets based on anchor words, filter effective anchor sets based on the combination relationship of keywords in the reference anchor sets, and determine anchoring features, including determining the number of effective anchor sets and determining the semantic matching degree between effective anchor sets.

[0050] The tag setting module is used to calculate the effective parameters of the information description for the business data based on the anchoring features, and to set effective tags for the business data.

[0051] The data analysis module is used to extract invalid business data, determine core semantic clusters based on the semantic correlation between anchor words, determine the search range based on the effective anchor set corresponding to each anchor word in the core semantic cluster, find missing keywords in the historical database based on each keyword in the search range, generate feedback statements based on the missing keywords, and send the feedback statements to the user terminal.

[0052] The storage module is used to receive response data from the user terminal and combine it with the business data to store it in the business database.

[0053] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention identifies anchor words in business data, filters effective anchor sets, determines anchor features, calculates effective parameters for information description of business data based on anchor features, sets validity labels for business data, and subsequently generates feedback data adaptively based on these validity labels for different types of business data. This invention considers the inconsistent quality of different business data when processing large amounts of data. By analyzing the validity of business data, it adaptively selects different strategies for generating feedback data, improving the accuracy of feedback data, thereby reducing manual inquiries and improving the usability of business data.

[0054] In particular, in practice, the quality of business data uploaded by users varies greatly. Some business data contains a large number of words unrelated to core business information, failing to reflect the true situation. Business personnel usually need to conduct follow-up inquiries to obtain supplementary explanations from users. Based on this, this invention filters effective anchor sets. The effective anchor sets are selected from reference anchor sets. The reference anchor sets consist of anchor words and adjacent keywords. Anchor words are usually highly related to business, and adjacent words are usually combined with business data to describe semantics. Therefore, setting reference anchor sets extracts only some semantic elements to consider the description of business data, initially reflecting the semantic elements in business data that have potential connections to the business. Subsequently, the effective anchor sets are filtered by splitting the reference anchor sets to form tag combinations. Further filtering of reference anchor sets that are semantically related and associated with the main business yields the effective anchor sets. Subsequently, the semantic matching degree between effective anchor sets is used to consider the effectiveness of the information description of business data, providing support for setting effectiveness tags, facilitating the identification of missing keywords, forming feedback statements, guiding users to provide feedback to complete business data, and ensuring the accuracy of business data.

[0055] In particular, the presence of numerous interfering words and invalid keywords in invalid business data can interfere with judgments, reduce accuracy, and waste computational resources. This invention addresses the reduction of invalid business data. It extracts invalid business data, determines core semantic clusters based on the semantic correlation between anchor words, and determines the reduction range based on the effective anchor sets corresponding to each anchor word within the core semantic clusters. This invention addresses the large number of potential keywords contained in invalid business data. It considers the relationships between various anchor words, excluding occasional anchor words that lack relevance to the overall content. It analyzes anchor words with close semantic relationships and high frequency of occurrence, as well as their corresponding effective anchor sets, thereby eliminating interfering and invalid keywords. Furthermore, based on this, it analyzes missing keywords to characterize potential phenomena related to the business data, generating query-style feedback to guide users in supplementing the business data. For example, the business data may contain descriptions of air conditioning conditions, but based on historical data, descriptions of some conditions are missing. Therefore, guiding users to supplement the business data reduces manual inquiries, obtains more comprehensive business data, and improves the reference value of the business data.

[0056] In particular, for valid business data, this invention extracts valid business data, traverses each valid anchor set, extracts keywords contained in the valid anchor sets, and directly constructs the search range. This method is based on the fact that the valid business data itself already has high semantic integrity and anchor feature validity. Therefore, there is no need to perform complex core semantic clustering and search range reduction processes. It directly uses the verified keywords in the valid anchor sets to construct the search range, thereby quickly locating missing information and generating feedback, ensuring the reliability of feedback data and improving the efficiency of business data processing. Attached Figure Description

[0057] Figure 1 This is a step diagram of the method for generating spatiotemporal tags for business data based on spatial entities, as described in an embodiment of the invention.

[0058] Figure 2 This is a logic block diagram illustrating the analysis of business data based on the aforementioned validity tags, as an embodiment of the invention.

[0059] Figure 3 This is a logic block diagram for anchor word determination in an embodiment of the invention.

[0060] Figure 4 A logic block diagram for filtering the valid anchor set in an embodiment of the invention; Detailed Implementation

[0061] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0062] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0063] It should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0064] Please see Figure 1 As shown, Figure 1 This is a flowchart illustrating the steps of a method for generating spatiotemporal tags for business data based on spatial entities, according to an embodiment of the present invention. The present invention provides a method for generating spatiotemporal tags for business data based on spatial entities, comprising:

[0065] Step S1: Obtain business data uploaded by the user terminal and perform spatiotemporal tagging processing on the business data, including setting spatiotemporal semantic tags for the business data;

[0066] Step S2: Identify anchor words in business data, construct several reference anchor sets based on anchor words, filter effective anchor sets based on the combination relationship of keywords in the reference anchor sets, and determine anchor features, including determining the number of effective anchor sets and determining the semantic matching degree between effective anchor sets.

[0067] Step S3: Calculate the effective parameters for the information description of the business data based on the anchored features, and set the validity labels for the business data;

[0068] Step S4: Analyze the business data based on the non-validity labels and generate feedback data, including...

[0069] Extract invalid business data, determine core semantic clusters based on the semantic correlation between anchor words, determine the search range based on the effective anchor sets corresponding to each anchor word in the core semantic clusters, find missing keywords in the historical database based on each keyword in the search range, generate feedback statements based on the missing keywords, and send the feedback statements to the user.

[0070] Step S5: Receive the response data from the user terminal, tagged it, and combine it with the business data to store it in the business database.

[0071] Specifically, there are no restrictions on the format of business data; business data can be in text or voice format. For business data in voice format, it needs to be converted into text before processing.

[0072] Specifically, please refer to Figure 2 As shown, Figure 2 The logical block diagram for analyzing business data based on the validity labels in this embodiment of the invention further includes analyzing business data based on the non-valid labels and generating feedback data, wherein...

[0073] Extract valid business data, traverse each valid anchor set, and use each valid anchor set as the search scope;

[0074] The system searches for missing keywords in the historical database based on the keywords within the search range, generates a feedback statement based on the missing keywords, and sends the feedback statement to the user.

[0075] Specifically, the process of setting spatiotemporal semantic tags for business data includes,

[0076] Identify the time and location of business data, and define the time and location as spatiotemporal semantic tags.

[0077] Please see Figure 3 As shown, Figure 3 This is a logical block diagram for anchor word determination in an embodiment of the invention. It identifies anchor words in business data and constructs several reference anchor sets based on these anchor words, including...

[0078] Perform semantic segmentation and keyword extraction on the text description of business data to identify several keywords;

[0079] Each keyword is compared with the anchor term database to determine the anchor terms;

[0080] Based on anchor words and adjacent keywords, determine the reference anchor set;

[0081] The anchor terminology is pre-defined.

[0082] Specifically, keywords are the original vocabulary set obtained from the textual description of business data through semantic segmentation and extraction techniques. They cover elements such as entities, actions, and attributes related to the business (e.g., "Haidian District," "drainage," "fault," "2024"). Their role is to initially condense the semantic elements of the business data, providing basic input for subsequent analysis.

[0083] Specifically, anchor words are determined through a pre-built database that combines the spatial entity characteristics of historical business data with their relevance to business scenarios. Those skilled in the art can analyze the business data in the historical database to extract frequently occurring keywords that are highly relevant to the business data, or they can manually set keywords related to the business scenario.

[0084] For example, if the main business is air conditioner maintenance, the anchor keyword library can be set with several keywords related to air conditioner maintenance, such as water leakage, power outage, and cooling, etc., which will not be elaborated further.

[0085] Specifically, the process of determining the reference anchor set based on the anchor term and its preceding and following keywords is as follows:

[0086] Perform semantic dependency parsing on business data text to locate anchor words in the sentence;

[0087] Adjacent keywords: Select the first non-auxiliary word / preposition keyword to the right of the anchor word (e.g., "fault" after the anchor word "power distribution room");

[0088] Preceding keyword: Select the first non-particle / preposition keyword to the left of the anchor word (e.g., "Haidian District" preceded by "flood season");

[0089] Among them, adjacent words must be in the same semantic clause as anchor words, and cross-sentence keywords will not be captured;

[0090] The preceding adjacent keywords, anchor words, and following adjacent keywords are determined as the corresponding reference anchor sets.

[0091] It is understandable that within the same semantic clause

[0092] If only subsequent adjacent keywords exist (such as anchor words at the beginning of a sentence), then a binary reference anchor set is generated;

[0093] If only the preceding adjacent keyword exists (such as the anchor word being at the end of the sentence), then a binary reference anchor set is generated;

[0094] Isolated anchors without adjacent words are considered invalid.

[0095] Understandably, in actual usage, keywords that appear consecutively with anchor words are usually used to describe or limit those anchor words. Therefore, setting a reference anchor set captures more elements that are relevant to the business.

[0096] Please see Figure 4 As shown, Figure 4 This is a logic block diagram for filtering valid anchor sets according to an embodiment of the invention. The process of filtering valid anchor sets based on the combination relationships of keywords within a reference anchor set includes:

[0097] Determine the first tag combination of anchor words and adjacent keywords;

[0098] Determine the combination of anchor words and adjacent keywords in the second tag;

[0099] Valid values ​​are calculated based on the occurrence probabilities of the first and second tag combinations in the historical database;

[0100] Based on the comparison results between the valid values ​​and the valid threshold, it is determined whether the reference anchor set is a valid anchor set.

[0101] Specifically, querying the occurrence probabilities of the first tag combination (the semantic relationship between the anchor term and its adjacent keywords) and the second tag combination (the semantic relationship between the anchor term and its preceding adjacent keywords) in the historical database essentially involves statistically analyzing the co-occurrence probabilities of the anchor term and its corresponding adjacent keywords in historical independent semantic clauses. The probability of any two keywords appearing in the same semantic clause can be obtained in advance by traversing and statistically analyzing the independent semantic clauses in the historical database to determine the occurrence probability of the first tag combination or the second tag combination. In practice, the average occurrence probability of the first tag combination and the second tag combination is determined as the valid value.

[0102] Specifically, the effective threshold is predetermined, with the aim of determining that the keywords within the reference anchor set have a strong correlation. In practice, several reference anchor sets in the historical database are statistically analyzed in advance, and the average occurrence probability of the reference anchor sets is calculated to represent the occurrence probability of each reference anchor set in the corpus under normal circumstances. In practice, the product of the average occurrence probability and the error coefficient is set as the effective threshold, and the error coefficient is in the range [0.85, 0.95].

[0103] It is understandable that if the binary reference anchor set is considered to exist only for the first tag combination, the probability of the first tag combination appearing is determined as an effective value, which will not be elaborated further.

[0104] Specifically, the historical database is predetermined. Those skilled in the art can pre-obtain several business data containing anchor words to construct the historical database. It is understood that the business data in the historical database should be filtered to remove business data with semantic and logical errors, which will not be elaborated further.

[0105] In practice, the quality of business data uploaded by users varies greatly. Some business data contains a large number of words unrelated to core business information, failing to reflect the true situation. Business personnel usually need to conduct follow-up inquiries to obtain supplementary explanations from users. Based on this, this invention filters effective anchor sets. The effective anchor sets are selected from a reference anchor set, which consists of anchor words and adjacent keywords. Anchor words are usually highly related to business, and adjacent words are usually combined with business data to describe semantics. Therefore, setting a reference anchor set extracts only some semantic elements to consider the description of business data, initially reflecting the semantic elements in the business data that have potential connections to the business. Subsequently, the effective anchor set is filtered by splitting the reference anchor set to form tag combinations. Further filtering of reference anchor sets that are semantically related and connected to the main business yields the effective anchor set. Subsequently, the semantic matching degree between effective anchor sets is used to consider the effectiveness of the information description of business data, providing support for setting effectiveness tags, facilitating the identification of missing keywords, forming feedback statements, guiding users to provide feedback to complete business data, and ensuring the accuracy of business data.

[0106] Specifically, the process of calculating effective parameters for the information description of the business data includes,

[0107] The ratio of the number of anchor sets to the predetermined threshold number of anchor sets is used as a quantitative consideration factor.

[0108] The ratio of the mean semantic matching degree to the predetermined semantic matching degree threshold is calculated as a semantic consideration factor;

[0109] We obtain the effective parameters for information description by weighting and summing the quantitative and semantic factors.

[0110] The semantic matching degree is determined based on the semantic relevance between keywords within the effective anchor set.

[0111] Calculate the semantic relevance between keywords in a single effective anchor set and keywords in another effective anchor set, solve for the mean semantic relevance, set the mean semantic relevance as the semantic matching degree, and then solve for the mean semantic matching degree.

[0112] In practice, the threshold for the number of anchor sets is predetermined. The average number of anchor sets in each business data of the historical database is used as a reference. The threshold for the number of anchor sets is set to be the product of the average number and the offset coefficient, with the offset coefficient in the range [0.75, 0.85].

[0113] The predetermined semantic matching threshold is determined in advance. Specifically, the average semantic matching degree between each reference anchor set in each business data in the historical database is pre-calculated, and the semantic matching threshold is set to be the product of the average semantic matching degree and the offset coefficient.

[0114] When the quantitative factors and semantic factors are weighted and summed, the weights are 0.6 and 0.4, respectively.

[0115] Specifically, the process of setting validity labels for business data includes,

[0116] If the valid parameters of the information description are greater than the preset validity threshold, then a validity label is set for the business data;

[0117] If the valid parameters of the information description are less than or equal to the preset validity threshold, then a non-valid label is set for the business data.

[0118] Specifically, the validity threshold is predetermined. The validity threshold is determined by the information description parameters calculated when the number of anchor sets is equal to the predetermined anchor set number threshold and the average semantic matching degree is equal to the predetermined semantic matching degree threshold.

[0119] Specifically, the process of extracting invalid business data, determining core semantic clusters based on the semantic relevance between anchor words, and determining the narrowing range based on the effective anchor sets corresponding to each anchor word within the core semantic clusters includes:

[0120] Clustering is performed based on the semantic relevance between anchor words to identify several anchor words that meet the clustering conditions, forming a core semantic cluster composed of each anchor word and the keywords between the anchor words;

[0121] Extract all valid anchor sets from the core semantic clusters, and use each of these valid anchor sets as the search scope;

[0122] The clustering condition is that the semantic correlation between anchor words is greater than a predetermined anchor semantic correlation threshold.

[0123] Specifically, there are no restrictions on how the semantic relevance between anchor words is calculated. For example, the anchor words can be vectorized and then the cosine similarity can be calculated, with the cosine similarity used as the semantic relevance.

[0124] The anchor semantic association threshold is predetermined. The average semantic association degree between anchor words in each business data in the historical database is predetermined, and the anchor semantic association threshold is set as the average semantic association degree.

[0125] Specifically, the process of finding missing keywords in the historical database based on keywords within the search scope, and generating feedback statements based on these missing keywords, includes:

[0126] In the historical database, keywords that have a preset relationship with all keywords within the search scope are retrieved to form a potential set of related keywords;

[0127] From the set of potentially related keywords, exclude keywords included in the search scope to obtain the set of missing keywords;

[0128] Generate a response statement in the form of a query that includes the missing keywords;

[0129] Specifically, if the semantic relevance between keywords is greater than a predetermined semantic relevance threshold and the keyword is a noun, then a relevance relationship is determined to exist.

[0130] The semantic relevance threshold is selected between 1.2 and 1.5 times the anchor semantic relevance threshold.

[0131] In practice, fault types are usually described in noun form and combined with other keywords. Therefore, it is important to consider nouns that are highly related to keywords. For example, the keyword "air conditioner" and another keyword "electric leakage" can usually describe the fault phenomenon.

[0132] During implementation, query statements are generated for missing keywords, such as whether there is a power leak, so that users can provide sufficient information and avoid repeated questioning of users. This allows for the direct acquisition of more comprehensive information, providing comprehensive data support for subsequent order dispatching, maintenance, and other derivative services.

[0133] Because invalid business data contains a large number of interfering words and invalid keywords, which can interfere with judgment, reduce accuracy, and waste computing power, this invention addresses the issue of reducing invalid business data. It extracts invalid business data, determines core semantic clusters based on the semantic correlation between anchor words, and determines the reduction range based on the effective anchor sets corresponding to each anchor word within the core semantic clusters. This invention addresses the large number of potential keywords contained in invalid business data by considering the correlation between anchor words. It excludes occasional anchor words that lack relevance to the overall content, and analyzes anchor words with close semantic relationships and high frequency of occurrence, as well as their corresponding effective anchor sets. This process eliminates interfering words and invalid keywords. Furthermore, it analyzes missing keywords to characterize potential phenomena related to business data, generating query-style feedback information to guide users to supplement business data. For example, business data may contain descriptions of air conditioning conditions, but based on historical data, descriptions of some conditions are missing. Therefore, guiding users to supplement business data reduces manual inquiries, obtains more comprehensive business data, and improves the referenceability of the business data.

[0134] On the other hand, a system for generating spatiotemporal labels for business data based on spatial entities is also provided, including,

[0135] The data acquisition module is used to acquire business data uploaded by the user terminal and perform spatiotemporal tagging processing on the business data, including setting spatiotemporal semantic tags for the business data.

[0136] The anchoring identification module is used to identify anchor words in business data, construct several reference anchor sets based on anchor words, filter effective anchor sets based on the combination relationship of keywords in the reference anchor sets, and determine anchoring features, including determining the number of effective anchor sets and determining the semantic matching degree between effective anchor sets.

[0137] The tag setting module is used to calculate the effective parameters of the information description for the business data based on the anchoring features, and to set effective tags for the business data.

[0138] The data analysis module is used to extract invalid business data, determine core semantic clusters based on the semantic correlation between anchor words, determine the search range based on the effective anchor set corresponding to each anchor word in the core semantic cluster, find missing keywords in the historical database based on each keyword in the search range, generate feedback statements based on the missing keywords, and send the feedback statements to the user terminal.

[0139] The storage module is used to receive response data from the user terminal and combine it with the business data to store it in the business database.

[0140] Specifically, there are no restrictions on the specific structure of the data acquisition module, anchoring identification module, tag setting module, and data analysis module. They can all be composed of logic components, including field-programmable processors, computers, or microprocessors in computers.

[0141] The specific structure of the storage module is not limited. The storage module can be a cloud-based or physical database, which will not be elaborated further.

[0142] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for generating spatiotemporal tags for business data based on spatial entities, characterized in that, include: Acquire business data uploaded by users and perform spatiotemporal tagging processing on the business data, including setting spatiotemporal semantic tags for the business data; Identify anchor words in business data, construct several reference anchor sets based on anchor words, filter effective anchor sets based on the combination relationship of keywords in the reference anchor sets, and determine anchor features, including determining the number of effective anchor sets and the semantic matching degree between effective anchor sets. Based on anchored features, calculate effective parameters for information description of business data and set validity labels for business data; Based on the non-validity labels, business data is analyzed to generate feedback data, including: Extract invalid business data, determine core semantic clusters based on the semantic correlation between anchor words, determine the search range based on the effective anchor sets corresponding to each anchor word in the core semantic clusters, find missing keywords in the historical database based on each keyword in the search range, generate feedback statements based on the missing keywords, and send the feedback statements to the user. The system receives response data from the user, tags it, combines it with business data, and stores it in the business database.

2. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 1, characterized in that, Identify anchor words in business data, and construct several reference anchor sets based on these anchor words, including: Perform semantic segmentation and keyword extraction on the text description of business data to identify several keywords; Each keyword is compared with the anchor term database to determine the anchor terms; Based on anchor words and adjacent keywords, determine the reference anchor set; The anchor terminology is pre-defined.

3. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 2, characterized in that, The process of selecting effective anchor sets based on the combination relationships of keywords within the reference anchor set includes: Determine the first tag combination of anchor words and adjacent keywords; Determine the combination of anchor words and adjacent keywords in the second tag; Valid values ​​are calculated based on the occurrence probabilities of the first and second tag combinations in the historical database; Based on the comparison results between the valid values ​​and the valid threshold, it is determined whether the reference anchor set is a valid anchor set.

4. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 1, characterized in that, The process of calculating the effective parameters for the information description of the business data includes, The ratio of the number of anchor sets to the predetermined threshold number of anchor sets is used as a quantitative consideration factor. The ratio of the mean semantic matching degree to the predetermined semantic matching degree threshold is calculated as a semantic consideration factor; We obtain the effective parameters for information description by weighting and summing the quantitative and semantic factors. The semantic matching degree is determined based on the semantic relevance between keywords within the effective anchor set.

5. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 1, characterized in that, The process of setting validity labels for business data includes, If the valid parameters of the information description are greater than the preset validity threshold, then a validity label is set for the business data; If the valid parameters of the information description are less than or equal to the preset validity threshold, then a non-valid label is set for the business data.

6. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 1, characterized in that, The process of extracting invalid business data, determining core semantic clusters based on the semantic relevance between anchor words, and determining the narrowing range based on the effective anchor sets corresponding to each anchor word within the core semantic clusters includes: Clustering is performed based on the semantic relevance between anchor words to identify several anchor words that meet the clustering conditions, forming a core semantic cluster composed of each anchor word and the keywords between the anchor words; Extract all valid anchor sets from the core semantic clusters, and use each of these valid anchor sets as the search scope; The clustering condition is that the semantic correlation between anchor words is greater than a predetermined anchor semantic correlation threshold.

7. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 5, characterized in that, It also includes analyzing business data based on the aforementioned invalidity tags to generate feedback data, wherein... Extract valid business data, traverse each valid anchor set, and use each valid anchor set as the search scope; The system searches for missing keywords in the historical database based on the keywords within the search range, generates a feedback statement based on the missing keywords, and sends the feedback statement to the user.

8. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 1, characterized in that, The process of finding missing keywords in the historical database based on keywords within the search scope, and generating feedback statements based on these missing keywords, includes: In the historical database, keywords that have a preset relationship with all keywords within the search scope are retrieved to form a potential set of related keywords; From the set of potentially related keywords, exclude keywords included in the search scope to obtain the set of missing keywords; Generate a response statement in the form of a query that includes the missing keywords; Specifically, if the semantic relevance between keywords is greater than a predetermined semantic relevance threshold and the keyword is a noun, then a relevance relationship is determined to exist.

9. The method for generating spatiotemporal tags for business data based on spatial entities according to claim 1, characterized in that, The process of setting spatiotemporal semantic tags for the business data. include, Identify the time and location of business data, and define the time and location as spatiotemporal semantic tags.

10. A system for generating spatiotemporal tags for business data based on spatial entities according to any one of claims 1-9, characterized in that, include, The data acquisition module is used to acquire business data uploaded by the user terminal and perform spatiotemporal tagging processing on the business data, including setting spatiotemporal semantic tags for the business data. The anchoring identification module is used to identify anchor words in business data, construct several reference anchor sets based on anchor words, filter effective anchor sets based on the combination relationship of keywords in the reference anchor sets, and determine anchoring features, including determining the number of effective anchor sets and determining the semantic matching degree between effective anchor sets. The tag setting module is used to calculate the effective parameters of the information description for the business data based on the anchoring features, and to set effective tags for the business data. The data analysis module is used to extract invalid business data, determine core semantic clusters based on the semantic correlation between anchor words, determine the search range based on the effective anchor set corresponding to each anchor word in the core semantic cluster, find missing keywords in the historical database based on each keyword in the search range, generate feedback statements based on the missing keywords, and send the feedback statements to the user terminal. The storage module is used to receive response data from the user terminal and combine it with the business data to store it in the business database.

Images