5G potential customer prediction method and system for cross marketing

Abstract

The invention discloses a 5G potential customer prediction method and system for cross marketing. Relates to the technical field of data analysis, and the method comprises the following steps: analyzing related data influencing the timeliness and accuracy of 5G potential customer data to obtain an evaluation coefficient influencing the timeliness of the 5G potential customer data and an evaluation coefficient influencing the accuracy of the 5G potential customer data; performing comprehensive evaluation on the evaluation coefficient influencing the timeliness and accuracy of the 5G potential customer data to obtain an evaluation coefficient influencing the quality of the 5G potential customer data, performing preliminary regulation and control according to the evaluation coefficient influencing the timeliness of the 5G potential customer data, and performing final regulation and control according to the evaluation coefficient influencing the quality of the 5G potential customer data. According to the invention, the influence on the 5G potential customer data quality is finally regulated and controlled according to the comparison result, the effect of improving the 5G potential customer data quality is achieved, and the problem that the 5G potential customer data quality is too poor in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular, to a method and system for predicting 5G potential customers for cross-marketing. Background Art

[0002] With the rapid development of 5G technology and the acceleration of the commercialization process, communication operators are facing unprecedented market opportunities and challenges. In order to seize the opportunity in the 5G market, operators need to accurately identify potential customers, formulate targeted marketing strategies, and improve customer experience and market share.

[0003] For example, a method and device for identifying potential customers based on transaction data disclosed in the invention patent with the publication number of CN115526661A includes: identifying the potential customers by using the customer types of existing customers who belong to different customers from the potential customers but have at least one transaction data in common with the potential customers.

[0004] For example, an analysis method for screening the initial customer group for potential high-quality Internet marketing communication disclosed in the invention patent with the publication number of CN115187300A includes: solving the problems that the traditional technology lacks a quantitative evaluation of the contribution degree of the initial customer group and thus cannot calculate and screen the initial customer group through a data model, as well as the problems of scattered resource investment, high cost, and low efficiency. The main solutions include recording the Internet marketing communication information and customer characteristic data of the customer group at the starting stage, storing all the data in a database; rating and grading the initial customer group from three aspects: the number of new customers attracted by the communication, the quality of the new customers attracted, and the quality of the initial customer group itself; performing variable screening on the grade score data of the initial customer group and the multi-dimensional customer characteristic data of the customer group, training and testing a logit regression model, and finally making a customer group screening score card; using the score card to score all the customer groups, and selecting the customer groups with the top ranking scores as the initial customer group for potential high-quality Internet marketing communication.

[0005] However, in the process of implementing the technical solutions of the present invention in the embodiments of the present application, it is found that the above technologies have at least the following technical problems:

[0006] In the prior art, predicting 5G potential customers requires collecting a large amount of relevant data, inputting these relevant data into a model or algorithm to predict 5G potential customers, and there is a problem of poor quality of 5G potential customer data. Summary of the Invention

[0007] The embodiments of the present application provide a method and system for predicting 5G potential customers for cross-marketing, solve the problem of poor quality of 5G potential customer data in the prior art, and achieve the effect of improving the quality of 5G potential customer data.

[0008] The embodiment of this application provides a method for predicting 5G potential customers in cross-marketing, including the following steps: collecting 5G potential customer data, relevant data affecting the timeliness of 5G potential customer data, and relevant data affecting the accuracy of 5G potential customer data, setting time rules and storage capacity thresholds for 5G potential customer data, and preprocessing the data according to the time rules and storage capacity thresholds; analyzing the relevant data affecting the timeliness of 5G potential customer data to obtain an evaluation coefficient for the timeliness of 5G potential customer data, analyzing the relevant data affecting the accuracy of 5G potential customer data to obtain an evaluation coefficient for the accuracy of 5G potential customer data, and comprehensively evaluating the evaluation coefficient for the timeliness of 5G potential customer data and the evaluation coefficient for the accuracy of 5G potential customer data to obtain an evaluation coefficient for the quality of 5G potential customer data; obtaining an evaluation threshold for the timeliness of 5G potential customer data and an evaluation threshold for the quality of 5G potential customer data from the database, comparing the evaluation coefficient for the timeliness of 5G potential customer data with the evaluation threshold for the timeliness of 5G potential customer data, and preliminarily regulating the timeliness of 5G potential customer data according to the comparison result, comparing the evaluation coefficient for the quality of 5G potential customer data with the evaluation threshold for the quality of 5G potential customer data, and finally regulating the quality of 5G potential customer data according to the comparison result; summarizing 5G potential customer data into a 5G potential customer data set, and obtaining a 5G potential customer list according to the 5G potential customer data set.

[0009] Further, the specific setting rules for setting time rules and storage capacity thresholds for 5G potential customer data are as follows: setting a time rule for timed clearing of 5G potential customer data, judging whether the 5G potential customer data has expired according to the time rule, if the 5G potential customer data has expired, performing a marking process on it, if the 5G potential customer data has not expired, comparing the capacity of 5G potential customer data with the storage capacity threshold; setting a storage capacity threshold for 5G potential customer data, comparing the capacity of 5G potential customer data with the storage capacity threshold, if the capacity of 5G potential customer data is greater than or equal to the storage capacity threshold, automatically triggering a data clearing operation, if the capacity of 5G potential customer data is less than the storage capacity threshold, continuously monitoring the capacity of 5G potential customer data.

[0010] Further, the specific analysis process for analyzing the relevant data affecting the timeliness of 5G potential customer data is as follows: the relevant data affecting the timeliness of 5G potential customer data includes the network delay of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data; analyzing the network delay of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data to obtain an evaluation coefficient for the timeliness of 5G potential customer data.

[0011] Further, the specific analysis process for analyzing the relevant data affecting the accuracy of 5G potential customer data is as follows: The relevant data for the accuracy of 5G potential customer data includes the incomplete rate of metadata records of 5G potential customer data, the data source marking ratio of 5G potential customer data, the field value inconsistency ratio of 5G potential customer data, the data backup frequency of 5G potential customer data, and the usage rate of standardized data formats for 5G potential customer data; analyze the incomplete rate of metadata records of 5G potential customer data, the data source marking ratio of 5G potential customer data, the field value inconsistency ratio of 5G potential customer data, the data backup frequency of 5G potential customer data, and the usage rate of standardized data formats for 5G potential customer data to obtain the evaluation coefficient for the accuracy of 5G potential customer data.

[0012] Further, the specific regulation process for initially regulating the timeliness of 5G potential customer data based on the comparison results is as follows: Compare the evaluation coefficient for the timeliness of 5G potential customer data with the evaluation threshold for the timeliness of 5G potential customer data. If the evaluation coefficient for the timeliness of 5G potential customer data is greater than or equal to the evaluation threshold for the timeliness of 5G potential customer data, monitor and track the entire process from data generation to its use by the model. If the evaluation coefficient for the timeliness of 5G potential customer data is less than the evaluation threshold for the timeliness of 5G potential customer data, refresh the data set regularly. For key data sources, implement a real-time or near-real-time data synchronization mechanism and use caching technology to store frequently accessed data.

[0013] Further, the specific regulation process for finally regulating the quality of 5G potential customer data based on the comparison results is as follows: Compare the evaluation coefficient for the quality of 5G potential customer data with the evaluation threshold for the quality of 5G potential customer data. If the evaluation coefficient for the quality of 5G potential customer data is greater than or equal to the evaluation threshold for the quality of 5G potential customer data, monitor data quality indicators and conduct data verification regularly. If the evaluation coefficient for the quality of 5G potential customer data is less than the evaluation threshold for the quality of 5G potential customer data, use statistical methods to detect outliers and missing values, sort the data in descending order, and use the average value of the front and back data to replace outliers or fill in missing values.

[0014] Further, the specific summarization process for summarizing 5G potential customer data into a 5G potential customer data set is as follows: 5G potential customer data includes customer Internet access duration, device replacement frequency, and call duration. Clean and denoise the customer Internet access duration, device replacement frequency, and call duration, summarize the cleaned and denoised data into a 5G potential customer data set, and store the 5G potential customer data set in a database.

[0015] Further, the specific process of obtaining the 5G potential customer list based on the output prediction values is as follows: perform weighted averaging on the output prediction values, directly obtain the 5G potential customer evaluation threshold from the database, aggregate all prediction values greater than the potential customer evaluation threshold into a preliminary 5G potential customer list, and remove duplicate customer records and customer records of customers who have already handled 5G services but have not been updated in the preliminary 5G potential customer list to obtain the 5G potential customer list.

[0016] Further, the specific method for obtaining the coefficient for evaluating the quality of 5G potential customer data is as follows: In the formula, DQ represents the coefficient for evaluating the quality of 5G potential customer data, which is used to evaluate the impact of the timeliness and accuracy of 5G potential customer data on the quality of 5G potential customer data. TL represents the coefficient for evaluating the timeliness of 5G potential customer data, α represents the weight factor of the coefficient for evaluating the timeliness of 5G potential customer data, AC represents the coefficient for evaluating the accuracy of 5G potential customer data, β represents the weight factor of the coefficient for evaluating the accuracy of 5G potential customer data, and e represents the natural constant.

[0017] The embodiment of the present application provides a 5G potential customer prediction system for cross-marketing, which includes a data collection module, a data analysis module, a threshold comparison module, and a prediction module:

[0018] The data collection module: is used to collect 5G potential customer data, relevant data affecting the timeliness of 5G potential customer data, and relevant data affecting the accuracy of 5G potential customer data, set time rules and storage capacity thresholds for the 5G potential customer data, and preprocess the data according to the time rules and storage capacity thresholds.

[0019] The data analysis module: is used to analyze the relevant data affecting the timeliness of 5G potential customer data to obtain the coefficient for evaluating the timeliness of 5G potential customer data, analyze the relevant data affecting the accuracy of 5G potential customer data to obtain the coefficient for evaluating the accuracy of 5G potential customer data, and comprehensively evaluate the coefficient for evaluating the timeliness of 5G potential customer data and the coefficient for evaluating the accuracy of 5G potential customer data to obtain the coefficient for evaluating the quality of 5G potential customer data.

[0020] Threshold comparison module: It is used to obtain the evaluation threshold for the timeliness of 5G potential customer data and the evaluation threshold for the quality of 5G potential customer data from the database, compare the evaluation coefficient for the timeliness of 5G potential customer data with the evaluation threshold for the timeliness of 5G potential customer data, and preliminarily regulate the timeliness of 5G potential customer data according to the comparison result. Then, compare the evaluation coefficient for the quality of 5G potential customer data with the evaluation threshold for the quality of 5G potential customer data, and finally regulate the quality of 5G potential customer data according to the comparison result.

[0021] Prediction module: It is used to aggregate 5G potential customer data into a 5G potential customer data set, and obtain a 5G potential customer list based on the 5G potential customer data set.

[0022] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0023] 1. By comparing the evaluation coefficient for the quality of 5G potential customer data with the evaluation threshold for the quality of 5G potential customer data, and finally regulating the quality of 5G potential customer data according to the comparison result, the effect of improving the quality of 5G potential customer data is achieved, and the problem of poor quality of 5G potential customer data in the prior art is effectively solved.

[0024] 2. By setting time rules and storage capacity thresholds for 5G potential customer data, and preprocessing the data according to the time rules and storage capacity thresholds, the accuracy of data acquisition is improved, and the problem of insufficient accuracy of data acquisition in the prior art is effectively solved.

[0025] 3. By directly obtaining the 5G potential customer evaluation threshold from the database, all prediction values greater than the potential customer evaluation threshold are aggregated into a preliminary 5G potential customer list, and duplicate customer records and customer records of customers who have already handled 5G services and have not been updated in the preliminary 5G potential customer list are removed to obtain a 5G potential customer list. The effect of improving marketing accuracy is achieved, and the problem of insufficient marketing accuracy in the prior art is effectively solved. Description of the Drawings

[0026] Figure 1 It is a flowchart of a method for predicting 5G potential customers for cross-marketing provided in the embodiments of the present application;

[0027] Figure 2 It is an image of the evaluation coefficient for the timeliness of 5G potential customer data in a method for predicting 5G potential customers for cross-marketing provided in the embodiments of the present application;

[0028] Figure 3Schematic structural diagram of a 5G potential customer prediction system for cross-marketing provided by an embodiment of the present application. Detailed implementation manners

[0029] By providing a method and system for predicting 5G potential customers for cross-marketing in an embodiment of the present application, the problem of too poor data quality of 5G potential customer data in the prior art is solved. By comparing the evaluation coefficient affecting the data quality of 5G potential customers with the evaluation threshold affecting the data quality of 5G potential customers, the data quality of 5G potential customers is finally regulated according to the comparison result, thereby achieving the effect of improving the data quality of 5G potential customers.

[0030] The technical solution in the embodiment of the present application is to solve the above problem of too poor data quality of 5G potential customer data, and the general idea is as follows:

[0031] By comparing the evaluation coefficient affecting the data quality of 5G potential customers with the evaluation threshold affecting the data quality of 5G potential customers, the data quality of 5G potential customers is finally regulated according to the comparison result, achieving the effect of improving the data quality of 5G potential customers.

[0032] To better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0033] As Figure 1As shown in the figure, it is a flowchart of a 5G potential customer prediction method for cross-marketing provided by an embodiment of the present application. The method includes the following steps: collecting 5G potential customer data, relevant data affecting the timeliness of 5G potential customer data, and relevant data affecting the accuracy of 5G potential customer data, setting time rules and storage capacity thresholds for the 5G potential customer data, and preprocessing the data according to the time rules and storage capacity thresholds; analyzing the relevant data affecting the timeliness of 5G potential customer data to obtain an evaluation coefficient for the timeliness of 5G potential customer data, analyzing the relevant data affecting the accuracy of 5G potential customer data to obtain an evaluation coefficient for the accuracy of 5G potential customer data, and comprehensively evaluating the evaluation coefficient for the timeliness of 5G potential customer data and the evaluation coefficient for the accuracy of 5G potential customer data to obtain an evaluation coefficient for the quality of 5G potential customer data; obtaining an evaluation threshold for the timeliness of 5G potential customer data and an evaluation threshold for the quality of 5G potential customer data from the database, comparing the evaluation coefficient for the timeliness of 5G potential customer data with the evaluation threshold for the timeliness of 5G potential customer data, and preliminarily regulating the timeliness of 5G potential customer data according to the comparison result, comparing the evaluation coefficient for the quality of 5G potential customer data with the evaluation threshold for the quality of 5G potential customer data, and finally regulating the quality of 5G potential customer data according to the comparison result; summarizing the 5G potential customer data into a 5G potential customer dataset, and obtaining a 5G potential customer list according to the 5G potential customer dataset.

[0034] In this embodiment, the process of obtaining the 5G potential customer list based on the 5G potential customer dataset is as follows: Collect 5G potential customer data and perform ETL integration processing on the 5G potential customer data, including data extraction, data cleaning, data transformation, and data loading. During the processing, ensure that data is obtained from the correct source, as well as the consistency of data format and definition, through methods such as data source verification, unified date format, unified numerical unit, and data type conversion. Take the customer's Internet access duration, device replacement frequency, and call duration of each customer as a sample in the 5G potential customer dataset, and take the multiple collected samples as the model sample dataset. Divide the model sample dataset into a training set and a validation set according to the ratio of x:y, where x ∈ [1, 9], y ∈ [1, 9], x and y are positive integers, and x + y = 10. Use the logistic regression statistical method to establish a logistic regression model, respectively use the LightGBM model, XGBoost model, and TextCNN model as the base classifiers, and use the training set for training. Then use the five-fold cross-validation method to train the base classifier model, and then use the validation set for verification. Input the 5G potential customer data and output the predicted values. Perform weighted averaging on the output predicted values, directly obtain the 5G potential customer evaluation threshold from the database, and summarize all the predicted values greater than the potential customer evaluation threshold into the 5G potential customer preliminary list. Remove the duplicate customer records and the customer records of the 5G services that have been processed but not updated in the 5G potential customer preliminary list to obtain the 5G potential customer list.

[0035] The logistic regression model has a base classifier model and a five-fold cross-validation function.

[0036] The specific process of using the logistic regression statistical method to establish a logistic regression model: Install the sci kit-learn library in python, directly call the logistic regression model, and use the cross_val_score function for five-fold cross-validation.

[0037] Furthermore, the specific setting rules for setting the time rule and storage capacity threshold for the 5G potential customer data are as follows: Set a time rule for timed clearing of the 5G potential customer data, and judge whether the 5G potential customer data has expired according to the time rule. If the 5G potential customer data has expired, perform a marking process on it. If the 5G potential customer data has not expired, perform a threshold comparison between the 5G potential customer data capacity and the storage capacity threshold; Set a storage capacity threshold for the 5G potential customer data, and compare the 5G potential customer data capacity with the storage capacity threshold. If the 5G potential customer data capacity is greater than or equal to the storage capacity threshold, automatically trigger a data clearing operation. If the 5G potential customer data capacity is less than the storage capacity threshold, continuously monitor the 5G potential customer data capacity.

[0038] In this embodiment, for example, the retention period of 5G potential customer data is set to 6 months. The system regularly (daily or weekly) checks the 5G potential customer data in the database. For each piece of data, the system checks the difference between the creation or last update time and the current time. If the data has exceeded the set retention period, the system marks it as "expired". For the data marked as "expired", the system sets an "unavailable" flag for these data in the data table and prepares for data deletion operations.

[0039] The capacity of 5G potential customer data refers to the storage space occupied by 5G potential customer data. The maximum threshold of the set database storage capacity is 80%. The system monitors in real time the storage space occupied by 5G potential customer data. If the 5G potential customer data capacity reaches or exceeds the storage capacity threshold, the system automatically triggers a data clearing operation and clears the data in chronological order.

[0040] By regularly clearing expired data, the database storage resources can be optimized, avoiding the problem of insufficient storage space caused by unlimited data growth. By reasonably managing data storage, the enterprise's investment in storage resources can be reduced and costs can be saved.

[0041] Furthermore, the specific analysis process for analyzing the relevant data affecting the timeliness of 5G potential customer data is as follows: The relevant data affecting the timeliness of 5G potential customer data includes the network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data. Analyze the network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data to obtain an evaluation coefficient for the timeliness of 5G potential customer data.

[0042] In this embodiment, the specific method for obtaining the evaluation coefficient for the timeliness of 5G potential customer data is as follows:

[0043]

[0044] In the formula, TL represents the evaluation coefficient for the timeliness of 5G potential customer data, which is used to evaluate the impact of the network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data on the timeliness of 5G potential customer data. Set several 5G potential customer data timeliness time monitoring points, and take 5 5G potential customer data timeliness time monitoring points as a 5G potential customer data timeliness time monitoring segment. X = 1, 2, 3,..., b, where b represents the total number of 5G potential customer data timeliness time monitoring segments, ND X represents the network latency of 5G potential customer data in the Xth 5G potential customer data timeliness time monitoring segment, and μ represents the weight factor of the network latency of 5G potential customer data, DBX The amount of data processed in each batch of 5G potential customer data, which is represented as the Xth 5G potential customer data timeliness time monitoring segment. ρ represents the weight factor of the amount of data processed in each batch of 5G potential customer data, DCF X The data cleaning frequency of 5G potential customer data, which is represented as the Xth 5G potential customer data timeliness time monitoring segment. τ represents the weight factor of the data cleaning frequency of 5G potential customer data, and e represents the natural constant.

[0045] If the network latency is high, network latency problems may be encountered during real-time or near-real-time data cleaning, resulting in untimely data cleaning. Increasing the data cleaning frequency will exacerbate the network latency problem because frequent data transmission will increase the network load. A large amount of data processed in each batch may lead to an increase in network latency. Appropriately adjusting the amount of data processed in each batch can help balance the network load and reduce the latency of data processing. If the data cleaning frequency is high, reduce the amount of data processed in each batch.

[0046] The weight factor of the network latency of 5G potential customer data, the weight factor of the amount of data processed in each batch of 5G potential customer data, and the weight factor of the data cleaning frequency of 5G potential customer data can be obtained from the dynamic information database, indicating the proportion of the network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data in the 5G potential customer data timeliness evaluation coefficient. The weight factor of the network latency of 5G potential customer data, the weight factor of the amount of data processed in each batch of 5G potential customer data, and the weight factor of the data cleaning frequency of 5G potential customer data can be obtained through a mapping relationship. For example, by establishing a mapping set of the network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data and their corresponding weights based on the relationship between the network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data in historical data and dynamic information, and obtaining the weight factor of the network latency of 5G potential customer data, the weight factor of the amount of data processed in each batch of 5G potential customer data, and the weight factor of the data cleaning frequency of 5G potential customer data corresponding to the mapping set by inputting the real-time network latency of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data.

[0047] In a specific embodiment, the data example of the 5G potential customer data timeliness evaluation coefficient is as follows in the table.

[0048] Table 1 Data example table of the 5G potential customer data timeliness evaluation coefficient

[0049]

[0050] When the weight factors of network latency, the amount of data processed in each batch, and the data cleaning frequency of 5G potential customer data are 0.2, 0.3, and 0.5 respectively, through Figure 2 and the data in Table 1, it can be seen that when the average value of the amount of data processed in each batch of 5G potential customer data in the b 5G potential customer data timeliness time monitoring segments and the average value of the data cleaning frequency of 5G potential customer data in the b 5G potential customer data timeliness time monitoring segments remain unchanged, the greater the average value of the network latency of 5G potential customer data in the b 5G potential customer data timeliness time monitoring segments, the smaller the impact on the 5G potential customer data timeliness evaluation coefficient.

[0051] Furthermore, the specific analysis process for analyzing the relevant data affecting the accuracy of 5G potential customer data is as follows: The relevant data affecting the accuracy of 5G potential customer data includes the incomplete rate of metadata records of 5G potential customer data, the data source marking ratio of 5G potential customer data, the field value inconsistency ratio of 5G potential customer data, the data backup frequency of 5G potential customer data, and the standardized data format usage rate of 5G potential customer data; Analyze the incomplete rate of metadata records of 5G potential customer data, the data source marking ratio of 5G potential customer data, the field value inconsistency ratio of 5G potential customer data, the data backup frequency of 5G potential customer data, and the standardized data format usage rate of 5G potential customer data to obtain the evaluation coefficient of the accuracy of 5G potential customer data.

[0052] In this embodiment, the specific method for obtaining the evaluation coefficient of the accuracy of 5G potential customer data is as follows:

[0053]

[0054] In the formula, AC represents the evaluation coefficient of the accuracy of 5G potential customer data, which is used to evaluate the impact of the incomplete rate of metadata records of 5G potential customer data, the data source marking ratio of 5G potential customer data, the field value inconsistency ratio of 5G potential customer data, the data backup frequency of 5G potential customer data, and the standardized data format usage rate of 5G potential customer data on the accuracy of 5G potential customer data. Set several 5G potential customer data accuracy time monitoring points, and take 5 5G potential customer data accuracy time monitoring points as a 5G potential customer data accuracy time monitoring segment. Z = 1, 2, 3,..., a, where a represents the total number of 5G potential customer data accuracy time monitoring segments, IMRR ZThe incomplete rate of metadata records of 5G potential customer data, denoted as the Z-th 5G potential customer data accuracy time monitoring segment, DSTR Z The data source marker ratio of 5G potential customer data, denoted as the Z-th 5G potential customer data accuracy time monitoring segment, FVIR Z The field value inconsistency ratio of 5G potential customer data, denoted as the Z-th 5G potential customer data accuracy time monitoring segment, DBF Z The data backup frequency of 5G potential customer data, denoted as the Z-th 5G potential customer data accuracy time monitoring segment The weight factor of the data backup frequency of 5G potential customer data, UROD Z The usage rate of the standardized data format of 5G potential customer data, denoted as the Z-th 5G potential customer data accuracy time monitoring segment. e represents the natural constant

[0055] The incomplete rate of metadata records refers to the proportion of incomplete data regarding data sources, structures, formats, and context information in the dataset to the total amount of data. The incomplete rate of metadata can be checked through data audits. Missing necessary metadata fields, existing metadata fields with empty or default values, metadata content inconsistent with or outdated from the actual data, and metadata not conforming to predefined data standards or formats are all recorded as incomplete metadata

[0056] The data source marker ratio refers to the proportion of data items with marked sources in the dataset and can be obtained by analyzing the dataset using data quality tools or custom scripts

[0057] The field value consistency ratio refers to the degree of consistency among field values in the dataset and can be obtained through data analysis tools or by writing scripts

[0058] The data backup frequency refers to the execution frequency of data backup operations and can be obtained by viewing the log records of data backups

[0059] The usage rate of the standardized data format refers to the proportion of data items using a unified standard format in the dataset and is obtained through data quality checks and analyzing the dataset

[0060] Due to the lack of sufficient metadata to correctly identify the data source, explicit data source tagging helps to track and complete the metadata. The higher the incomplete rate of metadata records, the lower the data source tagging ratio. Increasing the data source tagging ratio helps to identify and unify the field values of different data sources. The higher the data source tagging ratio, the lower the field value inconsistency ratio. Since unified data formats reduce errors and confusion during data integration, the higher the field value inconsistency ratio, the lower the usage rate of standardized data formats. Since data with a unified format is easier to back up and restore, the higher the usage rate of standardized data formats, the higher the data backup frequency.

[0061] The weight factor of the data backup frequency of 5G potential customer data can be obtained through a dynamic information database, indicating the proportion of the data backup frequency of 5G potential customer data in the 5G potential customer data accuracy evaluation coefficient. The weight factor of the data backup frequency of 5G potential customer data can be obtained through a mapping relationship. For example, a mapping set of the data backup frequency of 5G potential customer data and its corresponding weight is established through the relationship between the data backup frequency of 5G potential customer data in historical data and dynamic information, and the weight factor of the data backup frequency of 5G potential customer data corresponding to the input real-time data backup frequency of 5G potential customer data is obtained from the mapping set.

[0062] Furthermore, the specific regulation process for preliminarily regulating the timeliness of 5G potential customer data according to the comparison result is as follows: compare the 5G potential customer data timeliness evaluation coefficient with the 5G potential customer data timeliness evaluation threshold. If the 5G potential customer data timeliness evaluation coefficient is greater than or equal to the 5G potential customer data timeliness evaluation threshold, monitor and track the entire process from data generation to being used by the model. If the 5G potential customer data timeliness evaluation coefficient is less than the 5G potential customer data timeliness evaluation threshold, refresh the data set regularly. For key data sources, implement a real-time or near-real-time data synchronization mechanism and use caching technology to store frequently accessed data.

[0063] In this embodiment, monitor and track the entire life cycle of data from the source (such as user behavior logs, call records, etc.) to being finally used by the logistic regression model.

[0064] Regularly refreshing the dataset means formulating a data refresh plan, such as updating the dataset regularly every day or week to include the latest potential customer information. Implementing a real-time or near-real-time data synchronization mechanism means implementing a real-time data synchronization mechanism for key data sources, such as user real-time behavior data, to ensure the immediate update of data. Using message queues and data stream processing technologies to achieve fast data transmission and synchronization. Caching technology means using in-memory caching technology to store frequently accessed data, such as popular product information or user preferences, reducing direct access to the database and improving data reading speed. Regularly refreshing the dataset and real-time synchronization mechanism help optimize the use of storage resources and avoid storing outdated data.

[0065] Furthermore, the specific regulation process for finally regulating the quality of 5G potential customer data according to the comparison result is as follows: Compare the evaluation coefficient of the quality of 5G potential customer data with the evaluation threshold of the quality of 5G potential customer data. If the evaluation coefficient of the quality of 5G potential customer data is greater than or equal to the evaluation threshold of the quality of 5G potential customer data, monitor the data quality indicators and perform data verification regularly. If the evaluation coefficient of the quality of 5G potential customer data is less than the evaluation threshold of the quality of 5G potential customer data, use statistical methods to detect outliers and missing values, sort the data in descending order, and use the average value of the front and back data to replace outliers or fill in missing values.

[0066] In this embodiment, measure the data quality indicators, identify problems in the data, correct, standardize, and transform the data, convert the data from the original format to the standard format, and monitor the data quality in real-time or regularly. Through automated data quality control and verification, resources can be utilized more effectively and unnecessary human input can be reduced.

[0067] Use statistical methods to detect outliers and missing values. The statistical method refers to the Z-score statistical method. The specific steps are as follows: The Z-score standardizes the data so that each data point is expressed as a multiple of the standard deviation from the mean. Values with an absolute Z-score greater than 2 are considered outliers, and other values are judged as normal values. Sort the data in descending order, and use the average value of the front and back data to replace outliers or fill in missing values. By processing outliers and missing values, the quality and usability of the data are improved.

[0068] Furthermore, the specific summarization process for summarizing 5G potential customer data into a 5G potential customer dataset is as follows: 5G potential customer data includes customer Internet access duration, device replacement frequency, and call duration. Clean and denoise the customer Internet access duration, device replacement frequency, and call duration, and summarize the cleaned and denoised data into a 5G potential customer dataset, and store the 5G potential customer dataset in the database.

[0069] In this embodiment, data such as customer Internet access duration, device replacement frequency, and call duration are collected from sources such as the operator system, customer service records, and sales databases. The cleaned Internet access duration, device replacement frequency, and call duration data are merged to form a 5G potential customer dataset. The 5G potential customer dataset is stored in a secure database, such as a relational database or a big data platform, and database indexes are set for quick querying and analysis. Storing the data in the database enables the implementation of access control and backup strategies to protect customer data from unauthorized access or loss.

[0070] Further, the specific process of obtaining the 5G potential customer list based on the output prediction values is as follows: The output prediction values are weighted and averaged, and the 5G potential customer evaluation threshold is directly obtained from the database. All prediction values greater than the potential customer evaluation threshold are aggregated into a preliminary 5G potential customer list, and duplicate customer records and customer records of customers who have already subscribed to 5G services but have not been updated in the preliminary 5G potential customer list are removed to obtain the 5G potential customer list.

[0071] In this embodiment, a logistic regression model is used to score the 5G potential customer data to obtain prediction values. These prediction values are weighted and averaged, and the pre-set 5G potential customer evaluation threshold is directly obtained from the database. According to the evaluation threshold, customer records with prediction values greater than the threshold are screened out. The customer records with prediction values greater than the threshold are aggregated into a preliminary 5G potential customer list, and duplicate customer records and customer records of customers who have already subscribed to 5G services but have not been updated in the preliminary 5G potential customer list are removed to obtain the 5G potential customer list, ensuring the accuracy and integrity of customer information, avoiding multiple marketing campaigns for the same customer, and targeting customers with high prediction values for marketing can improve the conversion rate of marketing activities.

[0072] Further, the specific method for obtaining the coefficient for evaluating the quality of 5G potential customer data is as follows: In the formula, DQ represents the coefficient for evaluating the quality of 5G potential customer data, which is used to evaluate the impact of the timeliness and accuracy of 5G potential customer data on the quality of 5G potential customer data. TL represents the coefficient for evaluating the timeliness of 5G potential customer data, α represents the weight factor of the coefficient for evaluating the timeliness of 5G potential customer data, AC represents the coefficient for evaluating the accuracy of 5G potential customer data, β represents the weight factor of the coefficient for evaluating the accuracy of 5G potential customer data, and e represents the natural constant.

[0073] In this embodiment, the weight factors affecting the timeliness evaluation coefficient of 5G potential customer data and the weight factors affecting the accuracy evaluation coefficient of 5G potential customer data can be obtained from the dynamic information database, representing the proportions of the timeliness evaluation coefficient of 5G potential customer data and the accuracy evaluation coefficient of 5G potential customer data in the quality evaluation coefficient of 5G potential customer data. The weight factors affecting the timeliness evaluation coefficient of 5G potential customer data and the weight factors affecting the accuracy evaluation coefficient of 5G potential customer data can also be obtained through mapping relationships. For example, mapping sets of the timeliness evaluation coefficient of 5G potential customer data and the accuracy evaluation coefficient of 5G potential customer data and their corresponding weights are established respectively based on the relationships between the timeliness evaluation coefficient of 5G potential customer data and the accuracy evaluation coefficient of 5G potential customer data and the dynamic information in historical data. The weight factors affecting the timeliness evaluation coefficient of 5G potential customer data and the weight factors affecting the accuracy evaluation coefficient of 5G potential customer data corresponding to them in the mapping set are obtained by inputting the real-time timeliness evaluation coefficient of 5G potential customer data and the accuracy evaluation coefficient of 5G potential customer data.

[0074] As Figure 3 shown, it is a schematic structural diagram of a 5G potential customer prediction system for cross-marketing provided by an embodiment of the present application. A 5G potential customer prediction system for cross-marketing provided by an embodiment of the present application includes: a data collection module, a data analysis module, a threshold comparison module, and a prediction module:

[0075] The data collection module: is used to collect 5G potential customer data, relevant data affecting the timeliness of 5G potential customer data, and relevant data affecting the accuracy of 5G potential customer data, set time rules and storage capacity thresholds for the 5G potential customer data, and preprocess the data according to the time rules and storage capacity thresholds;

[0076] The data analysis module: is used to analyze the relevant data affecting the timeliness of 5G potential customer data to obtain the timeliness evaluation coefficient of 5G potential customer data, analyze the relevant data affecting the accuracy of 5G potential customer data to obtain the accuracy evaluation coefficient of 5G potential customer data, and comprehensively evaluate the timeliness evaluation coefficient of 5G potential customer data and the accuracy evaluation coefficient of 5G potential customer data to obtain the quality evaluation coefficient of 5G potential customer data;

[0077] Threshold comparison module: It is used to obtain the evaluation threshold for the timeliness of 5G potential customer data and the evaluation threshold for the quality of 5G potential customer data from the database, compare the evaluation coefficient for the timeliness of 5G potential customer data with the evaluation threshold for the timeliness of 5G potential customer data, and perform preliminary regulation on the timeliness of 5G potential customer data according to the comparison result. Compare the evaluation coefficient for the quality of 5G potential customer data with the evaluation threshold for the quality of 5G potential customer data, and perform final regulation on the quality of 5G potential customer data according to the comparison result;

[0078] Prediction module: It is used to summarize 5G potential customer data into a 5G potential customer data set and obtain a 5G potential customer list based on the 5G potential customer data set.

[0079] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0080] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0081] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0082] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus, causing a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 steps for implementing the functions specified in one block or a plurality of blocks.

[0083] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to cover the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0084] Obviously, those skilled in the art can make various changes and deformations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and deformations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and deformations.

Claims

1. A cross-marketing 5G potential customer prediction method, characterized in that: The following steps are involved: Collect 5G potential customer data, relevant data that affects the timeliness of 5G potential customer data, and relevant data that affects the accuracy of 5G potential customer data, set time rules and storage capacity thresholds for 5G potential customer data, and pre-process the data according to the time rules and storage capacity thresholds; Analyze the relevant data that affect the timeliness of 5G potential customer data to obtain the evaluation coefficient that affects the timeliness of 5G potential customer data; analyze the relevant data that affect the accuracy of 5G potential customer data to obtain the evaluation coefficient that affects the accuracy of 5G potential customer data; conduct a comprehensive evaluation on the evaluation coefficient that affects the timeliness of 5G potential customer data and the evaluation coefficient that affects the accuracy of 5G potential customer data to obtain the evaluation coefficient that affects the quality of 5G potential customer data; Obtain the timeliness assessment threshold and quality assessment threshold of 5G potential customer data from the database, compare the timeliness assessment coefficient of 5G potential customer data with the timeliness assessment threshold, and make preliminary adjustments to the timeliness of 5G potential customer data based on the comparison results; compare the quality assessment coefficient of 5G potential customer data with the quality assessment threshold, and make final adjustments to the quality of 5G potential customer data based on the comparison results; The 5G potential customer data is aggregated into a 5G potential customer data set, and a 5G potential customer list is obtained based on the 5G potential customer data set.

2. A 5G potential customer prediction method for cross-marketing as claimed in claim 1, characterized in that: The specific setting rules for setting the time rule and storage capacity threshold for 5G potential customer data are as follows: A time rule for scheduled clearing of 5G potential customer data is set, and whether the 5G potential customer data is expired is determined according to the time rule. If the 5G potential customer data is expired, it is marked; if the 5G potential customer data is not expired, a threshold comparison is performed between the 5G potential customer data capacity and the storage capacity threshold; A storage capacity threshold is set for 5G potential customer data, and the 5G potential customer data capacity is compared with the storage capacity threshold. If the 5G potential customer data capacity is greater than or equal to the storage capacity threshold, the data clearing operation is automatically triggered; if the 5G potential customer data capacity is less than the storage capacity threshold, the 5G potential customer data capacity is continuously monitored.

3. A 5G potential customer prediction method for cross-marketing as claimed in claim 1, characterized in that: The specific analysis process of analyzing the relevant data that affects the timeliness of 5G potential customer data is as follows: The relevant data that affects the timeliness of 5G potential customer data include the network delay of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the frequency of data cleaning of 5G potential customer data; The network delay of 5G potential customer data, the amount of data processed in each batch of 5G potential customer data, and the data cleaning frequency of 5G potential customer data are analyzed, and the evaluation coefficients affecting the timeliness of 5G potential customer data are obtained.

4. A 5G potential customer prediction method for cross-marketing as claimed in claim 1, characterized in that: The specific analysis process of analyzing the relevant data that affects the accuracy of 5G potential customer data is as follows: Data related to the accuracy of 5G potential customer data include the rate of incomplete metadata records of 5G potential customer data, the rate of data source labeling of 5G potential customer data, the rate of field value inconsistency of 5G potential customer data, the data backup frequency of 5G potential customer data, and the use rate of standardized data formats for 5G potential customer data; The incomplete metadata record rate of 5G potential customer data, the data source labeling ratio of 5G potential customer data, the field value inconsistency ratio of 5G potential customer data, the data backup frequency of 5G potential customer data and the usage rate of standardized data format of 5G potential customer data were analyzed to obtain the evaluation coefficients affecting the accuracy of 5G potential customer data.

5. A cross-marketing 5G potential customer prediction method as claimed in claim 1, characterized in that: The specific regulation process of performing preliminary regulation on the timeliness of data affecting 5G potential customers based on the comparison results is as follows: The coefficient affecting the timeliness assessment of 5G potential customer data is compared with the threshold for affecting the timeliness assessment of 5G potential customer data. If the coefficient affecting the timeliness assessment of 5G potential customer data is greater than or equal to the threshold for affecting the timeliness assessment of 5G potential customer data, the entire process from data generation to data use by the model is monitored and tracked. If the coefficient affecting the timeliness assessment of 5G potential customer data is less than the threshold for affecting the timeliness assessment of 5G potential customer data, the data set is refreshed regularly, and for key data sources, a real-time or near real-time data synchronization mechanism is implemented, and cache technology is used to store frequently accessed data.

6. A 5G potential customer prediction method for cross-marketing as claimed in claim 1, characterized in that: The specific regulation process of finally regulating the quality of data affecting 5G potential customers according to the comparison results is as follows: The coefficient affecting the quality assessment of 5G potential customer data will be compared with the threshold for affecting the quality assessment of 5G potential customer data. If the coefficient affecting the quality assessment of 5G potential customer data is greater than or equal to the threshold for affecting the quality assessment of 5G potential customer data, monitor the data quality indicators and perform data verification regularly. If the coefficient affecting the quality assessment of 5G potential customer data is less than the threshold for affecting the quality assessment of 5G potential customer data, use statistical methods to detect outliers and missing values, arrange the data in descending order, and use the average of the before and after data to replace outliers or fill missing values.

7. A 5G potential customer prediction method for cross-marketing as claimed in claim 1, characterized in that: The specific aggregation process of aggregating the 5G potential customer data into the 5G potential customer data set is as follows: The 5G potential customer data includes the customer's Internet access time, device replacement frequency and call duration. The customer's Internet access time, device replacement frequency and call duration are cleaned and denoised, and the cleaned and denoised data are aggregated into a 5G potential customer data set, and the 5G potential customer data set is stored in the database.

8. A cross-marketing 5G potential customer prediction method as claimed in claim 1, characterized in that: The specific process of obtaining the 5G potential customer list according to the output prediction value is as follows: The output prediction values ​​are weighted averaged, the 5G potential customer evaluation threshold is directly obtained from the database, all prediction values ​​greater than the potential customer evaluation threshold are summarized into a preliminary list of 5G potential customers, and duplicate customer records and customer records that have not been updated but have handled 5G services in the preliminary list of 5G potential customers are removed to obtain a list of 5G potential customers.

9. A cross-marketing 5G potential customer prediction method as claimed in claim 1, characterized in that: The specific method for obtaining the coefficient affecting the quality assessment of 5G potential customer data is as follows: In the formula, DQ is expressed as the evaluation coefficient affecting the quality of 5G potential customer data, which is used to evaluate the impact of the timeliness and accuracy of 5G potential customer data on the quality of 5G potential customer data, TL is expressed as the evaluation coefficient affecting the timeliness of 5G potential customer data, α is expressed as the weight factor affecting the timeliness evaluation coefficient of 5G potential customer data, AC is expressed as the evaluation coefficient affecting the accuracy of 5G potential customer data, β is expressed as the weight factor affecting the accuracy evaluation coefficient of 5G potential customer data, and e is expressed as a natural constant.

10. A 5G potential customer prediction system for cross-marketing, characterized in that: Including data collection module, data analysis module, threshold comparison module and prediction module: Data collection module: used to collect 5G potential customer data, relevant data affecting the timeliness of 5G potential customer data, and relevant data affecting the accuracy of 5G potential customer data, set time rules and storage capacity thresholds for 5G potential customer data, and pre-process the data according to the time rules and storage capacity thresholds; Data analysis module: used to analyze relevant data affecting the timeliness of 5G potential customer data to obtain the evaluation coefficient affecting the timeliness of 5G potential customer data, analyze relevant data affecting the accuracy of 5G potential customer data to obtain the evaluation coefficient affecting the accuracy of 5G potential customer data, conduct a comprehensive evaluation on the evaluation coefficient affecting the timeliness of 5G potential customer data and the evaluation coefficient affecting the accuracy of 5G potential customer data to obtain the evaluation coefficient affecting the quality of 5G potential customer data; Threshold comparison module: used to obtain the timeliness assessment threshold and quality assessment threshold of 5G potential customer data from the database, compare the timeliness assessment coefficient of 5G potential customer data with the timeliness assessment threshold of 5G potential customer data, and make preliminary adjustments to the timeliness of 5G potential customer data according to the comparison results, compare the quality assessment coefficient of 5G potential customer data with the quality assessment threshold of 5G potential customer data, and make final adjustments to the quality of 5G potential customer data according to the comparison results; Prediction module: used to aggregate 5G potential customer data into a 5G potential customer data set, and obtain a 5G potential customer list based on the 5G potential customer data set.

Images