Method for aggregating base service data and related map center server

Abstract

The application discloses a kind of base-level business data's convergence processing method and related map center server, the method includes: extracting from base-level business data to be positioned address record, and confirmed point location is marked as the trusted anchor point with anchor point credibility, obtain address processing data;Analysis to be positioned address record obtains the candidate position set containing candidate position;According to candidate position set and address processing data determine candidate position support relationship diagram;According to candidate position support relationship diagram, determine the candidate position credibility set, and according to candidate position credibility set and mutual exclusion relationship, determine the initial selection position of to-be-positioned address record, obtain initial selection position set;According to initial selection position set and candidate position support relationship diagram determine target position set, and output target position set.The application is conducive to improving the convergence accuracy of base-level business address data.

Description

Technical Field

[0001] This application relates to the fields of business data processing and geographic information processing technology, and in particular to a method for aggregating and processing basic business data and a related map center server. Background Technology

[0002] Business address data for grassroots units typically originates from self-built business systems, lightweight applications, report files, or image recognition results. The map center server needs to parse address records and aggregate the parsed locations onto the map to support spatial mapping, maintenance management, and querying.

[0003] Existing technologies typically compare the textual matching degree between each address record and map points of interest independently, selecting locations with higher matching degrees. When an address lacks a house number, uses a business abbreviation, or corresponds to multiple points of interest with the same name, the information provided by a single address record is insufficient, easily leading to the selection of incorrect locations and resulting in outliers, misplaced areas, or abnormal clustering in batch map uploads. Therefore, improving the positioning accuracy of fuzzy addresses has become a problem that needs to be solved in the aggregation and processing of grassroots business data. Summary of the Invention

[0004] This application proposes a method for aggregating and processing grassroots business data and a related map center server, aiming to determine the location corresponding to an fuzzy address by utilizing the group relationship between address records and confirmed points, thereby improving the accuracy of grassroots business data aggregation.

[0005] In a first aspect, embodiments of this application provide a method for aggregating and processing basic business data, applied to a map center server. The method includes: acquiring basic business data; extracting address records to be located from the basic business data; marking confirmed points as trusted anchors with anchor credibility to obtain address processing data; parsing the address records to be located to obtain a candidate location set containing candidate locations; establishing a support relationship between the candidate locations and the trusted anchors based on the candidate location set and the address processing data, and establishing a mutual exclusion relationship between the candidate locations corresponding to the address records to be located to obtain a candidate location support relationship graph; passing the anchor credibility along the support relationship according to the candidate location support relationship graph to obtain a candidate location credibility set; determining the initial selected location of the address records to be located based on the candidate location credibility set and the mutual exclusion relationship to obtain an initial selected location set; generating spatial anomaly detection results based on the initial selected location set; correcting the abnormal locations in the initial selected location set based on the spatial anomaly detection results and the candidate location support relationship graph to obtain a target location set; and outputting the target location set.

[0006] As can be seen, this embodiment converts confirmed locations into trusted anchor points with anchor point credibility, and transmits the anchor point credibility along the support relationship in the candidate location support relationship diagram, so that the selection of candidate locations depends not only on a single address record, but also on the group relationship constraints of confirmed locations and address data in the same batch; at the same time, by correcting the abnormal initial selection locations through spatial anomaly detection, it is possible to reduce the mismatch of locations caused by points of interest with the same name, missing addresses, and merchant abbreviations.

[0007] In one possible example, the step of establishing a support relationship between the candidate locations and the trusted anchors based on the candidate location set and the address processing data, and establishing a mutual exclusion relationship between the candidate locations corresponding to the address record to be located, to obtain a candidate location support relationship graph, includes: extracting address resolution evidence corresponding to the candidate locations from the candidate location set; extracting the trusted anchors and service grouping information from the address processing data; and combining the address resolution evidence, the trusted anchors, the service grouping information, and the candidate location set to obtain relationship construction data; establishing the support relationship between the candidate locations and between the candidate locations and the trusted anchors based on the relationship construction data to obtain an initial support relationship graph, wherein the support relationship represents administrative division consistency, address semantic consistency, spatial proximity, or service grouping consistency; for each address record to be located, connecting its corresponding different candidate locations in the initial support relationship graph to form the mutual exclusion relationship to obtain a constrained support relationship graph; and writing the address resolution evidence into the constrained support relationship graph to obtain the candidate location support relationship graph.

[0008] As can be seen, this example writes address resolution evidence and business grouping information into the graph structure at the same time, so that supporting relationships and mutually exclusive relationships can participate in the selection of candidate positions. This can both enhance the correct candidate positions by utilizing group relationships and prevent the same address record from being retained in multiple positions at the same time.

[0009] In one possible example, the step of constructing support relationships based on the relationship data, establishing support relationships between candidate locations and between candidate locations and trusted anchors, to obtain an initial support relationship graph, wherein the support relationships characterize administrative division consistency, address semantic consistency, spatial proximity, or business grouping consistency, includes: determining node pairs between candidate locations with the same administrative division and node pairs between candidate locations and trusted anchors from the relationship construction data, to obtain an administrative association node pair set; generating address semantic evidence, spatial proximity evidence, and business grouping evidence based on the administrative association node pair set, to obtain a candidate support edge set; and connecting the corresponding candidate locations or trusted anchors based on the candidate support edge set to obtain the initial support relationship graph.

[0010] As can be seen, this example first narrows down the scope of relationship construction according to administrative divisions, and then generates multiple types of supporting evidence between administratively related node pairs, which can reduce the possibility of forming incorrect supporting relationships between addresses with the same name across regions.

[0011] In one possible example, the step of generating address semantic evidence, spatial proximity evidence, and business grouping evidence based on the set of administratively associated node pairs to obtain a candidate support edge set includes: comparing the address elements of each node based on the set of administratively associated node pairs to obtain a first support evidence record set containing the address semantic evidence; generating the spatial proximity evidence based on the first support evidence record set and the coordinate positions of each node to obtain a second support evidence record set; generating the business grouping evidence based on the second support evidence record set and the business grouping information corresponding to each node to obtain a third support evidence record set; and filtering the administratively associated node pairs that meet the support conditions based on the third support evidence record set, and determining the filtered administratively associated node pairs as the candidate support edge set.

[0012] As can be seen, this example supplements the evidence in terms of address semantics, spatial proximity and business grouping in turn, so that the candidate support edges have traceable evidence sources and avoid establishing support relationships based solely on a single text similarity result.

[0013] In one possible example, the step of comparing the address elements of each node according to the administrative association node pair set to obtain a first supporting evidence record set containing the address semantic evidence includes: extracting the address elements corresponding to each node in the administrative association node pair set to obtain an address element pair set; aligning administrative division elements, road elements, house number elements, and business name elements step by step according to the address element pair set to obtain an address semantic matching record set; generating the address semantic evidence according to the address semantic matching record set, and associating the address semantic evidence with the corresponding administrative association node pair to obtain the first supporting evidence record set.

[0014] As can be seen, this example breaks down address text into address elements at different levels and generates address semantic evidence based on the hierarchical alignment results, which can distinguish node pairs that have the same administrative division but inconsistent road, house number, or business name.

[0015] In one possible example, the step of hierarchically aligning administrative division elements, road elements, house number elements, and business name elements according to the address element pair set to obtain an address semantic matching record set includes: performing hierarchical alignment of the administrative division elements in the address element pair set to obtain a first element alignment record set; performing name and hierarchical alignment of the road elements according to the first element alignment record set to obtain a second element alignment record set; performing sequential alignment of the house number elements and the business name elements according to the second element alignment record set to obtain a third element alignment record set; and aggregating consistent and conflicting elements in the third element alignment record set to obtain the address semantic matching record set.

[0016] As can be seen, this example performs alignment according to the hierarchical order of administrative division, road, house number, and business name, which can preserve consistent elements and conflicting elements, and prevent identical local names from masking address hierarchy conflicts.

[0017] In one possible example, the step of passing the anchor point confidence along the support relationship according to the candidate location support relationship graph to obtain a candidate location confidence set, and determining the initial selected location of the address record to be located according to the candidate location confidence set and the mutual exclusion relationship to obtain an initial selected location set includes: generating an anchor point confidence record set according to the trusted anchor point and the anchor point confidence; writing the anchor point confidence record set into the candidate location support relationship graph to obtain an initial node confidence set; passing the initial node confidence set from the trusted anchor point to the candidate location according to the support relationship in the candidate location support relationship graph to obtain the candidate location confidence set; sorting the candidate locations corresponding to each address record to be located according to the candidate location confidence set to obtain a candidate location sequence set; and determining the initial selected location of each address record to be located according to the candidate location sequence set and the mutual exclusion relationship to obtain the initial selected location set.

[0018] As can be seen, this example establishes a continuous processing chain of "anchor confidence - candidate position confidence - candidate position sequence - initial position", which allows the candidate position confidence to be directly used for position sorting and mutual exclusion selection, avoiding confidence becoming an isolated result that is not used.

[0019] In one possible example, the step of transferring the initial node credibility set from the trusted anchor point to the candidate location based on the support relationship in the candidate location support relationship graph to obtain the candidate location credibility set includes: determining the support path from the trusted anchor point to the candidate location based on the support relationship in the candidate location support relationship graph to obtain a credibility transfer path set; transferring the anchor point credibility along each support path based on the credibility transfer path set and the initial node credibility set, and aggregating the corresponding address semantic evidence, spatial proximity evidence, or service group evidence to obtain a path credibility record set; and for each candidate location, merging the support results pointing to that candidate location in the path credibility record set to obtain the candidate location credibility set.

[0020] As can be seen, this example records the process by which the credibility of the anchor point reaches the candidate position through different support paths, so that the credibility of the candidate position is supported by path evidence and can distinguish between the support of a single anchor point and the support of multiple anchor points.

[0021] In one possible example, the step of generating spatial anomaly detection results based on the initial selection location set, correcting the abnormal locations in the initial selection location set based on the spatial anomaly detection results and the candidate location support relationship graph, obtaining a target location set, and outputting the target location set includes: performing single-point anomaly detection and collective anomaly detection on the initial selection location set respectively, generating a location state record for each initial selection location, obtaining a location state record set, and determining the location state record set as the spatial anomaly detection result; based on the spatial anomaly detection results, obtaining alternative candidate locations with mutually exclusive relationships with the initial selection locations marked as anomalies from the candidate location support relationship graph, and determining a corrected location set based on the support relationships corresponding to the alternative candidate locations; based on the spatial anomaly detection results, retaining the normal initial selection locations from the initial selection location set, and merging the retained initial selection locations with the corrected location set to obtain the target location set.

[0022] As can be seen, this example performs single-point anomaly detection and collective anomaly detection after location selection, and uses mutual exclusion relationships to return candidate location support graphs to find alternative candidate locations, thereby forming a fallback location correction link.

[0023] Secondly, embodiments of this application provide a map center server, including: The data acquisition unit is used to acquire basic business data, extract the address records to be located from the basic business data, and mark the confirmed points as trusted anchors with anchor credibility to obtain address processing data. A candidate location generation unit is used to parse the address record to be located to obtain a candidate location set containing candidate locations; a relationship graph construction unit is used to establish a support relationship between the candidate locations and the trusted anchor points based on the candidate location set and the address processing data, and to establish a mutual exclusion relationship between the candidate locations corresponding to the address record to be located, to obtain a candidate location support relationship graph. The joint disambiguation unit is used to pass the anchor point confidence level along the support relationship according to the candidate location support relationship graph to obtain a candidate location confidence level set, and to determine the initial location of the address record to be located according to the candidate location confidence level set and the mutual exclusion relationship to obtain an initial location set; The point aggregation unit is used to generate spatial anomaly detection results based on the initial selection set of locations, correct the abnormal locations in the initial selection set of locations based on the spatial anomaly detection results and the candidate location support relationship diagram, obtain the target location set, and output the target location set.

[0024] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, a communication interface, and one or more programs, the one or more programs being stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps in the first aspect of embodiments of this application.

[0025] Fourthly, embodiments of this application provide a computer storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps described in the first aspect of this embodiment.

[0026] As can be seen, in this embodiment, compared to the existing processing method that independently compares the degree of text matching for each fuzzy address and directly selects candidate locations, this application can utilize the group spatial relationship between address records and confirmed points to perform candidate location disambiguation. The map center server of this application first obtains basic business data through a communication interface, extracts the address records to be located from the basic business data, and marks the confirmed points as trusted anchors with anchor point credibility, thus obtaining address processing data. Next, it parses the address records to be located to obtain a candidate location set. Then, it establishes support and mutual exclusion relationships based on the candidate location set and the address processing data to obtain a candidate location support relationship graph. Finally, it follows... The anchor point credibility is passed along the support relationship to obtain a candidate location credibility set. Based on the candidate location credibility set and the mutual exclusion relationship, a preliminary location set is determined. Finally, spatial anomaly detection results are generated to correct the abnormal locations in the preliminary location set, and the target location set is obtained and output. Since the candidate location credibility set is formed by the anchor point credibility along the support relationship and is directly used for the selection of mutually exclusive candidate locations, the positioning basis of the credible anchor point can be applied to candidate locations that lack complete address information. At the same time, the spatial anomaly detection results are used to backtrack and correct erroneous selections, thereby reducing mismatches of points of interest with the same name, missing addresses, and merchant abbreviations, which is conducive to improving the accuracy of the aggregation of grassroots business address data. Attached Figure Description

[0027] Figure 1 A schematic diagram of the hardware architecture for basic business data aggregation and processing provided in the embodiments of this application; Figure 2 A schematic diagram of the basic business data aggregation and processing software architecture provided in the embodiments of this application; Figure 3 This is a schematic diagram of a basic business address data aggregation scenario provided in an embodiment of this application; Figure 4 A flowchart illustrating the grassroots business data aggregation and processing method provided in this application embodiment; Figure 5 This is a schematic diagram of the candidate position support relationship diagram provided in the embodiments of this application; Figure 6 A schematic diagram illustrating the process of constructing the candidate position support relationship graph provided in the embodiments of this application; Figure 7 This is a schematic diagram illustrating the process of generating the initial support relationship diagram provided in the embodiments of this application; Figure 8 This is a schematic diagram illustrating the process of generating candidate support edges provided in an embodiment of this application; Figure 9 This is a schematic diagram illustrating the process of generating address semantic evidence provided in an embodiment of this application. Figure 10A flowchart illustrating the hierarchical alignment of address elements provided in this application embodiment; Figure 11 A schematic diagram illustrating the process of determining the initial location as provided in an embodiment of this application; Figure 12 A schematic diagram illustrating the process of anchor point credibility transfer provided in the embodiments of this application; Figure 13 A schematic diagram illustrating the process of abnormal location correction provided in an embodiment of this application; Figure 14 This is a schematic diagram of the map center server functional module provided in an embodiment of this application. Detailed Implementation

[0028] The embodiments of this application will now be described with reference to the accompanying drawings. The described embodiments are for illustrative purposes only and are not intended to limit the scope of protection of this application. Where there is no conflict, the various embodiments and their technical features can be combined with each other.

[0029] See Figure 1 The hardware architecture of the grassroots business data aggregation and processing system 100 includes a grassroots business terminal 110, a district / county business system 120, a map center server 130, an address resolution service device 140, a spatial database 150, an intranet map service device 160, a management terminal 170, and a communication network 180. The map center server 130 includes a processor 131, a memory 132, a communication interface 133, and a map interface 134. The grassroots business terminal 110 uploads business data through the district / county business system 120; the processor 131 calls the address resolution service device 140 to process address records, stores the processing results in the spatial database 150, and provides the target location to the intranet map service device 160 or the management terminal 170 through the map interface 134.

[0030] See Figure 2 The grassroots business data aggregation and processing system 100 includes a data access layer 201, an address processing layer 202, a relationship reasoning layer 203, and a map application layer 204. A data acquisition unit 210 is located in the data access layer 201; a candidate location generation unit 220 is located in the address processing layer 202; a relationship graph construction unit 230 and a joint disambiguation unit 240 are located in the relationship reasoning layer 203; and a point aggregation unit 250 is located in the map application layer 204. Each layer sequentially transmits processing results through address processing data, candidate location sets, candidate location supporting relationship graphs, preliminary location sets, and target location sets.

[0031] See Figure 3The district / county business system 120 sends a batch of basic business data 310 and a set of confirmed points 320 to the map center server 130. The map center server 130 parses the fuzzy addresses in the batch of basic business data 310 to obtain a set of candidate locations 330, combines it with the set of confirmed points 320 to form a candidate location support relationship graph 340, and further obtains a set of target locations 350 and a map display result 360. For example, when "a grocery store next to a certain avenue" corresponds to multiple map points of interest, the map center server 130 uses the group relationship between addresses in the same batch and confirmed points to determine the target location.

[0032] In another application scenario, different districts and counties upload address data recorded using road abbreviations, community names, or business names. The map center server can generate corresponding data batches for each batch and perform candidate location disambiguation within the administrative division of each data batch to prevent businesses with the same name in different districts and counties from interfering with each other.

[0033] See Figure 4 The grassroots business data aggregation and processing method can be implemented through the following steps S101 to S105, which can be executed by the map center server.

[0034] S101. Obtain basic business data, extract the address records to be located from the basic business data, and mark the confirmed points as trusted anchors with anchor credibility to obtain address processing data.

[0035] The basic business data can include address text, address image recognition text, business object identifiers, business grouping fields, and confirmed locations. Address records to be located are location records for which reliable coordinates have not yet been obtained. Trusted anchor points are confirmed locations that have been manually verified, verified through business operations, or have been used stably multiple times. Anchor point trustworthiness indicates the reliability of a trusted anchor point as a location reference and can be determined based on the source of confirmation, confirmation time, and consistency of historical usage.

[0036] S102. Parse the address record to be located to obtain a set of candidate locations containing candidate locations.

[0037] In step S102, character normalization, administrative division completion, and address element extraction can be performed on the address record to be located, and then the address resolution service can be called to obtain one or more candidate locations. The candidate location set is grouped according to the address record to be located, and the candidate coordinates, point of interest names, administrative divisions, road information, and resolution source can be saved for subsequent relationship construction.

[0038] S103. Based on the candidate location set and the address processing data, establish a support relationship between the candidate location and the trusted anchor point, and establish a mutual exclusion relationship between the candidate locations corresponding to the address record to be located, to obtain a candidate location support relationship diagram.

[0039] A candidate location support graph is a graph structure containing nodes and relationship edges. Support relationships indicate that two nodes are consistent in terms of administrative division, address semantics, spatial proximity, or business grouping. Mutual exclusion relationships indicate that different candidate locations for the same address record cannot simultaneously be the final location.

[0040] In one possible example, see Figure 5 The candidate location support relationship graph includes a candidate location node set 510, a trusted anchor node set 520, a support edge set 530, and a mutual exclusion edge set 540. The support edge set 530 connects related nodes within the candidate location node set 510, or connects the candidate location node set 510 with the trusted anchor node set 520; the mutual exclusion edge set 540 connects candidate location nodes corresponding to the same address record to be located.

[0041] S104. Based on the candidate location support relationship diagram, the anchor point confidence level is transmitted along the support relationship to obtain a candidate location confidence level set. Based on the candidate location confidence level set and the mutual exclusion relationship, the initial selection location of the address record to be located is determined to obtain an initial selection location set.

[0042] In step S104, the map center server can start from a trusted anchor node and pass the anchor confidence level along the relation edges with supporting evidence to the candidate location nodes. The candidate location confidence level is used to indicate the degree to which the corresponding candidate location is supported by trusted anchors and other candidate locations. The server sorts the candidate locations of the same address record to be located and retains an initial selection location under mutual exclusion constraints.

[0043] S105. Generate spatial anomaly detection results based on the initial selection location set, correct the abnormal locations in the initial selection location set based on the spatial anomaly detection results and the candidate location support relationship diagram, obtain the target location set, and output the target location set.

[0044] In step S105, single-point anomaly detection is used to identify initial locations that deviate from the corresponding administrative region or surrounding normal locations; collective anomaly detection is used to identify anomalies such as concentration, dispersion, or cross-regional distribution formed by a group of initial locations. For anomaly locations, the server returns a candidate location support relationship graph, searches for alternative candidate locations with mutually exclusive relationships, and completes the correction based on their support relationships.

[0045] As can be seen, in this example, the basic business data, address processing data, candidate location set, candidate location support relationship diagram, candidate location credibility set, preliminary location set and target location set are sequentially connected, which can avoid intermediate results becoming isolated features and improve the batch aggregation accuracy of fuzzy addresses.

[0046] See Figure 6The process of constructing the candidate position support relationship graph in step S103 can be achieved through the following steps S201 to S204.

[0047] S201. Extract the address resolution evidence corresponding to the candidate location from the candidate location set, extract the trusted anchor and service group information from the address processing data, and combine the address resolution evidence, the trusted anchor, the service group information and the candidate location set to obtain relationship construction data.

[0048] Address resolution evidence may include the standard address, point of interest name, administrative division, road, house number, and coordinates returned by the address resolution service. Business grouping information indicates the district, street, business area, or other business set to which the address record belongs. Relationship building data establishes a correspondence between the above information and the address record to be located and the candidate locations.

[0049] In one possible implementation, when some address records lack a business grouping field, temporary business grouping information can be generated based on their administrative division, the business groups of adjacent trusted anchors, and common fields of address records in the same batch. This implementation can solve the problem of not being able to form grouping evidence when the business grouping field is missing, allowing address records with missing fields to still participate in subsequent relationship construction.

[0050] S202. Based on the relationship, construct data, establish the support relationship between the candidate locations and between the candidate locations and the trusted anchor point, and obtain an initial support relationship graph. The support relationship represents the consistency of administrative division, address semantics, spatial proximity, or business grouping.

[0051] In step S202, a support relationship can correspond to one or more types of consistency. Different consistency evidences are stored separately to distinguish the source of evidence when determining support edges later. The support relationship between candidate locations is used to express the group consistency of addresses in the same batch; the support relationship between candidate locations and trusted anchor points is used to introduce the location basis of confirmed points.

[0052] S203. For each of the address records to be located, connect the different candidate positions corresponding to it in the initial support relationship graph to form the mutual exclusion relationship, and obtain the constraint support relationship graph.

[0053] In step S203, the mutual exclusion relationship does not mean that the two candidate locations are spatially mutually exclusive, but rather that they correspond to the same address record to be located, and ultimately only one candidate location can be selected. The constraint support relationship graph retains both support relationships and mutual exclusion relationships to support subsequent confidence ranking and unique location selection.

[0054] S204. Write the address resolution evidence into the constraint support relationship diagram to obtain the candidate location support relationship diagram.

[0055] In step S204, the address resolution evidence can be written into the node attributes of the candidate location node, or it can be associated with the corresponding support relationship. Therefore, subsequent location credibility transmission and abnormal location rollback can both be traced back to the address resolution basis.

[0056] As can be seen in this example, by constructing a unified data organization of candidate locations, trusted anchors, business grouping information, and address resolution evidence through relationships, a candidate location support relationship diagram with both group support capability and candidate mutual exclusion capability can be generated.

[0057] See Figure 7 The initial support relationship diagram generation process in step S202 can be achieved through the following steps S301 to S303.

[0058] S301. Determine the node pairs between the candidate locations with the same administrative division and the node pairs between the candidate locations and the trusted anchor points from the relationship construction data to obtain a set of administrative association node pairs.

[0059] Among these, "same administrative division" can refer to the same district / county level, the same street level, or the same administrative level that meets the requirements of the current map task. The set of administratively related node pairs is used to limit the scope of subsequent evidence generation and reduce invalid node combinations across administrative regions.

[0060] S302. Generate address semantic evidence, spatial proximity evidence, and business grouping evidence based on the administrative association node pair set to obtain a candidate support edge set.

[0061] Address semantic evidence is used to indicate the degree of consistency of address elements; spatial proximity evidence is used to indicate whether node coordinates are within a mutually referential spatial range; business grouping evidence is used to indicate whether nodes belong to the same business set. The candidate support edge set stores node pairs that satisfy the support conditions and their evidence types.

[0062] In one possible implementation, administrative division consistency can be set as a pre-screening condition, while address semantics, spatial proximity, and business grouping evidence can be set as subsequent supporting conditions. This implementation can solve the segmentation problem of merchants with the same name across regions easily forming erroneous associations, and reduce cross-regional false connections while retaining weak text matching candidates within the same region.

[0063] S303. Connect the candidate positions or the trusted anchor points corresponding to the candidate support edge set to obtain the initial support relationship graph.

[0064] In step S303, each candidate support edge records its connecting nodes and supporting evidence. A candidate location can be connected to multiple candidate locations or trusted anchor points simultaneously, with multiple independent support sources jointly providing the positioning basis.

[0065] This embodiment first forms a set of administratively related node pairs, and then generates a set of candidate support edges, which can control the scale of the graph structure and improve the regional consistency of support relationships.

[0066] See Figure 8 The candidate support edge generation process in step S302 can be implemented through the following steps S401 to S404.

[0067] S401. Based on the administrative association node pair set, compare the address elements of each node to obtain a first supporting evidence record set containing the address semantic evidence.

[0068] The first supporting evidence record set can record consistent address elements, conflicting address elements, missing address elements, and corresponding node pairs. Address semantic evidence does not require the address text to be completely identical, as long as there are no incompatible conflicts at its key levels.

[0069] S402. Generate the spatial proximity evidence based on the first supporting evidence record set and the coordinate positions of each node to obtain the second supporting evidence record set.

[0070] In step S402, spatial proximity can be determined based on whether the nodes are located near the same road, within the same community, or in the same business service area. The second supporting evidence record set supplements spatial proximity evidence while retaining the first supporting evidence record set.

[0071] S403. Generate the business group evidence based on the second supporting evidence record set and the business group information corresponding to each node, and obtain the third supporting evidence record set.

[0072] In step S403, the business grouping information can originate from grassroots business data or be inferred from administrative divisions and adjacent reliable anchor points. The third supporting evidence record set continues to retain the aforementioned address semantic evidence and spatial proximity evidence to form a continuous chain of evidence.

[0073] In one possible implementation, when address semantic evidence conflicts with traffic packet evidence, the update time and source of the traffic packet information can be examined. For traffic packet information that has not been updated for a long time, its support priority can be reduced. This implementation can resolve the subdivision problem of correct candidate locations being excluded due to traffic packet lag, which is beneficial for maintaining the consistency of evidence usage timeliness.

[0074] S404. Based on the third set of supporting evidence records, filter the administrative association node pairs that meet the supporting conditions, and determine the filtered administrative association node pairs as the candidate supporting edge set.

[0075] In step S404, the supporting conditions may require at least one valid piece of supporting evidence and the absence of key-level conflicts, or the evidence combination may be adjusted according to the business scenario. The screening results retain the source of the evidence for use in credibility transfer and anomaly correction.

[0076] As can be seen, in this example, supporting evidence is gradually supplemented in the order of address semantics, spatial proximity, and business grouping, making the formation process of candidate supporting edges clear, interpretable, and traceable.

[0077] See Figure 9 The address semantic evidence generation process can be achieved through the following steps S501 to S503.

[0078] S501. Extract the address elements corresponding to each node in the administrative association node pair set to obtain the address element pair set.

[0079] The address elements can include administrative division elements 910, road elements 920, house number elements 930, and business name elements 940. The address element pair set is organized according to administrative association node pairs, so that similar address elements of two nodes form a pair of elements to be compared.

[0080] S502. Align the administrative division elements, road elements, house number elements, and business name elements of the address element set step by step to obtain the address semantic matching record set.

[0081] In step S502, step-by-step alignment means first determining the upper-level address and then the lower-level address. When an incompatible conflict occurs at the upper level, subsequent alignment of the corresponding node pair can be stopped; when the upper level is consistent but the lower level is missing, the missing state can be retained instead of being directly determined as a mismatch. The address semantic matching record set 950 records the consistency, missing, and conflict situations of each level.

[0082] In one possible implementation, synonymous associations can be established between road abbreviations, historical road names, and business aliases. When two address elements differ in wording but correspond to the same standard entity, they are recorded as semantically consistent. This implementation can solve the subdivision problem where abbreviations and aliases cause direct character comparisons to fail, and can improve the adaptability of address semantic evidence to non-standard addresses.

[0083] S503. Generate the address semantic evidence based on the address semantic matching record set, and associate the address semantic evidence with the corresponding administrative association node pair to obtain the first supporting evidence record set.

[0084] In step S503, the address semantic evidence may include a consistency level, a conflict level, and a missing level, enabling subsequent processing to identify the specific source of the evidence rather than simply saving a single matching conclusion.

[0085] As can be seen in this example, by decomposing address elements and aligning them step by step, the semantic evidence of addresses can express the structural relationship between address levels, reducing false matches caused by identical local characters.

[0086] See Figure 10 The address element alignment process in step S502 can be achieved through the following steps S601 to S604.

[0087] S601. Perform hierarchical alignment of the administrative division elements in the address element pair set to obtain a first element alignment record set.

[0088] The administrative division elements can be organized according to the provincial, municipal, district / county, and street levels. The first element alignment record set records whether the corresponding levels are consistent, missing, or conflicting, and serves as the input for road element alignment.

[0089] S602. Align the road elements by name and level according to the first element alignment record set to obtain the second element alignment record set.

[0090] In step S602, road name alignment can identify full names, abbreviations, and historical names, while road hierarchy alignment can distinguish between main roads, branch roads, alleyways, or other road structures. The second element alignment record set inherits the administrative division alignment results and supplements the road alignment results.

[0091] S603. Align the doorplate element and the merchant name element sequentially according to the second element alignment record set to obtain the third element alignment record set.

[0092] In step S603, the address number, building number, and floor information can be aligned first, and then the business name can be aligned. When the address element is missing, the alignment result of the business name can be retained, but the partially identical business name cannot replace the already determined administrative division or road conflict.

[0093] S604. Collect the consistent elements and conflicting elements in the third element alignment record set to obtain the address semantic matching record set.

[0094] In step S604, consistent elements are used to form address semantic support, conflicting elements are used to limit erroneous support relationships, and missing elements are used to indicate that the current address evidence is incomplete. All of these results are written into the address semantic matching record set.

[0095] As can be seen, in this example, the previous alignment result is passed in a hierarchical manner, so that the processing of administrative divisions, roads, house numbers and business names is related before and after, which can preserve address structure information and improve the interpretability of semantic matching.

[0096] See Figure 11 The initial location determination process in step S104 can be achieved through the following steps S701 to S704.

[0097] S701. Generate an anchor point credibility record set based on the trusted anchor point and the anchor point credibility, and write the anchor point credibility record set into the candidate position support relationship graph to obtain an initial node credibility set.

[0098] The anchor point credibility record set records the credible anchor point identifier, anchor point credibility, and confirmation source. The initial node credibility set associates these records with the credible anchor point nodes in the candidate location support relationship graph, providing initial data for subsequent credibility transmission.

[0099] S702. Based on the support relationship in the candidate position support relationship graph, the initial node confidence set is transferred from the confidence anchor point to the candidate position to obtain the candidate position confidence set.

[0100] In step S702, the credibility transfer is constrained by the supporting relationships and their evidence. A candidate position can receive support results from one or more credible anchor points, or it can obtain indirect support through other candidate positions. The candidate position credibility set stores each candidate position and its supporting source.

[0101] S703. Sort the candidate locations corresponding to each address record to be located according to the candidate location confidence set to obtain a candidate location sequence set.

[0102] In step S703, the candidate location sequence is formed according to the confidence level of the candidate locations. For candidate locations with the same or similar confidence levels, the order can be determined based on the completeness of the address resolution evidence or the number of valid supporting sources.

[0103] S704. Determine the initial selected position for each of the address records to be located based on the candidate position sequence set and the mutual exclusion relationship, thereby obtaining the initial selected position set.

[0104] In step S704, one candidate position is reserved as an initial selection position for each candidate position sequence, and other candidate positions for the same address record to be located are excluded based on mutual exclusion relationships. The initial selection position is written into the initial selection position set for use in spatial anomaly detection.

[0105] As can be seen, in this example, by continuously using the anchor point credibility record set, the initial node credibility set, the candidate position credibility set, and the candidate position sequence set, the initial position has a clear source of credibility and a unique selection basis.

[0106] See Figure 12 The anchor point credibility transfer process can be implemented through the following steps S801 to S803.

[0107] S801. Determine the support path from the trusted anchor point to the candidate position based on the support relationship in the candidate position support relationship diagram to obtain a set of trustworthiness transmission paths.

[0108] The credibility transmission path 1220 consists of a credible anchor point and at least one candidate position. The server can limit the number of nodes the path passes through and can also exclude supporting edges containing key evidence conflicts, thereby preventing credibility from spreading continuously along irrelevant nodes.

[0109] S802. Based on the credibility transmission path set and the initial node credibility set, transmit the anchor point credibility along each of the supporting paths and aggregate the corresponding address semantic evidence, spatial proximity evidence or service group evidence to obtain a path credibility record set.

[0110] In step S802, the anchor point credibility record 1210 serves as the path starting point data. The path credibility record 1230 stores the corresponding credible anchor point, the supporting edges traversed, the evidence type of each supporting edge, and the candidate position it points to, enabling the credibility transmission result to be traced back.

[0111] S803. For each candidate location, merge the support results pointing to that candidate location in the path credibility record set to obtain the candidate location credibility set.

[0112] In step S803, the server merges the path credibility records 1230 pointing to the same candidate location to form a candidate location credibility record 1240. The independent support formed by different credibility anchors can jointly improve the credibility of the corresponding candidate location; paths with conflicting evidence can be marked or excluded.

[0113] As can be seen in this example, by recording the complete credibility transmission path, the credibility of the candidate position is not only a ranking result, but also has an interpretable anchor source and supporting evidence.

[0114] See Figure 13 The abnormal position correction process in step S105 can be achieved through the following steps S901 to S903.

[0115] S901. Perform single-point anomaly detection and collective anomaly detection on the initial selection location set respectively, generate a location status record for each initial selection location, obtain a location status record set, and determine the location status record set as the spatial anomaly detection result.

[0116] Among them, single-point anomaly detection can determine whether the initially selected location exceeds the corresponding administrative division, whether it deviates significantly from the trusted anchor points in the same group, or whether it falls into an area unsuitable for the business object. Collective anomaly detection can determine whether the initially selected locations in the same batch form an unreasonable concentrated, dispersed, or cross-regional distribution. Location status records are used to characterize whether the initially selected location is normal, abnormal, or pending confirmation.

[0117] S902. Based on the spatial anomaly detection result, obtain alternative candidate positions that have the mutually exclusive relationship with the initially selected positions marked as anomalies from the candidate position support relationship diagram, and determine the corrected position set based on the support relationship corresponding to the alternative candidate positions.

[0118] In step S902, the alternative candidate positions are derived from other candidate positions corresponding to the same address record to be located. The server excludes candidate positions that have been rejected by spatial anomaly detection and re-examines the supporting relationships of the remaining alternative candidate positions to determine the corrected position.

[0119] S903. Based on the spatial anomaly detection results, retain the normal initial selection positions from the initial selection position set, and merge the retained initial selection positions with the corrected position set to obtain the target position set.

[0120] In step S903, normal initial selection positions are not repeatedly selected as candidates; abnormal initial selection positions are replaced by corrected positions. If no corrected position meets the requirements, the corresponding address record can be marked as pending confirmation, instead of directly writing low-confidence candidate positions into the target location set.

[0121] As can be seen, in this example, the spatial anomaly detection results are fed back to the candidate location selection process by using mutual exclusion relationships, forming a closed loop of "initial selection-detection-back-correction", which can reduce the possibility of erroneous points entering the final convergence result.

[0122] See Figure 14 The map center server 130 includes a data acquisition unit 210, a candidate location generation unit 220, a relationship graph construction unit 230, a joint disambiguation unit 240, and a point aggregation unit 250.

[0123] The data acquisition unit 210 is used to acquire basic business data, extract the address records to be located from the basic business data, and mark the confirmed points as trusted anchors with anchor credibility to obtain address processing data.

[0124] The candidate location generation unit 220 is used to parse the address record to be located, obtain a candidate location set containing candidate locations, and send the candidate location set to the relationship graph construction unit 230.

[0125] The relationship graph construction unit 230 is used to establish a support relationship between the candidate positions and the trusted anchor points based on the candidate position set and the address processing data, and to establish a mutual exclusion relationship between the candidate positions corresponding to the address record to be located, so as to obtain a candidate position support relationship graph.

[0126] The joint disambiguation unit 240 is used to pass the anchor point confidence along the support relationship according to the candidate position support relationship graph to obtain a candidate position confidence set, and determine the initial position of the address record to be located according to the candidate position confidence set and the mutual exclusion relationship to obtain an initial position set.

[0127] The point aggregation unit 250 is used to generate spatial anomaly detection results based on the initial selection set of locations, correct the abnormal locations in the initial selection set of locations based on the spatial anomaly detection results and the candidate location support relationship diagram, obtain the target location set, and output the target location set.

[0128] Therefore, the data acquisition unit 210, the candidate location generation unit 220, the relationship graph construction unit 230, the joint disambiguation unit 240, and the point aggregation unit 250 successively reuse the output results of the previous unit, enabling the map center server 130 to complete the entire processing process from basic business data acquisition, fuzzy address candidate generation, group relationship reasoning to abnormal location correction.

[0129] In one possible example, regarding the step of establishing a support relationship between the candidate locations and the trusted anchor points based on the candidate location set and the address processing data, and establishing a mutual exclusion relationship between the candidate locations corresponding to the address record to be located, to obtain a candidate location support relationship graph, the relationship graph construction unit 230 is specifically used for: The address resolution evidence corresponding to the candidate location is extracted from the candidate location set, the trusted anchor and service group information are extracted from the address processing data, and the address resolution evidence, the trusted anchor, the service group information and the candidate location set are combined to obtain relationship construction data; Based on the relationship data, the support relationship is established between the candidate locations and between the candidate locations and the trusted anchor points to obtain an initial support relationship graph. The support relationship represents the consistency of administrative divisions, address semantics, spatial proximity, or business grouping. For each of the address records to be located, connect the different candidate positions corresponding to it in the initial support relationship graph to form the mutual exclusion relationship, and obtain the constraint support relationship graph. The address resolution evidence is written into the constraint support graph to obtain the candidate location support graph.

[0130] In one possible example, as the data is constructed based on the relationships, the supporting relationships are established between the candidate locations and between the candidate locations and the trusted anchors to obtain an initial supporting relationship graph. The supporting relationships characterize aspects such as administrative division consistency, address semantic consistency, spatial proximity, or business grouping consistency. The relationship graph construction unit 230 is specifically used for: From the relationship construction data, node pairs between candidate locations with the same administrative division and node pairs between the candidate locations and the trusted anchor points are determined to obtain a set of administrative association node pairs. Based on the administrative association node pair set, address semantic evidence, spatial proximity evidence, and business grouping evidence are generated to obtain a candidate support edge set; The initial support relationship graph is obtained by connecting the candidate positions or the trusted anchor points corresponding to the candidate support edge set.

[0131] In one possible example, regarding the generation of address semantic evidence, spatial proximity evidence, and business grouping evidence based on the administrative association node pair set to obtain a candidate support edge set, the relational graph construction unit 230 is specifically used for: By comparing the address elements of each node in the administrative association node pair set, a first supporting evidence record set containing the address semantic evidence is obtained; The spatial proximity evidence is generated based on the first supporting evidence record set and the coordinate positions of each node, thus obtaining the second supporting evidence record set; The third supporting evidence record set is obtained by generating the business group evidence based on the second supporting evidence record set and the business group information corresponding to each node. The administrative association node pairs that meet the support conditions are selected from the third supporting evidence record set, and the selected administrative association node pairs are determined as the candidate supporting edge set.

[0132] In one possible example, in the process of comparing the address elements of each node according to the administrative association node pair set to obtain a first supporting evidence record set containing the address semantic evidence, the relationship graph construction unit 230 is specifically used for: Extract the address elements corresponding to each node in the administrative association node pair set to obtain the address element pair set; Based on the address element pair set, the administrative division element, road element, house number element and business name element are aligned level by level to obtain the address semantic matching record set; The address semantic evidence is generated based on the address semantic matching record set, and the address semantic evidence is associated with the corresponding administrative association node pair to obtain the first supporting evidence record set.

[0133] In one possible example, regarding the stepwise alignment of administrative division elements, road elements, house number elements, and business name elements according to the address element set to obtain an address semantic matching record set, the relationship graph construction unit 230 is specifically used for: The administrative division elements in the address element pair set are hierarchically aligned to obtain a first element alignment record set; The road elements are aligned by name and level according to the first element alignment record set to obtain the second element alignment record set. The doorplate element and the merchant name element are sequentially aligned according to the second element alignment record set to obtain the third element alignment record set. By combining the consistent and conflicting elements in the third element alignment record set, the address semantic matching record set is obtained.

[0134] In one possible example, in the process of passing the anchor point confidence along the support relationship according to the candidate location support relationship graph to obtain a candidate location confidence set, and determining the initial location of the address record to be located according to the candidate location confidence set and the mutual exclusion relationship to obtain an initial location set, the joint disambiguation unit 240 is specifically used for: An anchor point credibility record set is generated based on the trusted anchor point and the anchor point credibility, and the anchor point credibility record set is written into the candidate position support relationship graph to obtain the initial node credibility set. Based on the support relationship in the candidate location support relationship graph, the initial node confidence set is transferred from the confidence anchor to the candidate location to obtain the candidate location confidence set; Based on the candidate location confidence set, the candidate locations corresponding to each address record to be located are sorted to obtain a candidate location sequence set; The initial selected position is determined for each of the address records to be located based on the candidate position sequence set and the mutual exclusion relationship, thus obtaining the initial selected position set.

[0135] In one possible example, regarding the transfer of the initial node confidence set from the confidence anchor to the candidate location based on the support relationships in the candidate location support relationship graph to obtain the candidate location confidence set, the joint disambiguation unit 240 is specifically used for: Based on the support relationships in the candidate location support relationship diagram, the support path from the trusted anchor point to the candidate location is determined, and a set of trustworthiness transmission paths is obtained; Based on the credibility transmission path set and the initial node credibility set, the anchor point credibility is transmitted along each of the supporting paths, and the corresponding address semantic evidence, spatial proximity evidence or service group evidence is collected to obtain a path credibility record set. For each candidate location, the support results pointing to that candidate location are merged from the path confidence record set to obtain the candidate location confidence set.

[0136] In one possible example, in the process of generating spatial anomaly detection results based on the initial location set, correcting the anomaly locations in the initial location set based on the spatial anomaly detection results and the candidate location support relationship diagram, obtaining the target location set, and outputting the target location set, the point aggregation unit 250 is specifically used for... Single-point anomaly detection and collective anomaly detection are performed on the initial selection set of locations respectively, and a location status record is generated for each initial selection location to obtain a location status record set. The location status record set is then determined as the spatial anomaly detection result. Based on the spatial anomaly detection results, alternative candidate positions that have the mutually exclusive relationship with the initially selected positions marked as anomalies are obtained from the candidate position support relationship graph, and a set of corrected positions is determined based on the support relationship corresponding to the alternative candidate positions. Based on the spatial anomaly detection results, normal initial selection positions are retained from the initial selection position set, and the retained initial selection positions are merged with the corrected position set to obtain the target position set.

[0137] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0138] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0139] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0140] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0141] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0142] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0143] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0144] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0145] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for aggregating and processing grassroots business data, characterized in that, Applied to a map center server, the method includes: Acquire basic business data, extract the address records to be located from the basic business data, and mark the confirmed points as trusted anchors with anchor credibility to obtain address processing data; The address record to be located is parsed to obtain a set of candidate locations containing candidate locations; Based on the candidate location set and the address processing data, a support relationship is established between the candidate location and the trusted anchor point, and a mutual exclusion relationship is established between the candidate locations corresponding to the address record to be located, thus obtaining a candidate location support relationship diagram. According to the candidate location support relationship diagram, the anchor point confidence is passed along the support relationship to obtain the candidate location confidence set, and the initial location of the address record to be located is determined according to the candidate location confidence set and the mutual exclusion relationship to obtain the initial location set; Based on the initial selection set of locations, spatial anomaly detection results are generated. Based on the spatial anomaly detection results and the candidate location support relationship diagram, the anomaly locations in the initial selection set are corrected to obtain the target location set, and the target location set is output.

2. The method as described in claim 1, characterized in that, The step of establishing a support relationship between the candidate locations and the trusted anchor points based on the candidate location set and the address processing data, and establishing a mutual exclusion relationship between the candidate locations corresponding to the address record to be located, to obtain a candidate location support relationship graph, includes: The address resolution evidence corresponding to the candidate location is extracted from the candidate location set, the trusted anchor and service group information are extracted from the address processing data, and the address resolution evidence, the trusted anchor, the service group information and the candidate location set are combined to obtain relationship construction data; Based on the relationship data, the support relationship is established between the candidate locations and between the candidate locations and the trusted anchor points to obtain an initial support relationship graph. The support relationship represents the consistency of administrative divisions, address semantics, spatial proximity, or business grouping. For each of the address records to be located, connect the different candidate positions corresponding to it in the initial support relationship graph to form the mutual exclusion relationship, and obtain the constraint support relationship graph. The address resolution evidence is written into the constraint support graph to obtain the candidate location support graph.

3. The method as described in claim 2, characterized in that, The step involves constructing data based on the relationships, establishing support relationships between candidate locations and between candidate locations and trusted anchor points, to obtain an initial support relationship graph. These support relationships characterize consistency in administrative divisions, semantic addressing, spatial proximity, or business grouping, including: From the relationship construction data, node pairs between candidate locations with the same administrative division and node pairs between the candidate locations and the trusted anchor points are determined to obtain a set of administrative association node pairs. Based on the administrative association node pair set, address semantic evidence, spatial proximity evidence, and business grouping evidence are generated to obtain a candidate support edge set; The initial support relationship graph is obtained by connecting the candidate positions or the trusted anchor points corresponding to the candidate support edge set.

4. The method as described in claim 3, characterized in that, The step of generating address semantic evidence, spatial proximity evidence, and business grouping evidence based on the administrative association node set to obtain a candidate support edge set includes: By comparing the address elements of each node in the administrative association node pair set, a first supporting evidence record set containing the address semantic evidence is obtained; The spatial proximity evidence is generated based on the first supporting evidence record set and the coordinate positions of each node, thus obtaining the second supporting evidence record set; The third supporting evidence record set is obtained by generating the business group evidence based on the second supporting evidence record set and the business group information corresponding to each node. The administrative association node pairs that meet the support conditions are selected from the third supporting evidence record set, and the selected administrative association node pairs are determined as the candidate supporting edge set.

5. The method as described in claim 4, characterized in that, The step of comparing the address elements of each node according to the administrative association node pair set to obtain a first supporting evidence record set containing the address semantic evidence includes: Extract the address elements corresponding to each node in the administrative association node pair set to obtain the address element pair set; Based on the address element pair set, the administrative division element, road element, house number element and business name element are aligned level by level to obtain the address semantic matching record set; The address semantic evidence is generated based on the address semantic matching record set, and the address semantic evidence is associated with the corresponding administrative association node pair to obtain the first supporting evidence record set.

6. The method as described in claim 5, characterized in that, The process involves aligning administrative division elements, road elements, house number elements, and business name elements hierarchically according to the address element set to obtain an address semantic matching record set, including: The administrative division elements in the address element pair set are hierarchically aligned to obtain a first element alignment record set; The road elements are aligned by name and level according to the first element alignment record set to obtain the second element alignment record set. The doorplate element and the merchant name element are sequentially aligned according to the second element alignment record set to obtain the third element alignment record set. By combining the consistent and conflicting elements in the third element alignment record set, the address semantic matching record set is obtained.

7. The method as described in claim 1, characterized in that, The process involves passing the anchor point confidence level along the support relationship diagram based on the candidate location to obtain a candidate location confidence level set, and determining the initial selected location of the address record to be located based on the candidate location confidence level set and the mutual exclusion relationship, thus obtaining an initial selected location set, including: An anchor point credibility record set is generated based on the trusted anchor point and the anchor point credibility, and the anchor point credibility record set is written into the candidate position support relationship graph to obtain the initial node credibility set. Based on the support relationship in the candidate location support relationship graph, the initial node confidence set is transferred from the confidence anchor to the candidate location to obtain the candidate location confidence set; Based on the candidate location confidence set, the candidate locations corresponding to each address record to be located are sorted to obtain a candidate location sequence set; The initial selected position is determined for each of the address records to be located based on the candidate position sequence set and the mutual exclusion relationship, thus obtaining the initial selected position set.

8. The method as described in claim 7, characterized in that, The step of transferring the initial node confidence set from the confidence anchor to the candidate position based on the support relationship in the candidate position support relationship graph to obtain the candidate position confidence set includes: Based on the support relationships in the candidate location support relationship diagram, the support path from the trusted anchor point to the candidate location is determined, and a set of trustworthiness transmission paths is obtained; Based on the credibility transmission path set and the initial node credibility set, the anchor point credibility is transmitted along each of the supporting paths, and the corresponding address semantic evidence, spatial proximity evidence or service group evidence is collected to obtain a path credibility record set. For each candidate location, the support results pointing to that candidate location are merged from the path confidence record set to obtain the candidate location confidence set.

9. The method as described in claim 1, characterized in that, The process of generating spatial anomaly detection results based on the initial selection set of locations, correcting the anomaly locations in the initial selection set based on the spatial anomaly detection results and the candidate location support graph, obtaining the target location set, and outputting the target location set includes: Single-point anomaly detection and collective anomaly detection are performed on the initial selection set of locations respectively, and a location status record is generated for each initial selection location to obtain a location status record set. The location status record set is then determined as the spatial anomaly detection result. Based on the spatial anomaly detection results, alternative candidate positions that have the mutually exclusive relationship with the initially selected positions marked as anomalies are obtained from the candidate position support relationship graph, and a set of corrected positions is determined based on the support relationship corresponding to the alternative candidate positions. Based on the spatial anomaly detection results, normal initial selection positions are retained from the initial selection position set, and the retained initial selection positions are merged with the corrected position set to obtain the target position set.

10. A map center server, characterized in that, include: The data acquisition unit is used to acquire basic business data, extract the address records to be located from the basic business data, and mark the confirmed points as trusted anchors with anchor credibility to obtain address processing data. A candidate location generation unit is used to parse the address record to be located to obtain a candidate location set containing candidate locations; The relationship graph construction unit is used to establish a support relationship between the candidate locations and the trusted anchor points based on the candidate location set and the address processing data, and to establish a mutual exclusion relationship between the candidate locations corresponding to the address record to be located, so as to obtain a candidate location support relationship graph. The joint disambiguation unit is used to pass the anchor point confidence level along the support relationship according to the candidate location support relationship graph to obtain a candidate location confidence level set, and to determine the initial location of the address record to be located according to the candidate location confidence level set and the mutual exclusion relationship to obtain an initial location set; The point aggregation unit is used to generate spatial anomaly detection results based on the initial selection set of locations, correct the abnormal locations in the initial selection set of locations based on the spatial anomaly detection results and the candidate location support relationship diagram, obtain the target location set, and output the target location set.

Images