Network marketing method based on intelligent matching of geographic data and new enterprise information

Abstract

The present invention is a network marketing method based on intelligently matching geographic data with newly opened business information. The method involves obtaining detailed information about business outlets and creating a basic snapshot information database, while also periodically acquiring and preprocessing newly opened business information. The network locations and business registration addresses are converted into longitude and latitude coordinates, and based on a preset matching distance range, the Haversine formula is used to filter out a set of matching network outlets that meet the criteria. The K-Means clustering algorithm is used to cluster the existing businesses corresponding to the matching network set, and the cluster centers are matched based on the characteristics of the newly opened businesses to obtain a list of optimal network outlets. Preset network outlet scoring indicators are used to calculate a comprehensive network outlet score based on the network scoring indicators, and the network outlet with the highest score is selected as the optimal matching network outlet. The newly opened business information is then pushed to the account manager corresponding to the optimal matching network outlet for marketing activities. If the information is rejected, it is re-matched. The present invention improves the accuracy and efficiency of network marketing and optimizes the service experience for newly opened businesses.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and mainly relates to a network marketing method based on intelligent matching of geographic data and new enterprise information. Background Art

[0002] In modern banking, marketing planning and customer relationship management are key to improving customer satisfaction and promoting business growth. With the development of financial technology, banks' marketing strategies are gradually shifting towards intelligent and data-driven approaches.

[0003] For example, the Chinese patent publication number CN115169859A discloses “an intelligent dispatch control method based on marketing tasks, comprising: S1, obtaining real-time marketing tasks and account manager information, classifying the marketing tasks according to the account manager by the intelligent marketing model, and calculating the classification confidence of the marketing tasks; S2, judging whether the classification confidence of the marketing tasks exceeds the confidence threshold, if so, directly dispatching it to the personal operation pool of the corresponding account manager, otherwise, going to step S3; S3, judging whether there is a management-account relationship between the marketing tasks, if so, dispatching it to the personal operation pool of the corresponding account manager, otherwise, going to step S4. Sent to the public operation pool of the branch; S4, the public operation pool of the branch distributes the marketing tasks therein to the personal operation pool of the corresponding account manager or receives marketing tasks from the personal operation pool of the account manager according to the received fine-tuning operation information. Compared with the existing technology, the present invention has the advantages of reducing distribution costs, improving distribution efficiency, and enhancing the intelligence of task distribution. However, the above invention has insufficient consideration of geographical location and distance information in the process of business opportunity allocation, which leads to insufficient refinement of business opportunity allocation, increases the risk of wasting business opportunities, and also increases the difficulty and cost of account managers and marketing personnel to perform tasks. In banking business, geographical location and corporate information are crucial to the successful execution of marketing tasks. Account managers need to accurately allocate marketing tasks according to the customer's geographical location, corporate size and business needs to ensure the effectiveness and pertinence of marketing activities. Therefore, there is an urgent need for a branch marketing method that takes geographical location and corporate information into consideration. Summary of the Invention

[0004] In order to solve the above-mentioned problems existing in the prior art, the present application provides a network marketing method based on intelligent matching of geographic data and new business information.

[0005] The technical solution of this application is as follows:

[0006] A method for network marketing based on intelligent matching of geographic data and newly opened business information, the method comprising:

[0007] Obtain relevant information about business outlets and create a basic snapshot information database for business outlets; regularly obtain new enterprise information from external systems, perform data preprocessing on the new enterprise information, and obtain a new enterprise information data set;

[0008] Based on the basic snapshot information database of business outlets and the information dataset of newly opened enterprises, an intelligent matching algorithm is used to obtain the best matching outlets for newly opened enterprises;

[0009] The newly opened enterprise information data set is pushed to the account manager of the corresponding best matching outlet, and the account manager conducts corresponding marketing activities.

[0010] As a preferred embodiment of the present invention, the business outlet-related information includes the outlet number, outlet name, outlet location, outlet type, types of outlet specialty services, number of outlet employees, average waiting time for corporate business calls at the queue machine, average daily number of corporate business transactions, number of corporate counters, and number of foreign exchange counters; the newly opened enterprise information includes the legal representative, establishment time, registered name, registered address, registered amount, industry type, and business scope.

[0011] As a preferred embodiment of the present invention, based on the basic snapshot information database of business outlets and the new enterprise information dataset, the best matching outlets for new enterprises are obtained by using an intelligent matching algorithm as follows:

[0012] Convert the branch locations in the basic snapshot information database of business branches and the registered addresses in the newly opened enterprise information dataset into structured latitude and longitude coordinate data;

[0013] Preset the matching distance range, and obtain the matching network point set according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range;

[0014] Clustering the existing enterprises corresponding to the matching network set using the K-Means clustering algorithm, matching the characteristics of the cluster centers of the newly opened and existing enterprises to obtain matching clustering results, and obtaining the optimal network list based on the matching clustering results; wherein the K-Means clustering algorithm uses the silhouette coefficient method to determine the optimal number of clusters K;

[0015] Calculate the outlet scoring indicators corresponding to the outlets in the optimal outlet list based on the basic snapshot information database of business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include the busyness of corporate business, distance priority, and existing industrial and commercial matching priority.

[0016] Determine the weights of the outlet scoring indicators based on user needs, calculate the comprehensive score of the outlet based on the outlet scoring indicator weights and the outlet scoring indicators, and select the outlet with the highest comprehensive score as the best matching outlet.

[0017] As a preferred embodiment of the present invention, the branch locations in the basic snapshot information library of business branches and the registered addresses in the newly opened enterprise information data set are converted into structured longitude and latitude coordinate data by using a geocoding service to convert the branch locations of business branches and the registered addresses of newly opened enterprises into longitude and latitude coordinate data; the converted longitude and latitude coordinate data are stored in a structured data table, wherein the longitude and latitude coordinate data includes ID data, longitude and latitude.

[0018] As a preferred embodiment of the present invention, a matching distance range is preset, and a matching network point set is obtained according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range:

[0019] Preset the matching distance range based on business needs and branch coverage, including the initial matching distance, matching distance step, and maximum matching distance; use the Haversine formula to calculate the geographical distance between the registered address of the newly opened enterprise and the branch location of each business branch, and match the branch list of the newly opened enterprise within the current matching distance range. If the matching result is empty, increase the matching distance step until the maximum matching distance is reached to obtain the matching branch set.

[0020] As a preferred embodiment of the present invention, the K-Means clustering algorithm is used to cluster the existing enterprises corresponding to the matching network set, and the characteristics of the cluster centers of the newly opened enterprises and existing enterprises are matched to obtain the matching clustering results, which are specifically:

[0021] Extracting business information of corresponding existing enterprises from the outlets in the matching outlet set, the business information including enterprise size, industry classification, and registered capital; standardizing the extracted business information using z-score standardization or min-max standardization to obtain standardized business information;

[0022] A range of K values ​​is set, and the optimal K value is found by performing multiple rounds of clustering. Specifically, for each K value, the standardized business information is clustered using the K-Means algorithm; the silhouette coefficient corresponding to each K value is calculated and recorded; the silhouette coefficients under different K values ​​are compared, and the K value with the largest silhouette coefficient is selected as the optimal cluster number K of the K-Means algorithm;

[0023] The optimal cluster number K is used as the K value of the K-Means clustering algorithm to perform cluster analysis on the standardized business information and obtain multiple cluster centers.

[0024] Calculate the Euclidean distance between the standardized industrial and commercial information of each newly opened enterprise and the cluster center of each existing enterprise to obtain the matching clustering results, and sort the matching clustering results in ascending order; assign the newly opened enterprise to the cluster corresponding to the cluster center with the highest ranking in the matching clustering results, extract the outlet information corresponding to all existing enterprises in the cluster, and obtain the best outlet list.

[0025] As a preferred embodiment of the present invention, the outlet scoring indicators corresponding to the outlets in the optimal outlet list are calculated based on the basic snapshot information database of the business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include the busyness of corporate business, distance priority, and existing industrial and commercial matching priority, specifically:

[0026] Based on the average waiting time for public business calls at the queuing machine, the average daily number of public business transactions, and the number of public counters in the basic snapshot information database of the business outlets, the expressions for the time factor, the number of transactions, and the number of counters are obtained. The public business busyness is calculated based on the time factor, the number of transactions, and the number of counters. The formula is:

[0027] Busyness of public services = (average waiting time for public services at the queuing machine / maximum acceptable waiting time) * ω1 + (average daily number of public services handled / maximum daily average number of services handled) * ω2 + (1 / number of public counters) * ω3;

[0028] Where ω1, ω2, and ω3 are the weights corresponding to the time factor, the number of services factor, and the number of counters factor, respectively. The time factor is expressed as the average waiting time for public services at the queue machine divided by the maximum acceptable waiting time. The number of services factor is expressed as the average daily number of public services handled / the maximum daily average number of services handled. The number of counters factor is expressed as 1 / the number of public counters.

[0029] Calculate the actual geographical distance between the latitude and longitude coordinates corresponding to the registered address of the newly opened enterprise and the location of the business outlets in the best outlet list, and standardize the calculated actual geographical distance to express it in the formula:

[0030] Normalized distance = 1-(actual geographical distance / maximum acceptable distance);

[0031] Using the normalized distance as the distance priority;

[0032] Convert the matching clustering results into matching priority, which can be expressed as follows:

[0033] Matching priority = 1 / (1 + Euclidean distance);

[0034] The comprehensive score of the outlets is expressed as follows:

[0035] Comprehensive branch score = corporate business busyness * ω'1 + distance priority * ω'2 + matching priority * ω'3;

[0036] Where ω'1, ω'2, and ω'3 are the weights corresponding to the busyness of public business, distance priority, and matching priority, respectively.

[0037] As a preferred embodiment of the present invention, the newly opened enterprise information data set is pushed to the customer manager of the corresponding best matching outlet. The customer manager of the best matching outlet chooses whether to manually return the newly opened enterprise information data set based on the actual situation of the outlet. If the newly opened enterprise information data set is returned, the newly opened enterprise is reallocated to the list to be allocated, and the best matching outlet other than the current best matching outlet is re-matched.

[0038] The present invention also provides a network marketing system based on intelligent matching of geographic data and newly opened enterprise information, the system comprising a data collection module, a geographic location encoding module, a distance calculation and network matching module, and a result output module, wherein:

[0039] The data acquisition module is used to obtain business outlet information from the API interface of the external system and create a basic snapshot information database of the business outlets; regularly obtain new enterprise information from the external system, perform data preprocessing on the new enterprise information, and obtain a new enterprise information data set;

[0040] The geographic location encoding module is used to convert the location of the outlets in the basic snapshot information database of the business outlets and the registered addresses in the newly opened enterprise information data set into structured latitude and longitude coordinate data using the API interface of the public map service, and transmit the data to the distance calculation and outlet matching module;

[0041] The distance calculation and network matching module is internally provided with an intelligent matching algorithm for obtaining the best matching network for a newly opened enterprise. Specifically:

[0042] Preset the matching distance range, and obtain the matching network point set according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range;

[0043] Clustering the existing enterprises corresponding to the matching network set using the K-Means clustering algorithm, matching the characteristics of the cluster centers of the newly opened and existing enterprises to obtain matching clustering results, and obtaining the optimal network list based on the matching clustering results; wherein the K-Means clustering algorithm uses the silhouette coefficient method to determine the optimal number of clusters K;

[0044] Calculate the outlet scoring indicators corresponding to the outlets in the optimal outlet list based on the basic snapshot information database of business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include the busyness of corporate business, distance priority, and existing industrial and commercial matching priority.

[0045] Determine the weights of the outlet scoring indicators based on user needs, calculate the comprehensive score of the outlet based on the outlet scoring indicator weights and outlet scoring indicators, and select the outlet with the highest comprehensive score as the best matching outlet;

[0046] The result output module is used to push the newly opened enterprise information data set to the account manager of the corresponding best matching outlet, and the account manager conducts corresponding marketing activities.

[0047] As a preferred embodiment of the present invention, the system is obtained by modular design based on the JAVA microservice architecture, and is internally provided with a Quartz timing tool. By combining the Quartz timing tool and the big data cloud platform intelligent scheduling component, the automatic operation of the data acquisition module, the geographic location encoding module, the distance calculation and network point matching module and the result output module is realized, and a data flow link corresponding to the automatic work is constructed. Based on the data flow link, a tracking mechanism and job lineage monitoring are realized.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1) This invention provides a network marketing method based on intelligently matching geographic data with newly opened business information. By converting the location information of business outlets and newly opened businesses into longitude and latitude coordinates, using the Haversine formula to calculate distances, and combining it with the K-Means clustering algorithm for feature matching, the method significantly improves the matching accuracy between newly opened businesses and business outlets, avoids matching errors that may occur in traditional methods, and enhances the effectiveness of marketing activities.

[0050] 2) This invention provides a network marketing method based on intelligently matching geographic data with new business information. It calculates a comprehensive score based on the busyness of corporate business, distance priority, and existing industrial and commercial matching priority. The network with the highest comprehensive score is selected as the best matching network. Dynamic adjustments are made based on feedback from relationship managers, optimizing the allocation of network resources. This allows new businesses to quickly and accurately obtain the most appropriate financial services while reducing ineffective resource waste.

[0051] 3) The present invention provides a network marketing method based on intelligent matching of geographic data and newly-established enterprise information, which directly pushes the newly-established enterprise information data set to the account manager of the best matching network. The account manager conducts marketing activities based on the actual situation. If the information is returned, it is redistributed and matched, which greatly improves the efficiency of marketing activities and reduces the time delay of information transmission and processing, so that newly-established enterprises can obtain the required services and support more quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The specific embodiments of the present invention are described below to facilitate understanding of the present invention by those skilled in the art. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations utilizing the concepts of the present invention are protected.

[0054] The present invention provides the following technical solution: a network marketing method based on intelligent matching of geographic data and new business information.

[0055] Example 1

[0056] This embodiment provides a network marketing method based on intelligent matching of geographic data and newly established business information, the method comprising:

[0057] S1. Obtaining relevant information about business outlets and creating a basic snapshot information database for business outlets;

[0058] S11. The branch information includes branch number, branch name, branch location, branch type, branch specialty services, number of branch employees, average waiting time for corporate business queuing, average daily number of corporate business transactions, number of corporate counters, and number of foreign exchange counters.

[0059] S2. Regularly obtain new enterprise information from an external system, perform data preprocessing on the new enterprise information, and obtain a new enterprise information data set;

[0060] Regularly obtain new business information from external systems. This information is usually uploaded in batches in file formats (such as CSV, JSON, etc.) and parsed into internal system database tables. This information includes the legal representative, establishment date, registered name, registered address, registered amount, industry type, and business scope.

[0061] In order to ensure the quality and availability of newly opened enterprise information, data preprocessing is performed on the acquired newly opened enterprise information. The data preprocessing includes data standardization and data cleaning. Specifically:

[0062] S21. Data standardization is to ensure that the data format is unified to facilitate subsequent processing and analysis, including field unification, data format conversion and data segmentation, among which:

[0063] The field unification includes field naming and data type unification. Field naming unification specifically involves revising the names of various fields to be consistent, such as "Company Name," "Registered Address," and "Date of Establishment." Data type unification specifically involves standardizing the data types of all fields, such as converting the "Date of Establishment" field to a unified date format (YYYY-MM-DD).

[0064] The data format conversion includes address standardization, phone number standardization and encoding conversion. Address standardization specifically standardizes address information and removes unnecessary symbols and redundant information, such as "No.123, ABC Street, City, Province, ZIP Code"; phone number standardization specifically standardizes phone number formats and removes spaces and specific characters; encoding conversion specifically converts different encoding formats (such as UTF-8 encoding) into a unified encoding format.

[0065] The data segmentation includes company name segmentation and address segmentation. The company name segmentation specifically involves segmenting the company name and extracting keywords to facilitate subsequent matching and analysis. The address segmentation specifically involves segmenting the address information and extracting information such as province, city, district, and street to facilitate the processing of geographic information.

[0066] S22. Data cleaning is to remove invalid and incomplete data to ensure data quality, including missing value processing, outlier processing, deduplication processing and invalid data elimination, among which:

[0067] The missing value processing includes missing value detection and missing value filling, specifically detecting whether there are missing values ​​in the data, such as if some fields are empty or invalid data, and filling the missing values ​​using appropriate strategies, such as using the mean, median, or a specific default value.

[0068] The outlier processing includes outlier detection and outlier processing, specifically detecting whether there are outliers in the data, such as obviously unreasonable establishment dates, addresses, etc., and processing the outliers, such as marking them as invalid data or correcting them;

[0069] The deduplication process includes duplicate data detection and duplicate data processing, specifically detecting whether there are duplicate records in the data and deleting the duplicate records to ensure the uniqueness of the data;

[0070] The invalid data elimination includes invalid data detection and invalid data processing, specifically detecting whether there is obviously invalid data in the data, such as empty addresses, companies whose registrations have been revoked, etc., and eliminating invalid data to ensure that the data is clean and usable;

[0071] S23. Storing the pre-processed new business information in a suitable database or data warehouse to obtain a new business information dataset for subsequent analysis and use. The suitable database may be a relational database, such as ORACLE, MySQL, or PostgreSQL. During storage, indexes are created for commonly used fields, such as company name and registered address, to improve query efficiency.

[0072] S3. Based on the basic snapshot information database of business outlets and the new enterprise information dataset, an intelligent matching algorithm is used to obtain the best matching outlets for new enterprises;

[0073] S31. Convert the branch locations in the basic snapshot information database of business branches and the registered addresses in the newly opened enterprise information dataset into structured latitude and longitude coordinate data;

[0074] Use geocoding services to convert the locations of business outlets and the registered addresses of newly established businesses into longitude and latitude coordinate data. Such geocoding services include, but are not limited to, Amap API, Google Maps API, and BaiduMap API. Store the converted longitude and latitude coordinate data in a structured data table, ensuring that each record contains the business outlet ID / company ID, longitude, and latitude.

[0075] Preferably, the converted coordinates are verified to ensure that the parsing results of the geocoding service are accurate; if the conversion fails, the address is marked as invalid and an error log is recorded for subsequent processing;

[0076] S32. Preset a matching distance range, and obtain a matching network point set based on the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range;

[0077] A reasonable matching distance range is preset based on business needs and network coverage. In this embodiment, the initial matching distance is set to 1 km, the matching distance step is 500 meters, and the maximum matching distance is 5 km. The Haversine formula or other geographic distance calculation method is used to calculate the geographic distance between the registered address of the newly opened enterprise and the location of each business network. The network list of the newly opened enterprise within the current matching distance range is matched. If the matching result is empty, the matching distance step is increased until the maximum matching distance is reached to obtain the matching network set.

[0078] S33. Clustering the existing enterprises corresponding to the matching network point set using a K-Means clustering algorithm, matching the characteristics of the cluster centers of the newly opened and existing enterprises to obtain matching clustering results, and obtaining an optimal network point list based on the matching clustering results; wherein the K-Means clustering algorithm uses a silhouette coefficient method to determine the optimal number of clusters K;

[0079] S331. Extracting business information of corresponding existing enterprises (existing corporate customers) from the outlets in the matching outlet set, including enterprise size, industry classification, registered capital, etc. Preferably, the extracted data also includes the outlet ID and corresponding corporate customer information;

[0080] Use z-score standardization or min-max standardization to standardize the extracted business information to eliminate the dimensional differences between different features;

[0081] S332. Using the silhouette coefficient method to determine the optimal number of clusters K is as follows:

[0082] Set a range of K values, for example, from 2 to 10, and prepare to perform multiple rounds of clustering to find the optimal K value; for each K value, use the K-Means algorithm to cluster the standardized business information; calculate and record the silhouette coefficient corresponding to each K value;

[0083] The silhouette coefficient measures the compactness and separation of the clustering results. The larger the value, the better the clustering effect. Therefore, the silhouette coefficients under different K values ​​are compared, and the K value with the largest silhouette coefficient is selected as the optimal cluster number K of the K-Means algorithm.

[0084] S333. Clustering the matching network point set using the K-Means clustering algorithm is specifically as follows:

[0085] Use the optimal cluster number K to run the K-Means clustering algorithm, perform cluster analysis on the standardized business information, and obtain multiple cluster centers;

[0086] S334. Standardize the business information (e.g., registered capital, industry classification, etc.) of newly established enterprises, i.e., their characteristics, based on the basic snapshot information database of business outlets, to make them comparable with the standardized data of existing enterprises;

[0087] Calculate the Euclidean distance between the standardized characteristics of each newly established enterprise and the cluster center of each existing enterprise to obtain matching cluster results, and sort the matching cluster results from smallest to largest. Assign the newly established enterprise to the cluster corresponding to the cluster center with the highest matching cluster result. Extract the corresponding network information of all existing enterprises in the cluster to obtain the optimal network list.

[0088] S34. Calculate outlet scoring indicators corresponding to outlets in the optimal outlet list based on the basic snapshot information database of business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include corporate business busyness, distance priority, and existing industrial and commercial matching priority.

[0089] S341, busyness of public business;

[0090] The busyness of corporate business is an important indicator for measuring the ability of business outlets to handle corporate business. It takes into account the efficiency and load of the outlets in handling corporate business. In this embodiment, the busyness of corporate business is calculated by the average waiting time for corporate business calls on the queuing machine, the average daily number of corporate business transactions, and the number of corporate counters. The average waiting time for corporate business calls on the queuing machine reflects the busyness of the outlet's corporate business. The longer the average waiting time, the busier the outlet's corporate business is, which affects the efficiency of new businesses in handling business at the outlet. The average daily number of corporate business transactions demonstrates the outlet's processing capacity for corporate business. The higher the average daily number of transactions, the stronger the outlet's processing capacity for corporate business, but it also means longer waiting times. The number of corporate counters: The number of corporate counters directly affects the outlet's capacity to handle corporate business. The more counters there are, the more corporate business that can be processed simultaneously, thereby reducing waiting time.

[0091] In summary, this embodiment considers the duration factor, the number of services, and the number of counters to calculate the busyness of corporate services. The duration factor is expressed as the average waiting time for corporate service calls at the queue machine / the maximum acceptable waiting time. The number of services is expressed as the average daily number of corporate services handled / the maximum daily average number of services handled. The number of counters is expressed as 1 / the number of corporate counters. Therefore, the busyness of corporate services is expressed as:

[0092] Busyness of public services = (average waiting time for public services at the queuing machine / maximum acceptable waiting time) * ω1 + (average daily number of public services handled / maximum daily average number of services handled) * ω2 + (1 / number of public counters) * ω3;

[0093] Where ω1, ω2, and ω3 are the weights corresponding to the duration factor, the number of transactions factor, and the number of counters factor, respectively. The weights are set according to actual needs.

[0094] S342, distance priority;

[0095] A distance priority is obtained based on the physical distance between the newly established enterprise and the business outlets, and the impact of the physical distance on the convenience of the enterprise's daily operations. Preferably, the closer the distance, the more convenient it is for the enterprise to handle business;

[0096] Therefore, the Haversine formula or other geographic distance calculation methods are used to calculate the actual geographic distance between the registered address of the newly-opened enterprise and the location of the business outlets in the optimal outlet list based on the latitude and longitude coordinate data. In order to convert the actual geographic distance into a unified scoring range (e.g., between 0 and 1), in this embodiment, the calculated actual geographic distance is normalized and expressed as follows:

[0097] Normalized distance = 1-(actual geographical distance / maximum acceptable distance);

[0098] Preferably, in order to more comprehensively evaluate the convenience of distance, the traffic conditions around the business outlets, such as the accessibility of public transportation facilities and road congestion, can be considered when calculating the distance priority. By taking the above traffic conditions as influencing factors, a traffic condition score is assigned to each business outlet based on the accessibility of public transportation, traffic congestion, etc. The traffic condition score ranges from 0 to 1, with 1 indicating the best traffic condition. The standardized distance is combined with the traffic condition score to form the final distance priority, which is expressed as follows:

[0099] Comprehensive distance priority = standardized distance × traffic status score;

[0100] S343, existing industrial and commercial matching priority;

[0101] The existing business matching priority is an indicator that measures the similarity between a newly opened enterprise and the existing customer base of existing enterprises of the business outlet. The higher the similarity, the more experience and professional services the outlet has in serving similar enterprises. In this embodiment, the matching clustering result in step S334 is converted into a priority score. Specifically, the calculated Euclidean distance, i.e., the similarity value, is converted into a priority score. A higher score indicates a higher matching priority. The formula is:

[0102] Matching priority = 1 / (1 + Euclidean distance);

[0103] S35. Determine the weights of the network scoring indicators based on user needs, calculate the comprehensive network score based on the network scoring indicator weights and the network scoring indicators, and select the network with the highest comprehensive score as the best matching network;

[0104] The comprehensive score of the outlets is expressed as follows:

[0105] Comprehensive branch score = corporate business busyness * ω'1 + distance priority * ω'2 + matching priority * ω'3;

[0106] Where ω'1, ω'2, and ω'3 are the weights corresponding to the busyness of public business, distance priority, and matching priority respectively;

[0107] The outlet with the highest comprehensive score in the best outlet list is selected as the best matching outlet;

[0108] S4. Use the Quartz timing tool to automatically push the newly opened enterprise information dataset to the account manager of the corresponding best matching outlet within the specified time period, and the account manager will carry out corresponding marketing activities;

[0109] Preferably, the account manager of the best matching outlet can choose to manually return the newly opened enterprise information data set, reallocate the newly opened enterprise to the pending allocation list, and re-match the remaining best matching outlets;

[0110] S5. Construct the data flow link corresponding to steps S1-S4 to ensure the update frequency and transmission timeliness of new business opportunity data; implement the tracking mechanism and job lineage monitoring based on the data flow link to track and monitor the method of pushing new customer data to outlets, wherein the tracking mechanism specifically relies on the big data cloud platform operation and maintenance center module to manually rerun the offline calculation job flow to recalculate the data of a certain time period in the past, and the system re-executes the data processing task of a previous specific business date or batch; the job lineage monitoring specifically monitors the operation status of upstream and downstream jobs and quickly locates the external dependencies of the job.

[0111] Example 2

[0112] This embodiment provides an application system for configuring dynamic forms based on agreements in a workflow engine. The system includes a data collection module, a geographic location encoding module, a distance calculation and network point matching module, and a result output module, wherein:

[0113] The data acquisition module is used to obtain business outlet information from the API interface of the external system and create a basic snapshot information database of the business outlets; regularly obtain new enterprise information from the external system, perform data preprocessing on the new enterprise information, and obtain a new enterprise information data set;

[0114] The geographic location encoding module is used to convert the location of the outlets in the basic snapshot information database of the business outlets and the registered addresses in the newly opened enterprise information data set into structured latitude and longitude coordinate data using the API interface of the public map service, and transmit the data to the distance calculation and outlet matching module;

[0115] The distance calculation and network matching module is internally provided with an intelligent matching algorithm for obtaining the best matching network for a newly opened enterprise. Specifically:

[0116] Preset the matching distance range, and obtain the matching network point set according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range;

[0117] Clustering the existing enterprises corresponding to the matching network set using the K-Means clustering algorithm, matching the characteristics of the cluster centers of the newly opened and existing enterprises to obtain matching clustering results, and obtaining the optimal network list based on the matching clustering results; wherein the K-Means clustering algorithm uses the silhouette coefficient method to determine the optimal number of clusters K;

[0118] Calculate the outlet scoring indicators corresponding to the outlets in the optimal outlet list based on the basic snapshot information database of business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include the busyness of corporate business, distance priority, and existing industrial and commercial matching priority.

[0119] Determine the weights of the outlet scoring indicators based on user needs, calculate the comprehensive score of the outlet based on the outlet scoring indicator weights and outlet scoring indicators, and select the outlet with the highest comprehensive score as the best matching outlet;

[0120] The result output module is used to push the newly opened enterprise information data set to the account manager of the corresponding best matching outlet, and the account manager conducts corresponding marketing activities;

[0121] Preferably, the system described in this embodiment is obtained by modular design based on the JAVA microservice architecture, and is internally provided with a Quartz timing tool. By combining the Quartz timing tool and the big data cloud platform intelligent scheduling component, the automatic operation of the data acquisition module, the geographic location encoding module, the distance calculation and network point matching module and the result output module is realized, and a data flow link corresponding to the automatic work is constructed, and a tracking mechanism and job lineage monitoring are realized based on the data flow link.

[0122] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention's description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A network marketing method based on intelligent matching of geographic data and new business information, characterized by: The method comprises: Obtain relevant information about business outlets and create a basic snapshot information database for business outlets; regularly obtain new enterprise information from external systems, perform data preprocessing on the new enterprise information, and obtain a new enterprise information data set; Convert the branch locations in the basic snapshot information database of business branches and the registered addresses in the newly opened enterprise information dataset into structured latitude and longitude coordinate data; Preset the matching distance range, and obtain the matching network point set according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range; Extracting the business information of corresponding existing enterprises from the outlets in the matching outlet set, and standardizing the extracted business information using z-score standardization or min-max standardization to obtain standardized business information; A range of K values ​​is set, and the optimal K value is found by performing multiple rounds of clustering. Specifically, for each K value, the standardized business information is clustered using the K-Means algorithm; the silhouette coefficient corresponding to each K value is calculated and recorded; the silhouette coefficients under different K values ​​are compared, and the K value with the largest silhouette coefficient is selected as the optimal cluster number K of the K-Means algorithm; The optimal cluster number K is used as the K value of the K-Means clustering algorithm to perform cluster analysis on the standardized business information and obtain multiple cluster centers. Calculate the Euclidean distance between the standardized business information of each newly established enterprise and the cluster center of each existing enterprise to obtain matching cluster results, and sort the matching cluster results from smallest to largest. Assign the newly established enterprise to the cluster corresponding to the cluster center with the highest ranking in the matching cluster results, extract the corresponding network information of all existing enterprises in the cluster, and obtain the optimal network list. Calculate the outlet scoring indicators corresponding to the outlets in the optimal outlet list based on the basic snapshot information database of business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include the busyness of corporate business, distance priority, and existing industrial and commercial matching priority. Determine the weights of the outlet scoring indicators based on user needs, calculate the comprehensive score of the outlet based on the outlet scoring indicator weights and outlet scoring indicators, and select the outlet with the highest comprehensive score as the best matching outlet; The newly opened enterprise information data set is pushed to the account manager of the corresponding best matching outlet, and the account manager conducts corresponding marketing activities.

2. The network marketing method based on intelligent matching of geographic data and new business information according to claim 1, characterized in that: The relevant information of the business outlets includes the outlet number, outlet name, outlet location, outlet type, types of specialty services of the outlet, number of outlet employees, average waiting time for corporate business calls at the queue machine, average daily number of corporate business transactions, number of corporate counters and number of foreign exchange counters; the information of the newly opened enterprise includes the legal representative, establishment time, registered name, registered address, registered amount, industry type and business scope.

3. The network marketing method based on intelligent matching of geographic data and new business information according to claim 2, characterized in that: The branch locations in the basic snapshot information library of business branches and the registered addresses in the newly opened enterprise information dataset are converted into structured longitude and latitude coordinate data. Specifically, the branch locations of business branches and the registered addresses of newly opened enterprises are converted into longitude and latitude coordinate data using geocoding services; the converted longitude and latitude coordinate data are stored in a structured data table, wherein the longitude and latitude coordinate data includes ID data, longitude and latitude.

4. The network marketing method based on intelligent matching of geographic data and new business information according to claim 3 is characterized in that: The matching distance range is preset, and the matching network point set is obtained according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range. Specifically: Preset the matching distance range based on business needs and branch coverage, including the initial matching distance, matching distance step, and maximum matching distance; use the Haversine formula to calculate the geographical distance between the registered address of the newly opened enterprise and the branch location of each business branch, and match the branch list of the newly opened enterprise within the current matching distance range. If the matching result is empty, increase the matching distance step until the maximum matching distance is reached to obtain the matching branch set.

5. The network marketing method based on intelligent matching of geographic data and new business information according to claim 4 is characterized in that: The business information includes enterprise size, industry classification and registered capital.

6. The network marketing method based on intelligent matching of geographic data and new business information according to claim 5 is characterized in that: The calculation steps of the network point scoring index are specifically as follows: Based on the average waiting time for public business calls at the queuing machine, the average daily number of public business transactions, and the number of public counters in the basic snapshot information database of the business outlets, the expressions for the time factor, the number of transactions, and the number of counters are obtained. The public business busyness is calculated based on the time factor, the number of transactions, and the number of counters. The formula is: Busyness of public services = (Average waiting time for public services at the queuing machine / Maximum acceptable waiting time) * + (average daily number of corporate transactions / maximum daily number of transactions) * + (1 / number of corporate counters) * ; Where, 、 and The weights of the time factor, the number of services, and the number of counters are respectively: the time factor is expressed as the average waiting time for public services at the queue machine / the maximum acceptable waiting time; the number of services is expressed as the average daily number of public services handled / the maximum daily average number of services handled; and the number of counters is expressed as 1 / the number of public counters; Calculate the actual geographical distance between the latitude and longitude coordinates corresponding to the registered address of the newly opened enterprise and the location of the business outlets in the best outlet list, and standardize the calculated actual geographical distance to express it in the formula: Normalized distance = 1 - (actual geographical distance / maximum acceptable distance); Using the normalized distance as the distance priority; Convert the matching clustering results into matching priority, which can be expressed as follows: Matching priority = 1 / (1 + Euclidean distance); The comprehensive score of the outlets is expressed as follows: Comprehensive score of branch = busyness of corporate business* +Distance Priority* +Match Priority* ; Where, 、 and They are the weights corresponding to the busyness of corporate business, distance priority and matching priority respectively.

7. The network marketing method based on intelligent matching of geographic data and new business information according to claim 6, characterized in that: The newly opened enterprise information dataset is pushed to the account manager of the corresponding best matching outlet. The account manager of the best matching outlet chooses whether to manually return the newly opened enterprise information dataset based on the actual situation of the outlet. If the newly opened enterprise information dataset is returned, the newly opened enterprise is reallocated to the waiting list and re-matched with the best matching outlet other than the current best matching outlet.

8. A network marketing system based on intelligent matching of geographic data and new business information, characterized by: The system includes a data collection module, a geographic location encoding module, a distance calculation and network point matching module, and a result output module, wherein: The data acquisition module is used to obtain business outlet information from the API interface of the external system and create a basic snapshot information database of the business outlets; regularly obtain new enterprise information from the external system, perform data preprocessing on the new enterprise information, and obtain a new enterprise information data set; The geographic location encoding module is used to convert the location of the outlets in the basic snapshot information database of the business outlets and the registered addresses in the newly opened enterprise information data set into structured latitude and longitude coordinate data using the API interface of the public map service, and transmit the data to the distance calculation and outlet matching module; The distance calculation and network matching module is internally provided with an intelligent matching algorithm for obtaining the best matching network for a newly opened enterprise. Specifically: Preset the matching distance range, and obtain the matching network point set according to the latitude and longitude coordinate data corresponding to the registered address, the latitude and longitude coordinate data corresponding to the network point location, and the matching distance range; Extracting the business information of corresponding existing enterprises from the outlets in the matching outlet set, and standardizing the extracted business information using z-score standardization or min-max standardization to obtain standardized business information; A range of K values ​​is set, and the optimal K value is found by performing multiple rounds of clustering. Specifically, for each K value, the standardized business information is clustered using the K-Means algorithm; the silhouette coefficient corresponding to each K value is calculated and recorded; the silhouette coefficients under different K values ​​are compared, and the K value with the largest silhouette coefficient is selected as the optimal cluster number K of the K-Means algorithm; The optimal cluster number K is used as the K value of the K-Means clustering algorithm to perform cluster analysis on the standardized business information and obtain multiple cluster centers. Calculate the Euclidean distance between the standardized business information of each newly established enterprise and the cluster center of each existing enterprise to obtain matching cluster results, and sort the matching cluster results from smallest to largest. Assign the newly established enterprise to the cluster corresponding to the cluster center with the highest ranking in the matching cluster results, extract the corresponding network information of all existing enterprises in the cluster, and obtain the optimal network list. Calculate the outlet scoring indicators corresponding to the outlets in the optimal outlet list based on the basic snapshot information database of business outlets and the newly opened enterprise information dataset. The outlet scoring indicators include the busyness of corporate business, distance priority, and existing industrial and commercial matching priority. Determine the weights of the outlet scoring indicators based on user needs, calculate the comprehensive score of the outlet based on the outlet scoring indicator weights and outlet scoring indicators, and select the outlet with the highest comprehensive score as the best matching outlet; The result output module is used to push the newly opened enterprise information data set to the account manager of the corresponding best matching outlet, and the account manager conducts corresponding marketing activities.

9. The network marketing system based on intelligent matching of geographic data and new business information according to claim 8, characterized in that: The system is modularly designed based on the JAVA microservice architecture and is equipped with a Quartz timing tool. By combining the Quartz timing tool and the intelligent scheduling component of the big data cloud platform, the system realizes the automatic operation of the data collection module, the geographic location encoding module, the distance calculation and network point matching module, and the result output module, and constructs a data flow link corresponding to the automatic operation. Based on the data flow link, the system realizes the number tracking mechanism and operation lineage monitoring.

Images