Regional earthquake safety evaluation multi-source heterogeneous data storage service system

Abstract

The invention discloses a regional earthquake safety evaluation multi-source heterogeneous data warehousing service system, which comprises a data set acquisition module, a triple determination module and an advanced matrix construction module, and is used for constructing a three-element advanced matrix according to N double-label data triples; the evaluation label determination module is used for determining an evaluation behavior label from the three-element advanced matrix; wherein the evaluation behavior tag is a binary tag and is used for marking the evaluation validity of adjacent data pairs, and the evaluation behavior tag has the values of'reasonable evaluation 'and'unnecessary evaluation'; the storage matrix construction module is used for constructing a multi-source heterogeneous data storage matrix according to the evaluation behavior tag and a three-element advanced matrix; the server generation module is used for storing the multi-source heterogeneous data storage matrix to a server to generate a multi-source data server; according to the method, redundant earthquake safety evaluation result data behaviors can be automatically judged, and ineffective information can be effectively controlled to enter into a database.

Description

Technical Field

[0001] The invention relates to a data storage service system, in particular to a regional earthquake safety evaluation multi-source heterogeneous data storage service system. Background Art

[0002] At present, the data types involved in regional earthquake safety assessment work show significant multi-source and heterogeneous characteristics, including but not limited to geological surveys, geophysical surveys, drilling analysis, seismic activity assessments, and multi-probability ground motion parameters. These data have great differences in source, structure, expression, and storage format, typically manifested as: geographic spatial data (points, lines, and surfaces), tabular data, map data, raw data files, and other types of coexistence, and are usually managed using ArcGIS, MapInfo, and other proprietary databases or file systems and other different carriers, resulting in inconsistent data structures, confusing field naming, high information duplication rates, and difficulty in structural comparison, which seriously restricts the efficiency of subsequent data storage, structural integration, analysis and evaluation, and query calls.

[0003] The patent document with patent publication number CN115964360B discloses a method and system for constructing an earthquake safety evaluation database, which can obtain classification results for representing earthquake safety level labels of the area to be evaluated. However, in the prior art, the management method for earthquake-related multi-source data lacks dynamic identification and organization means for data structure differences and time series relationships, and also lacks a quantitative judgment mechanism for data redundancy and evaluation necessity, resulting in problems such as repeated data storage, accumulation of invalid information, and waste of evaluation resources. In addition, there is currently a lack of a complete service method that can unify the label system, build a structured matrix, and support behavior identification and storage decision-making, which makes it difficult to meet the technical requirements for systematic, standardized and automated data processing in earthquake safety evaluation. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a regional earthquake safety assessment multi-source heterogeneous data warehousing service system, which solves the technical problems raised in the background technology by constructing a multi-source heterogeneous data warehousing matrix containing evaluation behavior labels.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A regional earthquake safety assessment multi-source heterogeneous data storage service system, including: A data set acquisition module is used to acquire a data set consisting of N multi-source heterogeneous data; A triplet determination module, used to determine N double-label data triplets according to the data set; An advanced matrix construction module, used to construct a three-element advanced matrix based on N double-label data triplets; An evaluation label determination module, used to determine an evaluation behavior label from the three-element advanced matrix; wherein the evaluation behavior label is a binary label used to mark the evaluation validity of adjacent data pairs, and its value is: "reasonable evaluation" and "unnecessary evaluation"; A storage matrix construction module is used to construct a multi-source heterogeneous data storage matrix based on the evaluation behavior labels and the three-element advanced matrix; A server generation module, used for storing the multi-source heterogeneous data storage matrix in a server to generate a multi-source data server; When the client sends a data query request, the multi-source data server responds to the data query request and generates a query vector; The multi-source data server receives the query vector as input, performs feature extraction on the multi-source heterogeneous data storage matrix, and outputs data features based on the data query request.

[0006] In some of the embodiments, the triplet determination module is specifically used for: S2-1, determining the stage continuous label and structure discrete label of the multi-source heterogeneous data from the data set; wherein the stage continuous label is used to characterize the collection stage of the multi-source heterogeneous data in the time dimension, and its minimum value is defined as the origin label; S2-2. Pair the multi-source heterogeneous data with their corresponding stage continuous labels and structure discrete labels to construct N double-label data triplets.

[0007] In some of the embodiments, determining the phase continuous labels and structure discrete labels of the multi-source heterogeneous data from the data set includes: S2-1-1, extract the marked collection timestamps of N multi-source heterogeneous data in the data set; S2-2-2. Assign stage-continuous labels to N multi-source heterogeneous data according to the time sequence of the acquisition timestamps; wherein the stage-continuous labels are used to characterize the acquisition stage of the multi-source heterogeneous data in the time dimension, and the minimum value thereof is defined as the origin label, corresponding to the multi-source heterogeneous data with the earliest acquisition time; S2-2-3, structural identification is performed on the data structures of N multi-source heterogeneous data to generate structural discrete labels; wherein the structural discrete labels are used to characterize the type attribution of the multi-source heterogeneous data at the structural level; In some of the embodiments, the advanced matrix building module is specifically used for: S3-1. Construct a matrix coordinate system in a two-dimensional plane, wherein the coordinate origin of the matrix coordinate system is represented by an origin label, and the coordinate Y axis is represented by a virtual connection axis of the stage continuous label; S3-2, performing matrix arrangement on the N double-label data triplets in the matrix coordinate system to construct a structure element matrix and a stage element matrix; S3-3, extracting an ordered sequence of N multi-source heterogeneous data from the structure element matrix; S3-4. Construct a three-element advanced matrix based on the ordered sequence of the multi-source heterogeneous data and the stage element matrix.

[0008] In some of the embodiments, performing matrix arrangement on N double-label data triplets in the matrix coordinate system to construct a structure element matrix and a phase element matrix includes: S3-2-1, extracting stage continuous labels and their corresponding multi-source heterogeneous data from N double-label data triplets; S3-2-2, fill the extracted stage continuous labels and their corresponding multi-source heterogeneous data into the first quadrant of the matrix coordinate system; wherein the origin label is located on the X-axis, and the stage continuous labels and their corresponding multi-source heterogeneous data are arranged equidistantly along the positive direction of the Y-axis; S3-2-3, defining the stage continuous labels in the first quadrant and their corresponding multi-source heterogeneous data as a structural element matrix; S3-2-4. Copy the elements of the second column vector in the structure element matrix to the second quadrant of the matrix coordinate system as the second column vector of the second quadrant; S3-2-5. Fill in the corresponding structural discrete labels according to the elements of the second column vector in the second quadrant as the first column vector of the second quadrant; S3-2-6. Define the structured discrete labels and their multi-source heterogeneous data located in the second quadrant as the stage element matrix.

[0009] In some of the embodiments, extracting an ordered sequence of N multi-source heterogeneous data from the structure element matrix includes: S3-3-1. The multi-source heterogeneous data corresponding to the origin labels in the structure element matrix are defined as the reference data, and the multi-source heterogeneous data corresponding to the non-origin labels are defined as the paired data; S3-3-2, associate each paired data with the reference data to obtain M heterogeneous data pairs, where M=N-1; A6-3. Calculate the content similarity of M heterogeneous data pairs; S3-3-4, sorting the M heterogeneous data pairs in descending order according to their content similarity values ​​to generate an ordered sequence of data pairs; S3-3-5, assigning sort numbers to the ordered sequence of data pairs; S3-3-6. Determine whether the sorting number of the heterogeneous data pair is a positive integer 1; if the sorting number is a positive integer 1, mark the corresponding heterogeneous data pair as the origin data pair; otherwise, mark it as a redundant data pair; S3-3-7. Delete all the reference data in the redundant data pairs, and define the data sequence of the M paired data and the origin data pairs after deletion as the ordered sequence of the multi-source heterogeneous data; wherein the ordered sequence of the multi-source heterogeneous data is characterized as an ordered sequence of the N multi-source heterogeneous data arranged in descending order based on content similarity.

[0010] In some of the embodiments, a three-element advanced matrix is ​​constructed according to the ordered sequence of the multi-source heterogeneous data and the stage element matrix, including: S3-4-1. Define the ordered sequence of N multi-source heterogeneous data as the first column vector of a three-element advanced matrix; S3-4-2, extracting the stage continuous labels corresponding to N multi-source heterogeneous data from the stage element matrix; S3-4-3. According to the correspondence between the first column vector of the three-element advanced matrix and the stage continuous labels, the stage continuous labels are rearranged to generate stage non-continuous labels; S3-4-4, define the stage non-continuous label as the second column vector of the three-element advanced matrix; S3-4-5. Fill in the content similarity according to the correspondence between the first column vector of the three-element advanced matrix and the content similarity, and define it as the third column vector of the three-element advanced matrix; wherein the matrix position corresponding to the reference data is filled with zero; S3-4-6. Pair the first column vector, the second column vector and the third column vector to construct the three-element advanced matrix.

[0011] In some embodiments, the evaluation tag determination module is specifically used to: S4-1, matching the first acquisition timestamp and the second acquisition timestamp of the adjacent stage continuous labels according to the stage continuous labels in the three-element advanced matrix; S4-2, defining the interval between the first acquisition timestamp and the second acquisition timestamp as an evaluation interval coefficient between adjacent multi-source heterogeneous data after normalization; S4-3, defining the product of the evaluation interval coefficient and the content similarity as the evaluation redundancy value; S4-4, comparing the evaluation redundancy value and the threshold value to generate a binary classification comparison result; S4-5. According to the binary classification comparison results, label the evaluation behavior labels of adjacent heterogeneous data pairs.

[0012] In some of the embodiments, based on the binary comparison results, the labeling conditions for labeling the evaluation behavior labels of adjacent heterogeneous data pairs are: If the result of the binary comparison is that the evaluation redundancy value is greater than the threshold, the data row corresponding to the second acquisition timestamp in the three-element advanced matrix is ​​marked as "unnecessary evaluation", and the data row corresponding to the first acquisition timestamp is marked as "reasonable evaluation"; otherwise, the data rows corresponding to the first acquisition timestamp and the second acquisition timestamp are both marked as "reasonable evaluation".

[0013] In some of the embodiments, a multi-source heterogeneous data storage matrix is ​​constructed based on the evaluation behavior labels and the three-element advanced matrix, including: S5-1, initializing the fourth column vector in the three-element advanced matrix; S5-2. Fill the evaluation behavior label into the fourth column vector to generate a multi-source heterogeneous data storage matrix.

[0014] The present invention provides a regional earthquake safety assessment multi-source heterogeneous data storage service system, which has the following beneficial effects: The present invention realizes the unified coding representation of multi-source heterogeneous data in the time dimension and the structural dimension by constructing a dual label system of stage continuous labels and structural discrete labels, and performs quadrant mapping of label information based on the matrix coordinate system to form a structural element matrix and a stage element matrix, and then constructs a three-element advanced matrix to complete the structural sorting, time classification and semantic alignment of the data. The label system and matrix construction method are not only suitable for data in various structural forms such as geographic spatial data, maps, tables, layers, etc., but also can uniformly organize and abstractly classify data from non-standard fields and heterogeneous sources, significantly improving the adaptability and accuracy of data standardization storage.

[0015] Based on data structure modeling, the present invention calculates the product of the content similarity of adjacent multi-source heterogeneous data and the collection time interval coefficient to form a reference index that can be used to judge repeatability and necessity, and uses this to label adjacent items in the data sequence with "reasonable evaluation" or "unnecessary evaluation". Specifically, when the collection time interval between two data is extremely short and their content expressions are highly similar, the system automatically determines that there is redundant evaluation behavior, only retains the data with an earlier time, and eliminates the duplicate data with a later time. This mechanism can dynamically identify "recent high-weight" redundant data and effectively control the influx of invalid information into the warehouse. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a structural block diagram of a regional earthquake safety assessment multi-source heterogeneous data storage service system of the present invention; Figure 2This is a data entry flow chart of a regional earthquake safety assessment multi-source heterogeneous data entry service system of the present invention; Figure 3 Schematic diagram of the labeling process of the evaluation behavior label of the present invention Figure 4 This is a schematic diagram of storing the regional evaluation results in the client in an embodiment of the present invention; Figure 5 The query diagram of the query request in the client in the embodiment of the present invention is shown in FIG. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1: Please refer to Figure 1-Figure 3 The present invention provides a regional earthquake safety assessment multi-source heterogeneous data storage service system, including: A data set acquisition module is used to acquire a data set consisting of N multi-source heterogeneous data; Exemplarily, in this embodiment, the multi-source heterogeneous data refers to regional evaluation data for regional earthquake safety evaluation.

[0019] According to the storage stage, it can be divided into actual material data and achievement data.

[0020] For actual material data, its contents may include: data from geological surveys, geophysical exploration, seismic activity analysis, drilling and engineering geological conditions analysis, and process data from earthquake hazard analysis.

[0021] For the results data, its contents may include: multi-probability seismic parameters of bedrock and surface of controlled drilling points, and multi-probability seismic parameter zoning map of target area.

[0022] Based on the data type, data can be divided into the following categories: Geographical spatial data: including points (Point), lines (Polyline) and surfaces (Polygon); Data table: that is, tabular data (Table); Figures: such as charts and graphs (Figure); Raw data: including data files and specialized formats (Data FileA, Apecialized FormatA), etc.

[0023] The geospatial data provided by different companies include data in ArcGIA format, data in other formats such as Mapinfo and MapGIA. These data are not standardized and unified in terms of feature coding and field rules. However, these data of different types, formats and storage methods are interrelated, forming a typical multi-source, heterogeneous data environment.

[0024] A triplet determination module, used to determine N double-label data triplets according to the data set; An advanced matrix construction module, used to construct a three-element advanced matrix based on N double-label data triplets; An evaluation label determination module, used to determine an evaluation behavior label from the three-element advanced matrix; wherein the evaluation behavior label is a binary label used to mark the evaluation validity of adjacent data pairs, and its value is: "reasonable evaluation" and "unnecessary evaluation"; A storage matrix construction module is used to construct a multi-source heterogeneous data storage matrix based on the evaluation behavior labels and the three-element advanced matrix; A server generation module, used for storing the multi-source heterogeneous data storage matrix in a server to generate a multi-source data server; When the client sends a data query request, the multi-source data server responds to the data query request and generates a query vector; The multi-source data server receives the query vector as input, performs feature extraction on the multi-source heterogeneous data storage matrix, and outputs data features based on the data query request.

[0025] Specifically, the data query request can be initiated based on any single or combined conditions such as structure tags, stage tags, evaluation behavior tags, etc., supporting precise query and fuzzy matching.

[0026] In this embodiment, by uniformly acquiring a data set consisting of N multi-source heterogeneous data, an original data foundation for regional earthquake safety assessment tasks is constructed. Among them, multi-source heterogeneous data are divided into actual material data and achievement data according to the storage stage, corresponding to the input support and result expression of the assessment process respectively; the former includes original data such as geological surveys, geophysical detection, and seismic activity analysis, and the latter covers model output results such as seismic parameters and zoning maps. The data types cover geospatial data (points, lines, and surfaces), tabular data, map data, and special format original data, and the sources include ArcGIS, Mapinfo, MapGIS and other system formats, showing typical structural heterogeneity and naming rule inconsistency problems. This embodiment provides comprehensive data support for subsequent label extraction, structural mapping, and assessment modeling through normalized access and structural classification processing of the above-mentioned complex heterogeneous data. Furthermore, the collected data will be included in the uniformly constructed storage matrix system to realize the semantic labeling organization of heterogeneous data, provide structured feature extraction capabilities for terminal query requests, and ultimately support the multi-source data server based on stage labels, structural labels, and assessment behavior labels for precise or fuzzy retrieval.

[0027] In this embodiment, the triplet determination module is specifically used for: S2-1, determining the stage continuous label and structure discrete label of the multi-source heterogeneous data from the data set; wherein the stage continuous label is used to characterize the collection stage of the multi-source heterogeneous data in the time dimension, and its minimum value is defined as the origin label; S2-2. Pair the multi-source heterogeneous data with their corresponding stage continuous labels and structure discrete labels to construct N double-label data triplets.

[0028] In this embodiment, by extracting the stage continuous labels and structural discrete labels of multi-source heterogeneous data and pairing them with the original data body, a set of double-label data triples consisting of N elements is constructed. The stage continuous label is used to identify the data collection stage in the time dimension, reflecting the temporal nature of the data; the structural discrete label is used to distinguish the data in the structural expression of the category, such as tables, maps, spatial data and other types. This step realizes the dual attribute calibration of the original heterogeneous data in the time dimension and the structural dimension, forming a unified labeled representation, so that the heterogeneous data can be orderly organized and compared before being stored, providing a data foundation for the final formation of a queryable, matchable, and reasonable storage matrix.

[0029] Further, determining the stage continuous labels and structure discrete labels of the multi-source heterogeneous data from the data set includes: S2-1-1. Extract the marked acquisition timestamps of N multi-source heterogeneous data in the data set; wherein, the acquisition timestamp is the time attribute attached to each multi-source heterogeneous data when it is constructed, which is used to characterize its acquisition time point; specifically, the acquisition timestamps of multi-source heterogeneous data have been marked by the original system, equipment or personnel during its generation or acquisition process. Therefore, in this step, there is no need to recalculate the acquisition time, and only read and extract the time attribute.

[0030] S2-2-2. Assign stage-continuous labels to N multi-source heterogeneous data according to the time sequence of the collection timestamps; wherein, the stage-continuous labels are used to characterize the collection stage of the multi-source heterogeneous data in the time dimension, and the minimum value is defined as the origin label, which corresponds to the multi-source heterogeneous data with the earliest collection time; the stage-continuous labels are essentially the stage numbers obtained after the standardized sorting of the collection timestamps, which are used to uniformly describe the relative collection stage of the data in the time dimension.

[0031] S2-2-3, structural identification is performed on the data structures of N multi-source heterogeneous data to generate structural discrete labels; wherein the structural discrete labels are used to characterize the type attribution of the multi-source heterogeneous data at the structural level; Among them, structure recognition refers to the process of identifying, classifying and categorizing the organizational structure, data representation and semantic features of multi-source heterogeneous data. Preferably, the structure type classification can be carried out according to the multi-dimensional features such as the data storage format, spatial attributes, field rules, data expression form, etc. The structure recognition can be implemented by rule matching, metadata parsing, data structure model comparison, etc., to perform structural standardization on data objects from different sources and formats, and classify them into predefined structural categories, thereby generating structural discrete labels.

[0032] In specific implementation, structure recognition can be divided into the following ways: If the data is a layer file with spatial attributes, it can be classified as geospatial data (points, lines, and surfaces); If the data is in pure tabular form (such as Excel, CSV, DBF), it is classified as tabular data; If the data is a drawing or image data (such as JPG, PDF, chart screenshots, etc.), it is classified as a drawing; If the data is in the original format (such as proprietary formats DataFileA, ShapeFile, Mapinfo, etc.), the meta-structure can be extracted according to the format rules; through the above structural attribute feature identification and classification, a corresponding structural discrete label is assigned to each piece of data.

[0033] Therefore, structural discrete labels belong to discrete identification variables; in the present invention, "stage continuous labels" are used to describe the sequential attributes of multi-source heterogeneous data in the time dimension, which are continuous identification variables; "structural discrete labels" are used to indicate the belonging categories of multi-source heterogeneous data in the structural feature dimension, which are discrete identification variables.

[0034] In this embodiment, the advanced matrix construction module is specifically used for: S3-1. Construct a matrix coordinate system in a two-dimensional plane, wherein the coordinate origin of the matrix coordinate system is represented by an origin label, and the coordinate Y axis is represented by a virtual connection axis of the stage continuous label; S3-2, performing matrix arrangement on the N double-label data triplets in the matrix coordinate system to construct a structure element matrix and a stage element matrix; S3-3, extracting an ordered sequence of N multi-source heterogeneous data from the structure element matrix; S3-4. Construct a three-element advanced matrix based on the ordered sequence of the multi-source heterogeneous data and the stage element matrix.

[0035] In this embodiment, a two-dimensional matrix coordinate system is introduced to perform spatial mapping on N double-label data triplets, thereby realizing the partitioned organization of structural attributes and time attributes in the matrix space. Specifically, the stage continuous labels form a virtual connection axis along the Y-axis direction, identifying the sequential logic of the data collection stage, and the origin label is located at the origin of the coordinates, which is used as a reference to align the data; the structural discrete labels are classified in the opposite quadrants of the matrix coordinate system with reference to the data body through structural mapping. The structural element matrix and stage element matrix thus constructed record the arrangement status of heterogeneous data in the structural and temporal dimensions, respectively. On this basis, the extracted ordered sequence of data is used as the result of structural similarity sorting, and is combined with the stage information and label information to form a three-element advanced matrix, realizing the spatial sorting modeling of multi-source heterogeneous data.

[0036] Furthermore, the N double-label data triplets are arranged in a matrix in the matrix coordinate system to construct a structure element matrix and a stage element matrix, including: S3-2-1, extracting stage continuous labels and their corresponding multi-source heterogeneous data from N double-label data triplets; S3-2-2, fill the extracted stage continuous labels and their corresponding multi-source heterogeneous data into the first quadrant of the matrix coordinate system; wherein the origin label is located on the X-axis, and the stage continuous labels and their corresponding multi-source heterogeneous data are arranged equidistantly along the positive direction of the Y-axis; S3-2-3. Define the stage continuous labels and their corresponding multi-source heterogeneous data located in the first quadrant as a structural element matrix; that is, the first column vector is the stage continuous label, the stage identification number starts from the X-axis and is arranged in the first quadrant toward the positive direction of the Y-axis; the second column vector is the corresponding multi-source heterogeneous data, and the multi-source heterogeneous data corresponding to its origin label is also located on the X-axis.

[0037] S3-2-4. Copy the elements of the second column vector in the structure element matrix to the second quadrant of the matrix coordinate system as the second column vector of the second quadrant; S3-2-5. Fill in the corresponding structural discrete labels according to the elements of the second column vector in the second quadrant as the first column vector of the second quadrant; S3-2-6. Define the structured discrete labels and their multi-source heterogeneous data located in the second quadrant as the stage element matrix.

[0038] In this embodiment, by arranging the double-label data triples in a quadrant matrix in a matrix coordinate system, a spatial distributed mapping expression of the stage continuous label and the structural discrete label is realized. Specifically, the stage continuous label and its corresponding data are arranged in the first quadrant, wherein the origin label is aligned with the starting point of the X-axis, and the remaining label data are arranged equidistantly along the positive direction of the Y-axis in sequence to form a structural element matrix; the first column of the matrix records the time stage, and the second column maps the multi-source heterogeneous data ontology to form the structural data distribution under the time axis. Furthermore, by copying and translating the data ontology in the structural element matrix to the second quadrant, and introducing the structural discrete label according to its corresponding relationship, a corresponding pairing of the structural label and the data is established, thereby forming a stage element matrix. The matrix layout logic uses the quadrant as the boundary and the label as the axis to realize the symmetrical expression of the double-label attributes in the coordinate system, which is helpful for independent modeling and cross-comparison of the structural distribution and the time stage in the same space.

[0039] Further, extracting an ordered sequence of N multi-source heterogeneous data from the structure element matrix includes: S3-3-1. The multi-source heterogeneous data corresponding to the origin labels in the structure element matrix are defined as the reference data, and the multi-source heterogeneous data corresponding to the non-origin labels are defined as the paired data; S3-3-2, associate each paired data with the reference data to obtain M heterogeneous data pairs, where M=N-1; A6-3. Calculate the content similarity of the M heterogeneous data pairs; wherein the content similarity is used to measure the similarity between two multi-source heterogeneous data in terms of information content, numerical distribution or expression results; In this embodiment, the content similarity is used to measure the similarity between two multi-source heterogeneous data in terms of information content, numerical distribution or expression characteristics. Preferably, the cosine similarity can be used to calculate the similarity after the feature vectorization of each type of data object is performed, and its expression is as follows: ; Among them, A and B represent the feature expression vectors of two multi-source heterogeneous data respectively. The similarity value ranges from [0, 1]. The higher the value, the higher the similarity of the two data in content expression. The content similarity is in scalar form and can be directly used for subsequent evaluation redundancy value calculation. If the multi-source heterogeneous data contains unstructured forms such as images, layers, and charts, they can be converted into vectors through feature extraction (such as image texture, layer vector features, table summary indicators, etc.) and then processed uniformly.

[0040] S3-3-4. Sort the M heterogeneous data pairs in descending order according to their content similarity values ​​to generate an ordered sequence of data pairs; wherein the ordered sequence of data pairs is used to represent the order in which all paired data are arranged from high to low according to the degree of similarity with the benchmark data; S3-3-5, assigning sort numbers to the ordered sequence of data pairs; Among them, the sorting number is a set of consecutive positive integers starting from 1, and is assigned to each multi-source heterogeneous data pair in order from high to low similarity; number 1 corresponds to the maximum similarity, and number M corresponds to the minimum similarity; the "sorting number" in this embodiment is used to identify the position of the data pairs arranged in descending order in the ordered sequence.

[0041] S3-3-6. Determine whether the sorting number of the heterogeneous data pair is a positive integer 1; if the sorting number is a positive integer 1, mark the corresponding heterogeneous data pair as the origin data pair; otherwise, mark it as a redundant data pair; S3-3-7. Delete all the reference data in the redundant data pairs, and define the data sequence of the M paired data and the origin data pairs after deletion as the ordered sequence of the multi-source heterogeneous data; wherein the ordered sequence of the multi-source heterogeneous data is characterized as an ordered sequence of the N multi-source heterogeneous data arranged in descending order based on content similarity.

[0042] In this embodiment, by taking the multi-source heterogeneous data corresponding to the origin label in the structural element matrix as the benchmark data, it is associated with the remaining N-1 paired data one by one to form a data pair set, and the data is compared and sorted according to the content similarity, so as to achieve cluster reconstruction of heterogeneous data at the content level. The content similarity is constructed based on feature vector expression. When processing different structural data such as maps, layers, charts, tables, etc., their texture features, vector features or summary indicators are extracted respectively and converted into standard vector representations, and then the cosine similarity is used for measurement calculation to ensure comparability between different data forms. After sorting in descending order, combined with the setting of sorting numbers and redundant marking logic, redundant benchmark data in multiple pairs of repeated evaluations are systematically eliminated, and only the most representative data of each group of content is retained, thereby generating a set of ordered sequences of multi-source heterogeneous data dominated by structural similarity.

[0043] Furthermore, according to the ordered sequence of the multi-source heterogeneous data and the stage element matrix, a three-element advanced matrix is ​​constructed, including: S3-4-1. Define the ordered sequence of N multi-source heterogeneous data as the first column vector of a three-element advanced matrix; According to the results of descending sorting of content similarity, the sorted ordered sequence of N multi-source heterogeneous data is extracted to form the first column vector of the three-element advanced matrix.

[0044] S3-4-2, extracting the stage continuous labels corresponding to N multi-source heterogeneous data from the stage element matrix; By traversing each multi-source heterogeneous data in the first column vector, the corresponding stage continuous label is found in the stage element matrix to form a one-to-one correspondence.

[0045] S3-4-3. According to the correspondence between the first column vector of the three-element advanced matrix and the stage continuous labels, the stage continuous labels are rearranged to generate stage non-continuous labels; S3-4-4, define the stage non-continuous label as the second column vector of the three-element advanced matrix; S3-4-5. Fill in the content similarity according to the correspondence between the first column vector of the three-element advanced matrix and the content similarity, and define it as the third column vector of the three-element advanced matrix; wherein the matrix position corresponding to the reference data is filled with zero; Specifically, content similarity is a representation of the degree of content similarity between the benchmark data and the paired data. In the ordered sequence of multi-source heterogeneous data (the first column vector of the three-element advanced matrix), even if the benchmark data is deleted, the retained paired data still has a corresponding relationship with the content similarity.

[0046] S3-4-6. Pair the first column vector, the second column vector and the third column vector to construct the three-element advanced matrix.

[0047] In this embodiment, based on the ordered sequence of multi-source heterogeneous data extracted in the previous stage and the stage element matrix, a three-element advanced matrix integrating sorting information, time labels and structural relevance is constructed. Specifically, the first column vector carries the main sequence of data sorted by content similarity, reflecting the distribution logic of the structural dimension; the second column vector generates non-continuous stage labels consistent with the current structural sorting by sequential mapping and rearrangement of the stage continuous labels, realizing the alignment and unification of the time stage information and the structural sequence; the third column vector is filled according to the similarity calculation results in the original data pair, and the numerical value expresses the content similarity between the two groups of data, and the corresponding numerical value of the reference data position is set to zero to highlight its reference role. Through the orderly pairing of the three column vectors, a composite data matrix with structural sorting, time stage and content similarity numerical values ​​is formed, which provides a complete three-dimensional input structure for evaluation behavior judgment, redundancy identification and label-driven warehousing based on sorting logic.

[0048] In this embodiment, the evaluation label determination module is specifically used to: S4-1, matching the first acquisition timestamp and the second acquisition timestamp of the adjacent stage continuous labels according to the stage continuous labels in the three-element advanced matrix; S4-2, defining the interval between the first acquisition timestamp and the second acquisition timestamp as an evaluation interval coefficient between adjacent multi-source heterogeneous data after normalization; S4-3, defining the product of the evaluation interval coefficient and the content similarity as the evaluation redundancy value; In this embodiment, content similarity is used to measure the closeness of two multi-source heterogeneous data at the content level, and is defined as a scalar value, usually in the interval [0,1]. The larger the value, the closer the content. This value can be multiplied by the evaluation interval coefficient to generate an evaluation redundancy value. Under the premise of a short time interval, the data content is highly repeated, which may be a resource-wasting evaluation behavior.

[0049] S4-4, comparing the evaluation redundancy value and the threshold value to generate a binary classification comparison result; S4-5. According to the binary classification comparison results, label the evaluation behavior labels of adjacent heterogeneous data pairs.

[0050] In this embodiment, based on the existing stage continuous labels and content similarity data in the three-element advanced matrix, a bivariate calculation is performed on adjacent multi-source heterogeneous data pairs. Specifically, first, by extracting the acquisition timestamps of adjacent data pairs, the time interval is calculated and normalized into the evaluation interval coefficient, which is used to quantify the compactness of the evaluation behavior in the time dimension; then combined with the content similarity index in the structural dimension, the evaluation redundancy value reflecting the "recent high weight" feature is generated by multiplying the two. As a scalar indicator, this redundancy value reflects that if the content similarity is still high when the time difference is small, the evaluation behavior may be redundant or waste resources. By comparing the redundancy value with the set threshold, a binary comparison result of "reasonable evaluation" and "unnecessary evaluation" is generated, and finally a clear evaluation behavior label is marked for each pair of adjacent data. This label can not only characterize whether the data has the meaning of retention, but also provide a judgment basis for the data selection of the storage matrix, and is one of the important screening labels in the storage stage.

[0051] Among them, according to the binary classification comparison results, the labeling conditions for labeling the evaluation behavior labels of adjacent heterogeneous data pairs are: If the binary comparison result is that the evaluation redundancy value is greater than the threshold, the data row corresponding to the second acquisition timestamp in the three-element advanced matrix is ​​marked as "unnecessary evaluation", and the data row corresponding to the first acquisition timestamp is marked as "reasonable evaluation"; otherwise, the data rows corresponding to the first acquisition timestamp and the second acquisition timestamp are both marked as "reasonable evaluation".

[0052] In this embodiment, by comparing the results of the binary classification based on the evaluation redundancy value and the preset threshold, the data rows corresponding to the adjacent data pairs in the three-element advanced matrix are accurately labeled with the evaluation behavior labels, thereby realizing the dynamic identification and marking control of the data retention value. Specifically, when the redundancy value is higher than the threshold, it means that there are highly repeated evaluation behaviors in a short time interval. The system defaults to retaining the data rows that are earlier in time, and marks the subsequent data rows as "unnecessary evaluation" to reduce the storage occupation of redundant information; on the contrary, if the redundancy value is lower than or equal to the threshold, it is considered that the two data have sufficient differentiation in content or acquisition timing, and both can be retained and uniformly labeled as "reasonable evaluation". This labeling mechanism combines the primary and secondary sorting of the time dimension with the numerical discrimination of the redundant value to realize the assistance of storage decision-making in a logically controllable and clearly identified manner, ensuring that the stored data is both structurally representative and avoids the accumulation of redundant content.

[0053] In this embodiment, a multi-source heterogeneous data storage matrix is ​​constructed based on the evaluation behavior labels and the three-element advanced matrix, including: S5-1, initializing the fourth column vector in the three-element advanced matrix; S5-2. Fill the evaluation behavior label into the fourth column vector to generate a multi-source heterogeneous data storage matrix.

[0054] In this embodiment, by initializing the fourth column vector in the three-element advanced matrix and filling the labeling result obtained by the evaluation behavior label into the column, the unified integration of structural sorting, time label, content similarity and behavior determination results is achieved, and a multi-source heterogeneous data storage matrix is ​​constructed. The storage matrix not only retains all the characteristic information of the original data in the time series, structure and similarity dimensions, but also introduces the behavior label as an explicit screening basis, providing a matrix data carrier with complete structure and clear semantics for subsequent data storage, call and visual query. Through this step, the original data completes the transformation from an unstructured, multi-source heterogeneous state to a directly callable storage standard format, and has good system adaptability, label drive and structural query capabilities, which significantly improves the organizational efficiency and intelligence level of earthquake safety evaluation data in the back-end platform.

[0055] Example 2: Reference Figure 4 and Figure 5 As shown, this embodiment 2 discloses the specific application of the client in the warehousing process described in embodiment 1 and the specific application of the query path of the data query request.

[0056] The client's data query path is as follows Figure 4 As shown, it is used to query various types of data features in the district evaluation results (i.e., the source heterogeneous data storage matrix), and each module therein is an adaptive display of the client interface.

[0057] And in Figure 5 For example, refer to Figure 5 , the query vector is used as input and data features based on the data query request can be output; the query vector can represent an area quantity query or a preparation point query, and the data feature can be the total area or the project information of the target point.

[0058] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.) means.

[0059] The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., DVD ), or semiconductor media. The semiconductor media may be a solid state drive.

[0060] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division of a waterway underwater terrain change analysis system and method. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0061] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A regional earthquake safety assessment multi-source heterogeneous data storage service system, characterized by: include: A data set acquisition module is used to acquire a data set consisting of N multi-source heterogeneous data; A triplet determination module, used to determine N double-label data triplets according to the data set; An advanced matrix construction module, used to construct a three-element advanced matrix based on N double-label data triplets; An evaluation label determination module, used to determine an evaluation behavior label from the three-element advanced matrix; wherein the evaluation behavior label is a binary label used to mark the evaluation validity of adjacent data pairs, and its value is: "reasonable evaluation" and "unnecessary evaluation"; A storage matrix construction module is used to construct a multi-source heterogeneous data storage matrix based on the evaluation behavior labels and the three-element advanced matrix; A server generation module, used for storing the multi-source heterogeneous data storage matrix in a server to generate a multi-source data server; When the client sends a data query request, the multi-source data server responds to the data query request and generates a query vector; The multi-source data server receives the query vector as input, performs feature extraction on the multi-source heterogeneous data storage matrix, and outputs data features based on the data query request.

2. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 1, characterized in that: The triplet determination module is specifically used for: S2-1, determining the stage continuous label and structure discrete label of the multi-source heterogeneous data from the data set; wherein the stage continuous label is used to characterize the collection stage of the multi-source heterogeneous data in the time dimension, and its minimum value is defined as the origin label; S2-2. Pair the multi-source heterogeneous data with their corresponding stage continuous labels and structure discrete labels to construct N double-label data triplets.

3. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 2, characterized in that: Determining phase continuous labels and structure discrete labels of multi-source heterogeneous data from the data set includes: S2-1-1, extract the marked collection timestamps of N multi-source heterogeneous data in the data set; S2-2-2. Assign stage-continuous labels to the N multi-source heterogeneous data according to the time sequence of the acquisition timestamps; wherein the stage-continuous labels are used to characterize the acquisition stage of the multi-source heterogeneous data in the time dimension, and the minimum value thereof is defined as the origin label, corresponding to the multi-source heterogeneous data with the earliest acquisition time; S2-2-3. Perform structural recognition on the data structures of N multi-source heterogeneous data and generate structural discrete labels; wherein the structural discrete labels are used to characterize the type attribution of the multi-source heterogeneous data at the structural level.

4. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 1, characterized in that: The advanced matrix building module is specifically used for: S3-1. Construct a matrix coordinate system in a two-dimensional plane, wherein the coordinate origin of the matrix coordinate system is represented by an origin label, and the coordinate Y axis is represented by a virtual connection axis of the stage continuous label; S3-2, performing matrix arrangement on the N double-label data triplets in the matrix coordinate system to construct a structure element matrix and a stage element matrix; S3-3, extracting an ordered sequence of N multi-source heterogeneous data from the structure element matrix; S3-4. Construct a three-element advanced matrix based on the ordered sequence of the multi-source heterogeneous data and the stage element matrix.

5. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 4, characterized in that: Arranging the N double-label data triplets in the matrix coordinate system to construct a structure element matrix and a stage element matrix includes: S3-2-1, extracting stage continuous labels and their corresponding multi-source heterogeneous data from N double-label data triplets; S3-2-2, fill the extracted stage continuous labels and their corresponding multi-source heterogeneous data into the first quadrant of the matrix coordinate system; wherein the origin label is located on the X-axis, and the stage continuous labels and their corresponding multi-source heterogeneous data are arranged equidistantly along the positive direction of the Y-axis; S3-2-3, defining the stage continuous labels in the first quadrant and their corresponding multi-source heterogeneous data as a structural element matrix; S3-2-4. Copy the elements of the second column vector in the structure element matrix to the second quadrant of the matrix coordinate system as the second column vector of the second quadrant; S3-2-5. Fill in the corresponding structural discrete labels according to the elements of the second column vector in the second quadrant as the first column vector of the second quadrant; S3-2-6. Define the structured discrete labels and their multi-source heterogeneous data located in the second quadrant as the stage element matrix.

6. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 5, characterized in that: Extracting an ordered sequence of N multi-source heterogeneous data from the structure element matrix includes: S3-3-1. The multi-source heterogeneous data corresponding to the origin labels in the structure element matrix are defined as the reference data, and the multi-source heterogeneous data corresponding to the non-origin labels are defined as the paired data; S3-3-2, associate each paired data with the reference data to obtain M heterogeneous data pairs, where M=N-1; A6-3. Calculate the content similarity of M heterogeneous data pairs; S3-3-4, sorting the M heterogeneous data pairs in descending order according to their content similarity values ​​to generate an ordered sequence of data pairs; S3-3-5, assigning sort numbers to the ordered sequence of data pairs; S3-3-6. Determine whether the sorting number of the heterogeneous data pair is a positive integer 1; if the sorting number is a positive integer 1, mark the corresponding heterogeneous data pair as the origin data pair; otherwise, mark it as a redundant data pair; S3-3-7. Delete all the reference data in the redundant data pairs, and define the data sequence of the M paired data and the origin data pairs after deletion as the ordered sequence of the multi-source heterogeneous data; wherein the ordered sequence of the multi-source heterogeneous data is characterized as an ordered sequence of the N multi-source heterogeneous data arranged in descending order based on content similarity.

7. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 6, characterized in that: According to the ordered sequence of the multi-source heterogeneous data and the stage element matrix, a three-element advanced matrix is ​​constructed, including: S3-4-1. Define the ordered sequence of N multi-source heterogeneous data as the first column vector of a three-element advanced matrix; S3-4-2, extracting the stage continuous labels corresponding to N multi-source heterogeneous data from the stage element matrix; S3-4-3. According to the correspondence between the first column vector of the three-element advanced matrix and the stage continuous labels, the stage continuous labels are rearranged to generate stage non-continuous labels; S3-4-4, define the stage non-continuous label as the second column vector of the three-element advanced matrix; S3-4-5. Fill in the content similarity according to the correspondence between the first column vector of the three-element advanced matrix and the content similarity, and define it as the third column vector of the three-element advanced matrix; wherein the matrix position corresponding to the reference data is filled with zero; S3-4-6. Pair the first column vector, the second column vector and the third column vector to construct the three-element advanced matrix.

8. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 1, characterized in that: The evaluation label determination module is specifically used for: S4-1, matching the first acquisition timestamp and the second acquisition timestamp of the adjacent stage continuous labels according to the stage continuous labels in the three-element advanced matrix; S4-2, defining the interval between the first acquisition timestamp and the second acquisition timestamp as an evaluation interval coefficient between adjacent multi-source heterogeneous data after normalization; S4-3, defining the product of the evaluation interval coefficient and the content similarity as the evaluation redundancy value; S4-4, comparing the evaluation redundancy value and the threshold value to generate a binary classification comparison result; S4-5. According to the binary classification comparison results, label the evaluation behavior labels of adjacent heterogeneous data pairs.

9. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 1, characterized in that: According to the binary comparison results, the labeling conditions for labeling the evaluation behavior labels of adjacent heterogeneous data pairs are: If the result of the binary comparison is that the evaluation redundancy value is greater than the threshold, the data row corresponding to the second acquisition timestamp in the three-element advanced matrix is ​​marked as "unnecessary evaluation", and the data row corresponding to the first acquisition timestamp is marked as "reasonable evaluation"; otherwise, the data rows corresponding to the first acquisition timestamp and the second acquisition timestamp are both marked as "reasonable evaluation".

10. A regional earthquake safety assessment multi-source heterogeneous data storage service system according to claim 1, characterized in that: According to the evaluation behavior labels and the three-element advanced matrix, a multi-source heterogeneous data storage matrix is ​​constructed, including: S5-1, initializing the fourth column vector in the three-element advanced matrix; S5-2. Fill the evaluation behavior label into the fourth column vector to generate a multi-source heterogeneous data storage matrix.

Images