Tobacco marketing abnormal address management and control method, system and device based on GIS and multi-dimensional data fusion and medium

Abstract

The invention discloses a tobacco marketing abnormal address management and control method, system and device based on GIS and multi-dimensional data fusion and a medium, and belongs to the technical field of tobacco industry informatization and internet marketing. The technical problems that in the prior art, address risk control dimensionality is single, query precision is insufficient, data linkage is lacked and visual analysis capacity is weak are solved. According to the technical scheme, the method comprises the steps that abnormal address data is constructed and queried based on a multi-dimensional label; gIS visualization and correlation analysis based on a mixed similarity model: GIS positioning and range loading, spatial candidate set filtering, text similarity fine screening, mixed weighted scoring and visualization rendering and statistics are realized through the mixed similarity model of spatial index coarse screening and text algorithm fine screening, and correlation order detail form display based on GIS range triggering is carried out.

Description

Technical Field

[0001] This invention relates to the fields of tobacco industry informatization and internet marketing technology, specifically a method, system, device, and medium for controlling abnormal addresses in tobacco marketing based on GIS and multidimensional data fusion. Background Technology

[0002] In the operation of "Internet+" marketing platforms in the tobacco industry, risk control of order addresses is a crucial link in ensuring compliant operations and preventing abnormal behaviors such as fraudulent orders and arbitrage. Existing abnormal address management functions mostly rely on simple text matching or static address blacklists. When more refined control is needed for specific areas or known risk points, existing technologies are insufficient. The main shortcomings of existing technologies are: ① Limited control dimensions: Existing technologies lack unified classification, management, and query capabilities for "risk control addresses" (point-like risks) and "electronic fences" (area-like risks), leading to unclear risk type definitions and a "one-size-fits-all" control strategy; ② Insufficient query precision: Relying on coarse-grained conditions such as "province and city" for queries, unable to drill down to the "street" level, and unable to perform precise positioning and analysis using GPS location. When two address texts are similar but their actual GPS locations are far apart, misjudgment is easy; ③ Lack of data linkage: It is impossible to intuitively link abnormal addresses with "number of members placing orders" and specific member order information. Risk management is disconnected from actual order data, making it impossible to assess the actual impact range of risk addresses (i.e., how many members placed orders at the risk point). ④ Weak visualization capabilities: Lack of integration with map tools (GIS) makes it impossible to visually display the location of abnormal addresses, risk range (electronic fence or risk control radius), and the distribution of "similar addresses" within that range on a map, resulting in low analysis efficiency. ⑤ Poor timeliness: The definition of the "inclusion time" for abnormal addresses is vague, failing to accurately reflect the latest time point when the address was identified as risky.

[0003] The root cause of these deficiencies lies in the fact that the existing system fails to deeply integrate Geographic Information System (GIS), multidimensional data (address database, orders, members), and business risk control rules (risk control addresses, electronic fences). The system architecture remains at the level of "information silo" management based on text and tables, lacking the utilization of the "spatial attributes" of address data, thus failing to achieve refined, visualized, and interconnected risk management based on real geographical locations.

[0004] With the deepening of the "Internet+" marketing model in the tobacco industry, order delivery address data has exploded, and its dimensions have expanded to include geospatial data containing GPS coordinates. To meet stringent compliance requirements (such as sensitive area control and prevention of malicious order manipulation), building a sophisticated abnormal address risk control system has become a rigid requirement for the industry. The core challenges of existing technologies in abnormal address control are: ① Primitive methods: Many platforms still rely on simple address text keywords or static blacklists, resulting in low query precision and accuracy. ② Lack of spatial analysis capabilities: The inability to utilize GIS to analyze the true spatial attributes of addresses makes it difficult to achieve ** Based on geographical boundaries or radius **Refined management using methods such as "electronic fences" or "risk point radius". ③ Severe data silos: The address risk control system is disconnected from the order and membership systems, making it impossible to assess the actual business impact of risky addresses (involving the number of members and orders).

[0005] Existing technologies commonly used to overcome the core difficulties in abnormal address management include: ① Mainstream map service providers' "electronic fence" APIs (LBS services): primarily used for "real-time" LBS location services, rather than ** Marketing Risk Control Business Analysis **The system lacks the ability to analyze and correlate historical data (such as similar addresses or the number of members placing orders).** ② Traditional "address blacklist" and "text similarity" risk control systems rely entirely on text, easily leading to high false positive and false negative rates. For example, they cannot distinguish between addresses with similar text but geographically distant locations (spatial false positives) or addresses with different text but geographically close locations (spatial false negatives), and cannot provide data on GIS maps. The "table-chart-person-order" four-in-one system **Linked Analysis Interface. ③ Data Mining Risk Control Model Based on "Spatial Clustering" (DBSCAN): Purely data-driven, lacking business initiative, and cannot be actively defined by operations personnel.** Business Rules **(For example, actively setting risk control addresses or electronic fences) typically requires T+1 or longer for batch calculations, making real-time management impossible and not providing** Penetrating The analysis interface does not allow real-time viewing of order details associated with risk points on the map.

[0006] In summary, the root cause of the shortcomings of the existing system lies in its failure to deeply integrate the GIS spatial analysis engine, refined business risk control rules (points / areas), and real-time e-commerce business data (members / orders), resulting in address management remaining at the "text" or "simple spatial" level, and a serious disconnect between risk control and business analysis. Summary of the Invention

[0007] The technical objective of this invention is to provide a method, system, device, and medium for controlling abnormal addresses in tobacco marketing based on the fusion of GIS and multidimensional data, in order to solve the problems of single address risk control dimensions, insufficient query accuracy, lack of data linkage, and weak visualization analysis capabilities in the existing technology.

[0008] The technical objective of this invention is achieved as follows: a method for controlling abnormal addresses in tobacco marketing based on the fusion of GIS and multidimensional data, the specific method of which is as follows:

[0009] Construction and query of abnormal address data based on multidimensional tags: By introducing spatial database indexing technology and aggregating multidimensional business tags, the "number of similar addresses" and "number of members involved" calculated by spatial clustering are stored and displayed together with abnormal address entries as persistent or semi-persistent attributes, realizing refined query from "province, city, district" to "street", and providing dynamically aggregated statistical data for the front-end instrument panel. At the same time, the main data table provides key management operations such as "add whitelist" and "remove abnormality".

[0010] GIS Visualization and Association Analysis Based on Hybrid Similarity Model: A hybrid similarity model that uses spatial index coarse screening and text algorithm fine screening to achieve GIS positioning and range loading, spatial candidate set filtering, text similarity fine screening, hybrid weighted scoring, and visualization rendering and statistics;

[0011] The GIS-triggered associated order details form display activates the high-risk control area on the map (whether it is a circular area represented by a risk control address or a polygonal area represented by an electronic fence), retrieves all associated orders and members, and displays the original brief information of the corresponding orders in a pop-up window in the form of a detailed form, providing the operations staff with the most direct and original data evidence.

[0012] As a preferred approach, the construction and querying of abnormal address data based on multi-dimensional tags are as follows:

[0013] Data Acquisition and Geocoding: Address data (D) is acquired in real-time or in batches from order systems, membership systems, and third-party risk control systems via API interfaces. i For all unstructured address text D text Call the geocoding service to obtain the address text D text The unique GPS coordinates G gps = (lng,lat), in the form: G gps =f GeoEncode (D text Simultaneously utilizing address resolution service f GeoParse D text The parsing is done as a structured "province, city, district, street" A struct A struct=f GeoParse (D text );

[0014] Spatial Data Modeling and Tagging: In the database, spatial modeling is performed on address data, and "risk control addresses" defined by operators are stored as Point geographic types, while "electronic fences" are stored as Polygon geographic types; simultaneously, newly defined "risk control addresses / fences" are processed through background batch processing tasks. k Spatial association calculations are performed to ultimately obtain the key business tags from the data tables in the database. These key business tags are: {GPS location (Point), province, city, district, street, abnormal address type (Enum: risk control / fence), inclusion time (Timestamp), number of similar addresses (Integer), number of members involved (Integer), ...}; where "inclusion time" T... rec The update logic ensures the timeliness of the data; T rec =max(T) risk ,T geo );

[0015] Building a Spatial Index: Constructing an efficient R-tree spatial index on the GPS location Point field. spatial R-tree spatial index I spatial It is a height-balanced tree specifically designed to optimize range queries on multidimensional spatial data, reducing the time complexity of spatial query operations from O(N) to...

[0016] Multidimensional query and dynamic management: When a user initiates a query containing "province, city, district, street" on the front end... street and "Abnormal address type" Q type When querying, the backend performs an efficient multidimensional query; specifically as follows:

[0017] The formula for the dashboard overview statistics is as follows:

[0018] C geo_total =Count(A | |A.type = 'fence' | A.street ∈ Q) street );

[0019] C risk_total =Count(A | A.type = 'Risk Control' | A.street ∈ Q) street );

[0020] C member_total =Sum(A. Number of members involved | ...);

[0021] Master data table display: The calculated fields are retrieved directly from the database, eliminating the need for real-time calculation and greatly improving query performance; the table displays {detailed address, number of similar addresses, number of members involved, abnormal address type, GPS location, and inclusion time}.

[0022] Dynamic management: The front-end table provides "Add to whitelist" and "Remove exception" buttons. Clicking these buttons will update the corresponding address entry A. i The status of the address (e.g., marking its "abnormal address type" as "whitelisted" or "normal") enables real-time management.

[0023] Even better, by using the newly defined "risk control address / fence" R in the background batch processing task team. k The spatial association calculation is performed as follows:

[0024] Similar address calculation: R is calculated using a hybrid weighted scoring model. k The set of similar addresses S within the risk control scope similar (R k The total number is: Similar_Count(R) k ) = Count(S similar (R k ));

[0025] Membership calculation involves: reverse querying the order database and statistical analysis in S. similar (R k The formula for calculating the total number of unique members who have placed and received orders is as follows:

[0026] Member_Count(R k )=Count(DISTINCT UserID|OrderAddress∈

[0027] S similar (R k )).

[0028] As a preferred approach, GIS visualization and association analysis based on a hybrid similarity model are as follows:

[0029] GIS positioning and risk control range loading: The user clicks on the target address A in the table. target , target address A target G target The coordinates are sent to the GIS visualization engine; the GIS visualization engine then displays the coordinates on the map. target Coordinate marking, and based on target address A target The "address type" loads its risk control range R scope The format is as follows:

[0030]

[0031] Spatial candidate set filtering (coarse screening): To avoid traversing the entire database address set, R-tree spatial index I is used. spatial Perform efficient spatial queries to retrieve geographically located objects in R scope All addresses within the space form a "space candidate set" S. candidates :S candidates ={A c |ST_Within(A c .Gps,R scope )}.

[0032] Text similarity calculation (refined screening): Traversing the candidate set S candidates Each candidate address A in c The candidate address A is calculated using the Levenshtein Distance algorithm. c Detailed address text A c .text and target address text A target Text similarity score between .text tags text The edit distance algorithm quantifies the difference by calculating the minimum number of "insertion," "deletion," or "replacement" operations required to transform one string into another. The formula is as follows: Among them, Score text It is a value in the range [0,1], the closer to 1, the more similar; only the Score text A is higher than the preset threshold θ c Only then was it determined to be a similar address S similar That is: S similar ={A c |A c ∈S candidates ∧Score text (A c Then construct a hybrid weighted scoring model, taking into account "spatial distance" (Dist). spatial And "text similarity" score text This yields the final list S of "similar addresses". similar The formula is: Dist s patial(Ac)=f H aversine(Ac.Gps,G t (arget); where the calculation process of the mixed weighted scoring model is as follows: final similarity S f inal(Ac)=w1*(1-Norm(Dist s patial))+w2*Scoret ext; where w1 and w2 are weighting coefficients (w1 + w2 = 1), and Norm is the normalization function; only S f inal(Ac) is higher than the preset threshold θ. c Only then was it determined to be a similar address: S s imilar = Ac|AcINS c andidatesANDS f inal(Ac)>theta;

[0033] Visualization, rendering, and statistics: S similar The addresses in the list are rendered as secondary markers on the map, and in A target The information window displays the total number of members placing orders within the corresponding risk control range. scope_member .

[0034] As a preferred option, the related order details form triggered by the GIS scope is displayed as follows:

[0035] Penetration query trigger: Based on the set of addresses S determined on the map final_list (i.e. A) target and all of its S similar ), query the order database, and retrieve all orders in S final_list Member M who placed an order in the address list i and its order summary information O j The brief information form should include at least the serial number, order number, member ID, recipient, mobile phone number, province name, city name, district / county name, and township name.

[0036] Related data retrieval: Retrieve all data in the above address set S final_list Member M who placed an order for deduplication i and its order summary information O j (including O) j Actual delivery address A j and its GPS coordinates G j );

[0037] Detailed form display and closed loop: Next to the map interface, a pop-up window opens a detailed form. The detailed form displays a list of retrieved detailed data in tabular form: [{Serial Number:1, Order Number:'...', Member ID:'...', Recipient:'...',...},...]. Operations personnel can accurately grasp the distribution of risky orders through the detailed form, and combined with the "Add to Whitelist" and "Remove Abnormality" buttons, they can immediately execute the "Handling" operation after completing "Query-Analysis-Drill Down", forming a complete risk control closed loop of "Analysis-Decision-Execution".

[0038] A tobacco marketing abnormal address management system based on GIS and multidimensional data fusion, the system includes a data layer, a service layer, a communication layer and a presentation layer;

[0039] The data layer is used for spatial data storage;

[0040] The service layer is used for business clusters and technical applications;

[0041] The communication layer includes API interfaces, data buses, and access and security controls;

[0042] The presentation layer is used to realize a four-in-one visual interactive interface of "table-chart-person-order";

[0043] The service layer includes:

[0044] The data acquisition and standardization module is used to obtain address data (D) in real time or in batches from the order system, membership system, and third-party risk control system via API interfaces. i For all unstructured address text D text Call the geocoding service to obtain its unique GPS coordinates G gps = (lng,lat), and in the database, spatial modeling is performed on the address data, that is, the "risk control address" defined by the operators is stored as a Point geographic type, and the "electronic fence" is stored as a Polygon geographic type. The newly defined "risk control address / fence" R is processed by the background batch processing task team. k Perform spatial association calculations and simultaneously build an efficient R-tree spatial index I on the Gps location (Point) field. spatial ;

[0045] The multidimensional query and statistics module allows users to initiate queries containing "province, city, district, and street" on the front end. street and "Abnormal address type" Q type When querying, the backend executes efficient multidimensional queries;

[0046] The spatial distance and statistics module is used by users to click on the target address A in the table. target Its G target The coordinates are sent to the GIS visualization engine and used with the R-tree spatial index. spatial Perform efficient spatial queries to retrieve geographically located objects in R scope All addresses within the space form a "space candidate set" S. candidates ={A c |ST_Within(A c .Gps,R scope )}, then iterate through S candidates Each candidate address A in cThe detailed address text A is calculated using the "edit distance" algorithm. c .text and target address text A target Text similarity score between .text tags text Construct a hybrid weighted scoring model that comprehensively considers "spatial distance" Dist spatial And "text similarity" score text This yields the final list S of "similar addresses". similar , will S similar The addresses in the list are rendered as secondary markers on the map, and in A target The information window displays the total number of members placing orders within the corresponding risk control range. scope_member ;

[0047] A GIS visualization and analysis engine used to mark it on a map and, based on A target The "address type" loads its risk control range R scope ;

[0048] The associated order penetration module is used to retrieve all associated orders and members, and displays the original brief information of the detailed orders in a pop-up form.

[0049] As a preferred option, the newly defined "risk control address / fence" R is processed through the background batch task team. k The spatial association calculation is performed as follows:

[0050] Similar address calculation: R is calculated using a hybrid weighted scoring model. k The set of similar addresses S within the risk control scope similar (R k The total number is: Similar_Count(R) k ) = Count(S similar (R k ));

[0051] Membership calculation involves: reverse querying the order database and statistical analysis in S. similar (R k The formula for calculating the total number of unique members who have placed and received orders is as follows:

[0052] Member_Count(R k )=Count(DISTINCT UserID|OrderAddress∈

[0053] S similar (R k ));

[0054] The working process of the multidimensional query and statistics module is as follows:

[0055] ① Dashboard overview statistics, the formula is as follows:

[0056] C geo_total =Count(A | |A.type = 'fence' | A.street ∈ Q) street );

[0057] C risk_total =Count(A | A.type = 'Risk Control' | A.street ∈ Q) street );

[0058] C member_total =Sum(A. Number of members involved | ...);

[0059] ② Master data table display: The calculated fields are retrieved directly from the database, eliminating the need for real-time calculation and greatly improving query performance; the table displays {detailed address, number of similar addresses, number of members involved, abnormal address type, GPS location, and inclusion time}.

[0060] ③ Dynamic Management: The front-end table provides "Add to Whitelist" and "Remove Exception" buttons. Clicking these buttons will update the corresponding address entry A. i The status of the address (e.g., marking its "abnormal address type" as "whitelisted" or "normal") enables real-time management.

[0061] More specifically, the workflow of the associated order penetration module is as follows:

[0062] ① Penetration query trigger: Based on the set of addresses S determined on the map final_list (i.e. A) target and all of its S similar ), query the order database, and retrieve all orders in S final_list Member M who placed an order in the address list i and its order summary information O j The brief information form should include at least the serial number, order number, member ID, recipient, mobile phone number, province name, city name, district / county name, and township name.

[0063] ② Related data retrieval: Retrieve all data in the above address set S final_list Member M who placed an order for deduplication i and its order summary information O j (including O) j Actual delivery address A j and its GPS coordinates G j );

[0064] ③ Detailed Form Display and Closed Loop: Next to the map interface, a detailed form opens in a pop-up window. The detailed form displays a list of retrieved detailed data in tabular form: [{Serial Number:1, Order Number:'...', Member ID:'...', Recipient:'...',...},...]. Operations personnel can accurately grasp the distribution of risky orders through the detailed form, and combined with the "Add to Whitelist" and "Remove Abnormality" buttons, they can immediately execute the "Handling" operation after completing "Query-Analysis-Drill Down", forming a complete risk control closed loop of "Analysis-Decision-Execution".

[0065] An electronic device includes: a memory and at least one processor;

[0066] The memory contains computer programs;

[0067] The at least one processor executes the computer program stored in the memory, causing the at least one processor to perform the above-described method for controlling abnormal addresses in tobacco marketing based on GIS and multidimensional data fusion.

[0068] A computer-readable storage medium storing a computer program that can be executed by a processor to implement the above-described method for controlling abnormal addresses in tobacco marketing based on GIS and multidimensional data fusion.

[0069] The method, system, equipment, and medium for controlling abnormal addresses in tobacco marketing based on GIS and multidimensional data fusion of the present invention have the following advantages:

[0070] (i) After locating the target abnormal address on the map, this invention automatically calculates the "risk control range" of the target abnormal address and retrieves all "similar addresses" whose GPS locations fall within the "risk control range". Then, by clicking on the "risk control range" on the map, it retrieves and displays in a pop-up form all members who placed orders within the "risk control range" and their brief order information. The brief order information includes the order number, member ID, recipient, and mobile phone number.

[0071] (II) This invention integrates multi-dimensional data of GPS, address type, and members for abnormal address management: The abnormal address data object simultaneously contains text information of "province, city, district, and street", "GPS location" (latitude and longitude), "abnormal address type" (risk control / fence), and "latest collection time" (time stamp). The dashboard dynamically displays "number of electronic fences", "number of risk control addresses", and "number of members who placed orders" in real time, and changes in conjunction with the filtering conditions. It also supports precise matching queries of "abnormal address type" and "province, city, district, and street". The "address type" and "GPS location" are displayed intuitively in a table, and the "number of similar addresses" and "number of members involved" associated with the corresponding risk point / area are also displayed.

[0072] (III) Abnormal Address Sorting and Management Mechanism Based on the Latest Inclusion Timestamp: This mechanism ensures that the "Inclusion Time" column always displays the latest time when the corresponding abnormal address became a "Risk Control Address" or "Electronic Fence," using this as the default reverse sorting basis. This allows operations personnel to always be aware of the latest risk addresses immediately. The latest time point T... new =max(T) risk ,T geo );

[0073] (iv) By constructing a management and control system that integrates multi-dimensional data and supports the linkage of "table-graph", this invention breaks through the data silos and analysis bottlenecks of existing technologies, and realizes the leap from text-based passive defense to space-based proactive analysis.

[0074] (V) This invention achieves a technological leap from “text risk control” to “spatial risk control” and from “passive interception” to “active analysis” by constructing an analysis model that integrates GIS spatial index, hybrid similarity model, geodesic distance calculation and multidimensional business data (orders, members). It solves the problems of single address risk control dimension, insufficient query accuracy, lack of data linkage and weak visualization analysis capability in the existing technology.

[0075] (vi) This invention realizes the functions of two major business modules: data collection and standardization, and multidimensional query and statistics. In response to the two different forms of risk in tobacco marketing, namely "risk control address" (point) and "electronic fence" (area), a unified data modeling, storage and efficient query method is proposed. That is, by introducing spatial database indexing technology and combining it with multidimensional business tags, the performance bottleneck of traditional databases in efficiently handling spatial queries such as "point within area" and "point within radius" is solved. It is not just a simple query, but the "number of similar addresses" and "number of members involved" calculated by spatial clustering are used as persistent or semi-persistent attributes and stored and displayed together with abnormal address entries. It realizes refined query from "province, city, district" to "street" and provides dynamically aggregated statistical data for the front-end dashboard. At the same time, it provides key management operations such as "add whitelist" and "remove abnormality" in the main data table.

[0076] (vii) This invention innovatively proposes a hybrid similarity model of "spatial index coarse screening and text algorithm fine screening", which overcomes the misjudgment problem of "spatial similarity but different text" or "text similarity but spatial distance" in traditional methods, and at the same time connects the "risk address" and "actually affected orders" with a visual analysis bridge.

[0077] (viii) This invention enables a penetrating query from “risk range” to “specific order”; when the operator clicks on the highlighted risk control range on the map (whether it is a circular area represented by the risk control address or a polygonal area represented by the electronic fence), it is activated to retrieve all associated orders and members, and displays the original brief information of these orders in a pop-up window in the form of a detailed form, providing the operator with the most direct and original data evidence. Attached Figure Description

[0078] The invention will be further described below with reference to the accompanying drawings.

[0079] Appendix Figure 1 This is a schematic diagram of the structure of a tobacco marketing abnormal address control system based on the fusion of GIS and multidimensional data.

[0080] Appendix Figure 2 Flowchart for multidimensional query and management of abnormal addresses;

[0081] Appendix Figure 3 A flowchart for GIS correlation analysis and drop-through display;

[0082] Appendix Figure 4 This is a diagram showing the specific location of the abnormal address. Detailed Implementation

[0083] The following detailed description of the method, system, equipment, and medium for controlling abnormal addresses in tobacco marketing based on GIS and multidimensional data fusion, with reference to the accompanying drawings and specific embodiments, is provided in the specification.

[0084] Example 1:

[0085] This embodiment provides a method for controlling abnormal addresses in tobacco marketing based on the fusion of GIS and multidimensional data. The method is as follows:

[0086] S1. Construction and query of abnormal address data based on multi-dimensional tags: By introducing spatial database indexing technology and aggregating multi-dimensional business tags, the "number of similar addresses" and "number of members involved" calculated by spatial clustering are stored and displayed together with the abnormal address entries as persistent or semi-persistent attributes, realizing refined query from "province, city, district" to "street", and providing dynamically aggregated statistical data for the front-end instrument panel. At the same time, the main data table provides key management operations such as "add whitelist" and "remove abnormality".

[0087] S2. GIS Visualization and Association Analysis Based on Hybrid Similarity Model: A hybrid similarity model of "spatial index coarse screening and text algorithm fine screening" is used to realize GIS positioning and range loading, spatial candidate set filtering, text similarity fine screening, hybrid weighted scoring, and visualization rendering and statistics.

[0088] S3. Display of related order details form triggered by GIS range: Activate the high-risk control range already defined on the map (whether it is a circular area represented by the risk control address or a polygonal area represented by the electronic fence), retrieve all related orders and members, and display the original brief information of the corresponding orders in a pop-up window in the form of a detailed form, providing the most direct and original data evidence for the operations personnel.

[0089] As attached Figure 1 As shown, the construction and query of abnormal address data based on multi-dimensional tags in step S1 of this embodiment is as follows:

[0090] S101, Data Acquisition and Geocoding: Obtain address data in real-time or in batches from the order system, membership system, and third-party risk control system via API interfaces. i For all unstructured address text D text Call the geocoding service to obtain the address text D text The unique GPS coordinates G gps = (lng,lat), in the form: G gps =f GeoEncode (D text Simultaneously utilizing address resolution service f GeoParse D text The parsing is done as a structured "province, city, district, street" A struct A struct =f GeoParse (D text );

[0091] S102. Spatial Data Modeling and Tagging: In the database, spatial modeling is performed on the address data, and the "risk control address" defined by the operations personnel is stored as a Point geographic type, and the "electronic fence" is stored as a Polygon geographic type; at the same time, the newly defined "risk control address / fence" is processed through the background batch task team. k Spatial association calculations are performed to ultimately obtain the key business tags from the data tables in the database. These key business tags are: {GPS location (Point), province, city, district, street, abnormal address type (Enum: risk control / fence), inclusion time (Timestamp), number of similar addresses (Integer), number of members involved (Integer), ...}; where "inclusion time" T... rec The update logic ensures the timeliness of the data; T rec =max(T) risk ,T geo );

[0092] S103. Constructing a Spatial Index: Building an efficient R-tree spatial index on the Gps location Point field.spatial R-tree spatial index I spatial It is a height-balanced tree specifically designed to optimize range queries on multidimensional spatial data, reducing the time complexity of spatial query operations from O(N) to...

[0093] S104. Multidimensional Query and Dynamic Management: When a user initiates a query containing "province, city, district, street" on the front end... street and "Abnormal address type" Q type When querying, the backend performs an efficient multidimensional query; specifically as follows:

[0094] S10401, Dashboard overview statistics, the formula is as follows:

[0095] C geo_total =Count(A | |A.type = 'fence' | A.street ∈ Q) street );

[0096] C risk_total =Count(A | A.type = 'Risk Control' | A.street ∈ Q) street );

[0097] C member_total =Sum(A. Number of members involved | ...);

[0098] S10402, Master Data Table Display: Directly retrieves pre-calculated fields from the database, eliminating the need for real-time calculation and greatly improving query performance; the table displays {detailed address, number of similar addresses, number of members involved, abnormal address type, GPS location, and inclusion time}.

[0099] S10403, Dynamic Management: The front-end table provides "Add to Whitelist" and "Remove Exception" buttons. Clicking these buttons will update the corresponding address entry A. i The status of the address (e.g., marking its "abnormal address type" as "whitelisted" or "normal") enables real-time management.

[0100] In step S102 of this embodiment, the newly defined "risk control address / fence" R is processed by the background batch processing task team. k The spatial association calculation is performed as follows:

[0101] S10201, Similar Address Calculation: R is calculated using a hybrid weighted scoring model. k The set of similar addresses S within the risk control scope similar (R k The total number is: Similar_Count(R) k ) = Count(S similar (Rk ));

[0102] S10202, Membership Calculation Involved: Reverse query of the order database, statistics in S similar (R k The formula for calculating the total number of unique members who have placed and received orders is as follows:

[0103] Member_Count(R k )=Count(DISTINCT UserID|OrderAddress∈

[0104] S similar (R k )).

[0105] The GIS visualization and association analysis based on the hybrid similarity model in step S2 of this embodiment is as follows:

[0106] S201, GIS Positioning and Risk Control Range Loading: When the user clicks on the target address A in the table... target , target address A target G target The coordinates are sent to the GIS visualization engine; the GIS visualization engine then displays the coordinates on the map. target Coordinate marking, and based on target address A target The "address type" loads its risk control range R scope The format is as follows:

[0107]

[0108] S202, Spatial Candidate Set Filtering (Coarse Screening): To avoid traversing the entire database address set, R-tree spatial index I is used. spatial Perform efficient spatial queries to retrieve geographically located objects in R scope All addresses within the space form a "space candidate set" S. candidates :S candidates ={A c |ST_Within(A c .Gps,R scope )}.

[0109] S203, Text Similarity Calculation (Refined Screening): Traversing the Candidate Set S candidates Each candidate address A in c The candidate address A is calculated using the Levenshtein Distance algorithm. c Detailed address text A c .text and target address text A target Text similarity score between .text tagstext The edit distance algorithm quantifies the difference by calculating the minimum number of "insertion," "deletion," or "replacement" operations required to transform one string into another. The formula is as follows: Among them, Score text It is a value in the range [0,1], the closer to 1, the more similar; only the Score text A is higher than the preset threshold θ c Only then was it determined to be a similar address S similar That is: S similar ={A c |A c ∈S candidates ∧Score text (A c Then construct a hybrid weighted scoring model, taking into account "spatial distance" (Dist). spatial And "text similarity" score text This yields the final list S of "similar addresses". similar The formula is: Dist s patial(Ac)=f H aversine(Ac.Gps,G t (arget); where the calculation process of the mixed weighted scoring model is as follows: final similarity S f inal(Ac)=w1*(1-Norm(Dist s patial))+w2*Score t ext; where w1 and w2 are weighting coefficients (w1 + w2 = 1), and Norm is the normalization function; only S f inal(Ac) is higher than the preset threshold θ. c Only then was it determined to be a similar address: S s imilar = Ac|AcINS c andidatesANDS f inal(Ac)>theta;

[0110] S204, Visualization Rendering and Statistics: This section discusses S... similar The addresses in the list are rendered as secondary markers on the map, and in A target The information window displays the total number of members placing orders within the corresponding risk control range. scope_member .

[0111] As attached Figure 3 and 4 As shown, the associated order details form triggered by the GIS range in step S3 of this embodiment is displayed as follows:

[0112] S501, Penetration Query Trigger: Based on the set of addresses S determined on the map. final_list (i.e. A) target and all of its S similar ), query the order database, and retrieve all orders in S final_list Member M who placed an order in the address list i and its order summary information O j The brief information form should include at least the serial number, order number, member ID, recipient, mobile phone number, province name, city name, district / county name, and township name.

[0113] S502, Related Data Retrieval: Retrieve all data in the above address set S final_list Member M who placed an order for deduplication i and its order summary information O j (including O) j Actual delivery address A j and its GPS coordinates G j );

[0114] S503, Detailed Form Display and Closed Loop: Next to the map interface, a detailed form opens in a pop-up window. The detailed form displays a list of retrieved detailed data in tabular form: [{Serial Number: 1, Order Number: '...', Member ID: '...', Recipient: '...', ...}, ...]. Operations personnel can accurately grasp the distribution of risky orders through the detailed form, and combined with the "Add to Whitelist" and "Remove Abnormality" buttons, they can immediately execute the "Handling" operation after completing "Query-Analysis-Drill Down", forming a complete risk control closed loop of "Analysis-Decision-Execution".

[0115] Example 2:

[0116] As attached Figure 1 As shown, this embodiment provides a tobacco marketing abnormal address management system based on GIS and multidimensional data fusion. The system includes a data layer, a service layer, a communication layer, and a presentation layer.

[0117] The data layer is used for spatial data storage;

[0118] The service layer is used for business clusters and technical applications;

[0119] The communication layer includes API interfaces, data buses, and access and security controls;

[0120] The presentation layer is used to realize a four-in-one visual interactive interface of "table-chart-person-order";

[0121] The service layer includes:

[0122] The data acquisition and standardization module is used to obtain address data (D) in real time or in batches from the order system, membership system, and third-party risk control system via API interfaces. i For all unstructured address text D text Call the geocoding service to obtain its unique GPS coordinates G gps = (lng,lat), and in the database, spatial modeling is performed on the address data, that is, the "risk control address" defined by the operators is stored as a Point geographic type, and the "electronic fence" is stored as a Polygon geographic type. The newly defined "risk control address / fence" R is processed by the background batch processing task team. k Perform spatial association calculations and simultaneously build an efficient R-tree spatial index I on the Gps location (Point) field. spatial ;

[0123] The multidimensional query and statistics module allows users to initiate queries containing "province, city, district, and street" on the front end. street and "Abnormal address type" Q type When querying, the backend executes efficient multidimensional queries;

[0124] The spatial distance and statistics module is used by users to click on the target address A in the table. target Its G target The coordinates are sent to the GIS visualization engine and used with the R-tree spatial index. spatial Perform efficient spatial queries to retrieve geographically located objects in R scope All addresses within the space form a "space candidate set" S. candidates ={A c |ST_Within(A c .Gps,R scope )}, then iterate through S candidates Each candidate address A in c The detailed address text A is calculated using the "edit distance" algorithm. c .text and target address text A target Text similarity score between .text tags text Construct a hybrid weighted scoring model that comprehensively considers "spatial distance" Dist spatial And "text similarity" score text This yields the final list S of "similar addresses". similar , will S similar The addresses in the list are rendered as secondary markers on the map, and in A target The information window displays the total number of members placing orders within the corresponding risk control range. scope_member ;

[0125] A GIS visualization and analysis engine used to mark it on a map and, based on A target The "address type" loads its risk control range R scope ;

[0126] The associated order penetration module is used to retrieve all associated orders and members, and displays the original brief information of the detailed orders in a pop-up form.

[0127] In this embodiment, the newly defined "risk control address / fence" R is processed through a background batch processing task team. k The spatial association calculation is performed as follows:

[0128] ① Similar address calculation: R is calculated using a hybrid weighted scoring model. k The set of similar addresses S within the risk control scope similar (R k The total number is: Similar_Count(R) k ) = Count(S similar (R k ));

[0129] ②Membership calculation: Reverse query the order database and calculate the statistics in S similar (R k The formula for calculating the total number of unique members who have placed and received orders is as follows:

[0130] Member_Count(R k )=Count(DISTINCT UserID|OrderAddress∈

[0131] S similar (R k )).

[0132] The working process of the multidimensional query and statistics module in this embodiment is as follows:

[0133] ① Dashboard overview statistics, the formula is as follows:

[0134] C geo_total =Count(A | |A.type = 'fence' | A.street ∈ Q) street );

[0135] C risk_total =Count(A | A.type = 'Risk Control' | A.street ∈ Q) street );

[0136] C member_total =Sum(A. Number of members involved | ...);

[0137] ② Master data table display: The calculated fields are retrieved directly from the database, eliminating the need for real-time calculation and greatly improving query performance; the table displays {detailed address, number of similar addresses, number of members involved, abnormal address type, GPS location, and inclusion time}.

[0138] ③ Dynamic Management: The front-end table provides "Add to Whitelist" and "Remove Exception" buttons. Clicking these buttons will update the corresponding address entry A. i The status of the address (e.g., marking its "abnormal address type" as "whitelisted" or "normal") enables real-time management.

[0139] The working process of the associated order penetration module in this embodiment is as follows:

[0140] ① Penetration query trigger: Based on the set of addresses S determined on the map final_list (i.e. A) target and all of its S similar ), query the order database, and retrieve all orders in S final_list Member M who placed an order in the address list i and its order summary information O j The brief information form should include at least the serial number, order number, member ID, recipient, mobile phone number, province name, city name, district / county name, and township name.

[0141] ② Related data retrieval: Retrieve all data in the above address set S final_list Member M who placed an order for deduplication i and its order summary information O j (including O) j Actual delivery address A j and its GPS coordinates G j );

[0142] ③ Detailed Form Display and Closed Loop: Next to the map interface, a detailed form opens in a pop-up window. The detailed form displays a list of retrieved detailed data in tabular form: [{Serial Number:1, Order Number:'...', Member ID:'...', Recipient:'...',...},...]. Operations personnel can accurately grasp the distribution of risky orders through the detailed form, and combined with the "Add to Whitelist" and "Remove Abnormality" buttons, they can immediately execute the "Handling" operation after completing "Query-Analysis-Drill Down", forming a complete risk control closed loop of "Analysis-Decision-Execution".

[0143] Example 3:

[0144] This embodiment also provides an electronic device, including: a memory and a processor;

[0145] The memory stores the instructions executed by the computer.

[0146] The processor executes computer execution instructions stored in the memory, causing the processor to execute the tobacco marketing abnormal address control method based on GIS and multidimensional data fusion in any embodiment of the present invention.

[0147] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor can be a microprocessor or any conventional processor.

[0148] Memory is used to store computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and by accessing data stored in the memory. Memory can mainly include a program storage area and a data storage area. The program storage area can store the operating system, at least one application program required for a function, etc.; the data storage area can store data created based on the use of the terminal, etc. In addition, memory can also include high-speed random access memory, and can also include non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart memory cards (SMC), secure digital cards (SD cards), flash memory cards, at least one disk storage device, flash memory devices, or other volatile solid-state storage devices.

[0149] Example 4:

[0150] This embodiment also provides a computer-readable storage medium storing multiple instructions, which are loaded by a processor to cause the processor to execute the tobacco marketing abnormal address control method based on GIS and multidimensional data fusion in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium may be provided, on which software program code implementing the functions of any of the above embodiments is stored, and the computer (or CPU or MPU) of the system or device may read and execute the program code stored in the storage medium.

[0151] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0152] Storage media embodiments for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RYM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0153] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0154] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion unit connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion unit execute some and all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management method, characterized in that, The method is specifically as follows: Based on multi-dimensional label, abnormal address data construction and query: by introducing spatial database index technology, set multi-dimensional business label, "similar address number" and "involved member number" calculated by spatial clustering are stored and displayed as persistent or semi-persistent attributes with abnormal address entries, realize fine query from "province-city-district" to "street", and provide dynamic aggregated statistical data for front-end dashboard, at the same time, provide key management operations of "adding white list" and "removing abnormal" in main data table; Based on hybrid similarity model, GIS visualization and correlation analysis: through hybrid similarity model of "spatial index coarse screening, text algorithm fine screening", realize GIS positioning and range loading, spatial candidate set filtering, text similarity fine screening, hybrid weighted scoring, and visualization rendering and statistics; Based on GIS range trigger, associated order detail form display: activate the high-risk control range drawn on the map, retrieve all associated orders and members, and display the corresponding order original brief information in the form of a detailed form pop-up window, provide the most direct and original data evidence for operation personnel. 2.The GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management method according to claim 1, characterized in that, Based on multi-dimensional label, abnormal address data construction and query are as follows: Data collection and geocoding: real-time or batch address data D from order system, member system and third-party risk control system through API interface i , for all unstructured address text D text , call geocoding service to obtain address text D text Unique GPS coordinates G gps =(lng, lat), in the form of: G gps =f GeoEncode (D text ); at the same time, use address resolution service f GeoParse to parse D text into structured "province, city, district, street" A struct , A struct =f GeoParse (D text ); Spatial data modeling and labeling: In the database, address data is spatially modeled, and "risk control addresses" defined by operation personnel are stored as Point geographic types, and "electronic fences" are stored as Polygon geographic types; at the same time, through a background batch processing task, new definitions of "risk control addresses / fences" R k Perform spatial correlation calculation to finally obtain the key business labels of the data table in the database; wherein the key business labels of the data table are {GPS location (Point), province, city, district, street, abnormal address type (Enum: risk control / fence), collection time (Timestamp), similar address quantity (Integer), involved member number (Integer),...}; wherein the update logic of "collection time" T rec ensures the timeliness of the data; T rec =max(T risk ,T geo ) Build spatial index: build efficient R-tree spatial index I on GpsLocationPoint field spatial R-tree spatial index I spatial is a highly balanced tree, which is specifically designed to optimize range queries on multi-dimensional spatial data, making the time complexity of spatial query operations from O(N) to Multi-dimensional query and dynamic management: when the user initiates a query containing "province-city-district-street" Q street and "abnormal address type" Q type in the front end, the back end executes efficient multi-dimensional queries; as follows: Dashboard overview statistics, formula as follows: C geo_total = Count(A || A.type = 'fence' A.street E Q street ); C risk_total = Count(A || A.type = 'fraud' A.street E Q street ); C member_total =Sum(A. Number of members involved | ...); Main data table display: directly retrieve the calculated fields from the database; table display {detailed address, similar address number, involved member number, abnormal address type, GPS position, collection time}; Dynamic management: The front-end table provides "add white list" and "remove exception" buttons, which will update the status of the corresponding address entry A after being clicked, thereby achieving instant management. i Dynamic management: The front-end table provides "add white list" and "remove exception" buttons, which will update the status of the corresponding address entry A after being clicked, thereby achieving instant management. 3.The GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management method according to claim 2, characterized in that, The new defined "risk control address / fence" R is passed to the background batch task queue k The spatial correlation calculation is performed as follows: Similar address calculation: R is calculated by mixing the weighted scoring model k Similar address set S within the risk control range similar (R k ), the total number is: Similar_Count(R k ) = Count(S similar (R k )) Member calculation: Reverse query order library, count the total number of members in S similar (R k ) set of orders, the formula is as follows: Member_Count(R k ) = Count(DISTINCT UserID | OrderAddress ∈ S similar (R k )).

4. The GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management method according to claim 1, characterized in that, Based on hybrid similarity model, GIS visualization and correlation analysis are as follows: GIS positioning and risk range loading: user clicks target address A in table target , sends target address A target 's G target IS coordinates to GIS visualization engine; The GIS visualization engine places a coordinate marker on the map G target and loads its risk range R according to the "address type" of the target address A target , in the form of: scope ​ Spatial candidate set filtering: Utilize R-tree spatial index I spatial , perform efficient spatial query to get all addresses that geographically fall within R scope , form "spatial candidate set" S candidates : S candidates = {A c | ST_Within(A c .Gps, R scope}. Text similarity calculation: Traversing the candidate set S candidates Each candidate address A in c The candidate address A is calculated using the "edit distance" algorithm. c Detailed address text A c .text and target address text A target Text similarity score between .text tags text The edit distance algorithm quantifies the difference by calculating the minimum number of "insertion," "deletion," or "replacement" operations required to transform one string into another. The formula is as follows: Among them, Score text It is a value in the range [0,1], the closer to 1, the more similar; only the Score text A is higher than the preset threshold θ c Only then was it determined to be a similar address S similar That is: S similar ={A c |A c ∈S candidates ∧Score text (A c Then construct a hybrid weighted scoring model, taking into account "spatial distance" (Dist). spatial And "text similarity" score text This yields the final list S of "similar addresses". similar The formula is: Dist s patial(Ac)=f H aversine(Ac.Gps,G t (arget); where the calculation process of the mixed weighted scoring model is as follows: final similarity S f inal(Ac)=w1*(1-Norm(Dist s patial))+w2*Score t ext; where w1 and w2 are weighting coefficients (w1 + w2 = 1), and Norm is the normalization function; only S f inal(Ac) is higher than the preset threshold θ. c Only then was it determined to be a similar address: S s imilar = Ac|AcINS c andidatesANDS f inal(Ac)>theta; Visual rendering and statistics: S similar The addresses in the list are rendered as secondary markers on the map, and in the information window of A target , the total number of members C scope_member in the corresponding risk control range is displayed. 5.The GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management method according to claim 1, characterized in that, Based on GIS range trigger, associated order detail form display is as follows: Penetration query trigger: according to the address set S determined on the map final_list , query the order library to retrieve all members M final_list whose orders are in S i address list and order brief information O j ; wherein the brief information form includes at least serial number, order number, member ID, consignee, mobile phone number, province name, city name, district name and township name; Correlation data retrieval: retrieve all the members M in the above address set S final_list and their order brief information O i and their order brief information O j ; Detail form display and closed loop: in the form of a pop-up window beside the map interface, open a detailed form, the detailed form displays the retrieved detailed data list in the form of a table: [{serial number: 1, order number: '...', member ID: '...', consignee: '...',...},...]; operation personnel accurately master the distribution of risk orders through the detailed form, and combine the "add white list" and "remove abnormal" buttons, so that operation personnel can immediately execute "disposal" operation after completing "query-analysis-drill down", form a complete risk control closed loop of "analysis-decision-execution".

6. A tobacco marketing abnormal address management and control system based on GIS and multi-dimensional data fusion, characterized in that, The system comprises a data layer, a service layer, a communication layer and a display layer; The data layer is used for spatial data storage; The service layer is used for business cluster and technical application; The communication layer comprises API interface, data bus and permission and security control; The display layer is used for realizing visual interactive interface of "table-map-person-single"; The service layer comprises: Data collection and standardization module, used to obtain address data D from order system, member system and third-party risk control system in real time or batch through API interface i ; for all unstructured address text D text , call the geocoding service to obtain its unique GPS coordinates G gps =(lng, lat), and in the database, spatialize the address data, that is, store the "risk control address" defined by the operation personnel as Point geographic type, and store the "electronic fence" as Polygon geographic type, and through the background batch processing task, define the "risk control address / fence" R k Perform spatial correlation calculation, and build efficient R-tree spatial index I on the Gps location (Point) field spatial ; A multi-dimensional query and statistics module is used for the user to initiate a query containing "province-city-district-street" Q street and "abnormal address type" Q type in the front end, and the back end executes efficient multi-dimensional query; A spatial distance and statistics module is used to click on a target address A in a table by a user target , and its G target -coordinate is sent to a GIS visualization engine, and a R-tree spatial index I spatial is used to perform efficient spatial queries to obtain all addresses that fall within R scope , forming a "spatial candidate set” S candidates ={A c ∣ST_Within(A c .Gps,R scope )}; each candidate address A candidates in S c is traversed, and a "text similarity” Score c between a detailed address text A target .text and a target address text A text .text is calculated using an "edit distance” algorithm; a hybrid weighted scoring model is constructed to consider "spatial distance” Dist spatial and "text similarity” Score text , and a final "similar address” list S similar is obtained; addresses in the list S similar are rendered as secondary markers on a map, and a total number of members C target in a corresponding risk control range is displayed in an information window of A scope_member ; GIS visualization analysis engine for marking them on a map and loading their risk control range R according to the "address type" of A target ; and scope ; The associated order penetration module is used for retrieving all associated orders and members, and displaying the original brief information of detailed orders in the form of a detailed form pop-up window. 7.The GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management and control system according to claim 6, characterized in that, The new defined "risk control address / fence" R is passed to the background batch task queue k The spatial correlation calculation is performed as follows: Similar address calculation: R is calculated by mixing the weighted scoring model k Similar address set S within the risk control range similar (R k ), the total number is: Similar_Count(R k ) = Count(S similar (R k )); Member calculation: Reverse query order library, count the total number of members in S similar (R k ) set of orders, the formula is as follows: Member_Count(R k ) = Count(DISTINCT UserID | OrderAddress ∈ S similar (R k )); The working process of the multi-dimensional query and statistical module is as follows: ①Dashboard overview statistics, formula as follows: C geo_total = Count(A || A.type = 'fence' A.street E Q street ); C risk_total = Count(A || A.type = 'fraud' A.street E Q street ); C member_total = Sum ( A. Involved Members |... ); ②Main data table display: directly retrieve the calculated fields from the database; table display {detailed address, similar address number, involved member number, abnormal address type, GPS position, collection time}; ③Dynamic management: The front-end table provides "add white list" and "remove exception" buttons. After clicking, the corresponding address entry A will be updated, so as to realize instant management. i management: The front-end table provides "add white list" and "remove exception" buttons. After clicking, the corresponding address entry A will be updated, so as to realize instant management. 8.The GIS and multi-dimensional data fusion-based tobacco marketing abnormal address management and control system according to claim 6 or 7, characterized in that, The working process of the associated order penetration module is as follows: ①Penetration query trigger: according to the address set S determined on the map final_list , query the order library to retrieve all members M who have placed orders at the S final_list address list and their order summary information O i ; wherein the summary information form at least includes serial number, order number, member ID, recipient, mobile phone number, province name, city name, district name and township name j ; (ii) Correlation data retrieval: retrieve all the data in the above address set S final_list The de-duplication member M i and order brief information O j ; ③Detail form display and closed loop: beside the map interface, a detailed form is opened in the form of a pop-up window, and the detailed form displays the detailed data list retrieved in the form of a table: [{serial number: 1, order number: '...', member ID: '...', consignee: '...',...},...]; the operation personnel accurately master the distribution of the risk orders through the detailed form, and combine the "add white list" and "remove exception" buttons, so that the operation personnel can immediately perform the "disposal" operation after completing "query-analysis-drill down", and form a complete risk control closed loop of "analysis-decision-execution".

9. An electronic device, comprising: Comprise: a memory and at least one processor; wherein the memory has stored thereon a computer program; the at least one processor executes the computer program stored in the memory, so that the at least one processor executes the tobacco marketing abnormal address management and control method based on GIS and multi-dimensional data fusion as claimed in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that, the computer program stored in the computer readable storage medium can be executed by the processor to realize the tobacco marketing abnormal address management and control method based on GIS and multi-dimensional data fusion as claimed in any one of claims 1 to 5.

Images