Intelligent marketing distribution system and method based on deep learning and related equipment

Abstract

The invention discloses an intelligent marketing distribution system and method based on deep learning and related equipment. The method comprises the following steps: acquiring a channel error code and short message subsequent data; performing data statistics and analysis on the obtained error codes of the corresponding channels, storing the error codes in a database, and storing short message subsequent data in the database; automatically classifying each channel error code through a deep machine learning algorithm, and obtaining mark information of each channel error code; associating each channel error code with the error code classification type and taking the associated channel error code as channel error code classification information; based on the short message subsequent data, obtaining a recommended channel group and the number of channels in the recommended channel group through data analysis and mining; and distributing the short message marketing task according to the channel weight of the recommended channel group, and if distribution fails, automatically triggering a corresponding personalized strategy mechanism according to the channel error code classification information. According to the method, a personalized solution for the error is provided for the user who fails to reach the channel, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the fields of data processing and computer technology, and in particular to a deep learning-based intelligent marketing distribution system, method, and related equipment. Background Art

[0002] With the rapid development of artificial intelligence (AI) technology, the market for intelligent marketing distribution systems is also expanding. These systems primarily analyze and filter data at the customer and individual channel levels to achieve intelligent marketing. However, for users whose channels fail to reach them, these systems lack personalized solutions, fail to further explore the value of data resources, and thus fail to improve business profitability.

[0003] Based on this, this application provides an intelligent marketing distribution system, method and related equipment based on deep learning to solve the above technical problems. Summary of the Invention

[0004] The present invention provides an intelligent marketing distribution system, method and related equipment based on deep learning, which are used to solve the problem that for users who fail to be reached through channels, personalized solutions cannot be provided for the error and the value of data resources cannot be further mined.

[0005] To achieve the above objectives, this application is implemented through the following technical solutions:

[0006] This application provides an intelligent marketing distribution system based on deep learning, which includes:

[0007] A data collection module, the data collection module is used to obtain channel error codes and SMS post-process data; perform data statistics and analysis on the obtained multiple channel error codes and save them to a database, and save the SMS post-process data to the database;

[0008] An error code identification and classification module, which is used to automatically classify each of the channel error codes through a deep machine learning algorithm, obtain tag information of each of the channel error codes, and associate each of the channel error codes with an error code classification type as channel error code classification information;

[0009] A channel recommendation module, configured to obtain a recommended channel group and the number of channels in the recommended channel group through data analysis and mining based on the SMS post-process data;

[0010] An intelligent distribution module is used to distribute SMS marketing tasks for each marketing batch according to the channel weight of the recommended channel group. If the distribution fails, the corresponding personalized strategy mechanism is automatically triggered according to the channel error code classification information. The personalized strategy mechanism is used to adopt different SMS data retransmission processing strategies for channel error codes of different error code classification types.

[0011] In one embodiment, the channel recommendation module is further configured to obtain a salvage channel group and the number of channels in the salvage channel group through data analysis and mining based on the SMS post-process data in the database;

[0012] The intelligent distribution module is also used to execute the SMS data retransmission processing strategy. If the SMS data enters the channel of the retrieval channel group, it will determine whether to continue to distribute it to the channel in the retrieval channel group based on the actual retrieval times and the maximum preset retrieval times.

[0013] In one embodiment, the data collection module includes:

[0014] A first collection unit, which is used to collect error codes generated in real time through the Kafka processing platform and ES during the SMS data sending process and use them as channel error codes;

[0015] A second collecting unit, the second collecting unit is used to collect SMS post-process data;

[0016] A first preprocessing unit, configured to perform data preprocessing on the channel error code, extracting field information, performing internal data conversion, and removing invalid information, and storing the SMS post-process data in a Redis cluster;

[0017] The data statistics and analysis unit is used to perform data statistics and analysis on the channel error code after data preprocessing and store it in the channel error information record table of the database; it is also used to store the SMS post-process data stored in the Redis cluster in the database.

[0018] In one embodiment, the error code identification and classification module includes:

[0019] a rule table establishing unit, configured to perform statistics and analysis on error codes according to the channel error information record table, and obtain a rule table including error code classification rules;

[0020] An information marking unit, configured to automatically mark the channel error codes returned by each channel according to the error code classification rule using a decision tree algorithm or a support vector machine (SVM) algorithm to obtain marking information;

[0021] a relationship establishing unit, configured to identify and establish a record relationship between each channel error code and an error code classification type according to the tag information, thereby forming channel error code classification information;

[0022] A record relationship maintenance unit is used to maintain and manage the channel error code classification information.

[0023] In one embodiment, the channel recommendation module includes:

[0024] A data analysis and mining unit, which is used to analyze and mine the SMS post-process data using Hadoop technology to obtain an indicator report, which is used to provide different dimensions of SMS marketing tasks;

[0025] A scoring calculation unit is used to calculate the scores of each channel in different dimensions of the indicator report based on a business algorithm, including project click cost, project click rate, project gross profit ratio, price, and project success rate; and calculate the final channel score based on the weight of each dimension; wherein the weight of each dimension is obtained based on actual business conditions;

[0026] The channel group acquisition unit is used to sort the final channel scores, acquire the recommended channel group and the number of channels in the recommended channel group, and acquire the salvaged channel group and the number of channels in the salvaged channel group.

[0027] In one embodiment, the intelligent distribution module includes:

[0028] a configuration unit, configured to obtain corresponding processing strategies and retransmission strategies according to error code classification types of the channel error codes;

[0029] A second pre-processing unit, which is used to perform verification pre-processing on the SMS marketing task to be distributed, including filtering blacklist, frequency, and encryption and decryption;

[0030] a marketing distribution unit, configured to send SMS data via the recommended channel group, wherein the number of numbers sent by each channel in the recommended channel group is calculated based on a channel weight;

[0031] A retransmission judgment unit is used to automatically trigger a corresponding personalized strategy mechanism based on the channel error code returned by the channel that failed to send and the classification information of the channel error code. The processing strategies include immediate retransmission, delayed retransmission, no retransmission, immediate retry, delayed retry and no retry. When the processing strategy is retransmission, the SMS data is returned to the channel in the retrieval channel group for forwarding. If a channel in the retrieval channel group fails to send, it is determined whether the processing strategy of the error code returned by the failed channel is retransmission. If it is retransmission, the forwarding is continued using other channels in the retrieval channel group. If it is not retransmission, the forwarding is stopped.

[0032] In one embodiment, the intelligent distribution module further includes:

[0033] The information return unit is used to receive the error codes returned by each channel in real time and save the error codes to the ES.

[0034] This application also provides an intelligent marketing distribution method based on deep learning error code classification, the intelligent marketing distribution method comprising:

[0035] Obtaining channel error codes and SMS post-process data; performing data statistics and analysis on the obtained multiple channel error codes and saving them to a database, and saving the SMS post-process data to the database;

[0036] Automatically classify each of the channel error codes using a deep machine learning algorithm to obtain tag information for each of the channel error codes; associate each of the channel error codes with an error code classification type and use the associated information as channel error code classification information;

[0037] Based on the SMS post-process data, obtaining a recommended channel group and the number of channels in the recommended channel group through data analysis and mining;

[0038] For each marketing batch, the SMS marketing task is distributed according to the channel weight of the recommended channel group. If the distribution fails, the corresponding personalized strategy mechanism is automatically triggered according to the channel error code classification information. The personalized strategy mechanism is used to adopt different SMS data retransmission processing strategies for channel error codes of different error code classification types.

[0039] The present application also provides an electronic device, which includes a memory and at least one processor, wherein the memory stores a computer program, and the at least one processor is configured to implement the steps of the method described above when executing the computer program.

[0040] The present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, the steps of the method described above are implemented.

[0041] The technical solution provided by this application provides the following beneficial effects: First, channel error codes and SMS post-process data are collected. This data includes various error codes encountered during SMS sending and user responses to SMS messages. Through data statistics and analysis, the error codes are categorized and stored in a database. The SMS post-process data is also stored for subsequent analysis and decision support.

[0042] Then, using a deep machine learning algorithm, the collected channel error codes are automatically categorized and each error code is assigned a tag. This tag information includes the characteristics of each error code and is associated with the corresponding error code classification type to form channel error code classification information. Based on SMS post-process data analysis, a recommended channel group and the number of channels within it are determined. This recommended channel group is selected based on performance and effectiveness data to improve the success rate of SMS marketing.

[0043] Finally, for each marketing batch, SMS marketing tasks are allocated based on the weight of each channel in the recommended channel group. If delivery fails, a personalized policy mechanism is automatically triggered based on the channel error code classification information. This personalized policy mechanism can adopt different resend strategies based on the error code type, such as immediate resend, delayed resend, or abandoned resend.

[0044] This provides personalized solutions for users whose channels fail to reach them. Through personalized policy mechanisms, appropriate action can be taken for different error scenarios, reducing the number of failed or delayed SMS messages received by users and improving the user experience. The collection of channel error codes and SMS post-process data further unlocks the value of data resources. Intelligent distribution and error handling reduce duplicate messages caused by errors, lowering SMS marketing costs. Automated error code classification and intelligent distribution mechanisms reduce manual intervention and improve processing speed and efficiency. The application of deep machine learning algorithms improves the accuracy of error code classification and enables more precise error handling. Allocating SMS tasks based on channel weight optimizes resource utilization and increases the return on investment of marketing campaigns.

[0045] Other features and advantages disclosed in this embodiment will be described in the following description. Alternatively, some features and advantages may be inferred or unambiguously determined from the description, or may be learned by practicing the above-mentioned techniques of this disclosure. To make the above-mentioned objectives, features, and advantages of this disclosure more clearly understood, the following describes a detailed description of preferred embodiments in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present application is further described below with reference to the accompanying drawings and examples.

[0047] Figure 1 A flow chart of an intelligent marketing distribution method provided in an embodiment of the present application is shown.

[0048] Figure 2 A flow chart of data collection, analysis and statistics in the intelligent marketing distribution process provided by an embodiment of the present application is shown.

[0049] Figure 3 A flowchart of error code identification and classification in the intelligent marketing distribution process provided by an embodiment of the present application is shown.

[0050] Figure 4 A flow chart of channel recommendation in the intelligent marketing distribution process provided by an embodiment of the present application is shown.

[0051] Figure 5 A flow chart of intelligent distribution in the intelligent marketing distribution process provided in an embodiment of the present application is shown.

[0052] Figure 6 A structural diagram of the intelligent marketing distribution system provided in an embodiment of the present application is shown.

[0053] Figure 7 A structural diagram of the data collection module provided in an embodiment of the present application is shown.

[0054] Figure 8 A structural diagram of the error code identification and classification module provided in an embodiment of the present application is shown.

[0055] Figure 9 A schematic diagram of the structure of the channel recommendation module provided in an embodiment of the present application is shown.

[0056] Figure 10 A structural diagram of the intelligent distribution module provided in an embodiment of the present application is shown.

[0057] Figure 11 The figure shows a structural block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate preferred embodiments of the present invention. However, the present invention may be implemented in many other forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present disclosure.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0060] The following is a brief description of the technical field and related terms of the embodiments of the present application.

[0061] Intelligent marketing distribution systems have been widely adopted in the international market. Among related technologies, these systems have been applied in a variety of fields, including e-commerce, finance, education, and healthcare. The technologies behind these systems are constantly innovating, including machine learning, natural language processing, and knowledge graphs. The application of these technologies enables intelligent marketing distribution systems to more accurately analyze user behavior and preferences, enabling more targeted marketing. Furthermore, as the intelligence level of intelligent marketing distribution systems continues to increase, they can provide businesses and users with more intelligent and personalized marketing solutions.

[0062] Machine Learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. Given a set of tasks T and a performance metric P, a computer program can learn from experience E. If its performance in task T can be measured by P, then it improves with experience E. Machine learning specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of artificial intelligence.

[0063] Deep learning is a special type of machine learning that achieves tremendous power and flexibility by learning to represent the world using nested hierarchies of concepts, where each concept is defined in relation to simpler concepts, and more abstract representations are computed using less abstract methods. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning through demonstration.

[0064] This application provides an intelligent marketing distribution system, method and related equipment based on deep learning. The following will first describe the method implementation example, and then describe the system, etc.

[0065] Method embodiment.

[0066] See also Figure 1 , Figure 1A flow chart of an intelligent marketing distribution method provided in an embodiment of the present application is shown.

[0067] This embodiment provides an intelligent marketing distribution method based on deep learning, which includes:

[0068] S101, obtaining channel error codes and SMS post-process data; performing data statistics and analysis on the obtained multiple channel error codes and saving them to a database, and saving the SMS post-process data to the database;

[0069] S102, automatically classifying each of the channel error codes using a deep machine learning algorithm to obtain tag information of each of the channel error codes; associating each of the channel error codes with an error code classification type and using the associated information as channel error code classification information;

[0070] S103, obtaining a recommended channel group and the number of channels in the recommended channel group through data analysis and mining based on the SMS post-process data;

[0071] S104, for each marketing batch, distributes SMS marketing tasks according to the channel weight of the recommended channel group. If the distribution fails, the corresponding personalized strategy mechanism is automatically triggered according to the channel error code classification information. The personalized strategy mechanism is used to adopt different SMS data retransmission processing strategies for channel error codes of different error code classification types.

[0072] This technical solution first collects channel error codes and SMS post-process data. This data includes various error codes encountered during SMS sending and user responses to SMS messages. Multiple error codes from the corresponding channels are collected and analyzed, categorized, and stored in a database. SMS post-process data is also stored for subsequent analysis and decision support.

[0073] Then, using a deep machine learning algorithm, the collected channel error codes are automatically categorized and each error code is assigned a tag. This tag information includes the characteristics of each error code and is associated with the corresponding error code classification type to form channel error code classification information. Based on SMS post-process data analysis, a recommended channel group and the number of channels within it are determined. This recommended channel group is selected based on performance and effectiveness data to improve the success rate of SMS marketing.

[0074] Finally, for each marketing batch, SMS marketing tasks are assigned based on the weight of each channel in the recommended channel group. If delivery fails, the personalized strategy mechanism is automatically triggered based on the channel error code classification information. The personalized strategy mechanism can adopt different retransmission strategies based on the type of error code, such as immediate retransmission, delayed retransmission, or abandoned retransmission. Channel weights can be pre-acquired and pre-determined based on historical data analysis (the performance of each channel in past marketing campaigns). Performance includes key performance indicators such as delivery rate, open rate, click-through rate, and conversion rate.

[0075] The beneficial effects of this embodiment include: reduced manual intervention and improved processing speed and efficiency through automated error code classification and intelligent distribution mechanisms. The application of deep machine learning algorithms improves the accuracy of error code classification, making error handling more precise. Allocating SMS tasks based on channel weights optimizes resource utilization and improves the return on investment of marketing activities. Through personalized policy mechanisms, appropriate handling measures can be taken for different error situations, reducing the number of failed or delayed SMS messages received by users and improving the user experience. Intelligent distribution and error handling reduce duplicate sending due to errors, reducing the cost of SMS marketing.

[0076] In summary, the intelligent marketing distribution method realizes the automation and intelligence of marketing distribution by combining deep learning and data analysis technology, improves efficiency and effectiveness, and reduces costs and risks.

[0077] In one embodiment, S103 further includes: obtaining a salvage channel group and the number of channels in the salvage channel group through data analysis and mining based on the SMS post-process data in the database;

[0078] S104 also includes: when executing the SMS data retransmission processing strategy, if the SMS data enters a channel of the retrieval channel group, determining whether to continue to distribute it to the channel in the retrieval channel group according to the actual retrieval times and the maximum preset retrieval times.

[0079] Determine the number of channels to include in the salvage channel group to ensure efficient message retransmission in subsequent steps. If the initial delivery of SMS data fails and the personalized policy mechanism is executed, if the SMS data requires further processing, it is transferred to the salvage channel group. The actual number of salvage attempts for each retransmission is recorded and compared with the preset maximum number of salvage attempts. Based on this comparison, a decision is made as to whether to continue message delivery to the channels in the salvage channel group. If the actual number of salvage attempts does not exceed the preset maximum number of salvage attempts, delivery attempts will continue. If the maximum number of salvage attempts has been reached or exceeded, delivery will cease, avoiding unnecessary resource waste.

[0080] The beneficial effects of this embodiment include: intelligently identifying the retrieval channel group enables secondary processing of SMS messages that fail to be delivered, increasing the chances of successful delivery. By controlling the number of retrievals, resource allocation is optimized, ensuring maximum marketing effectiveness within limited resources. A preset maximum number of retrievals serves as a safety mechanism, preventing resource waste caused by unlimited retransmissions. Through intelligent distribution and retrieval mechanisms, a balance between cost and efficiency is achieved, maximizing cost-effectiveness.

[0081] See also Figure 2 , Figure 2 The following is a flow chart showing the process of data collection, analysis and statistics in the intelligent marketing distribution process provided by an embodiment of the present application. In one embodiment, S101 includes:

[0082] S201: During the SMS data sending process, the error codes generated are collected in real time by the Kafka processing platform and ES and used as channel error codes;

[0083] S202, collecting SMS post-process data;

[0084] S203, performing data preprocessing on the channel error code, extracting field information, performing internal data conversion, and removing invalid information, and storing the SMS post-process data in a Redis cluster;

[0085] S204, performing data statistics and analysis on the channel error code after data preprocessing and storing the data in the channel error information record table of the database; and storing the SMS post-process data stored in the Redis cluster in the database.

[0086] This means that during SMS data transmission, error codes generated are monitored and collected in real time by the Kafka processing platform and Elasticsearch (ES). As a distributed stream processing platform, Kafka efficiently processes high-throughput data streams, while Elasticsearch (ES) is used to store, search, and analyze this data. The collected error codes are marked as channel error codes for subsequent analysis and processing.

[0087] At the same time, SMS post-process data is collected, including information such as user response to SMS, SMS delivery status, user click behavior, etc., which is used to evaluate the effectiveness of SMS marketing activities.

[0088] Data preprocessing is performed on the collected channel error codes. This includes extracting key fields, performing internal data conversion, and removing invalid information. This ensures data quality and usability, laying the foundation for subsequent analysis. Furthermore, SMS backend data is stored in a Redis cluster to improve data processing speed and efficiency.

[0089] After preprocessing, channel error codes are statistically analyzed to identify error patterns, frequencies, and potential causes. The analysis results are stored in the database's channel error information record table, providing data support for subsequent error code classification and intelligent decision-making. Simultaneously, SMS backend data stored in the Redis cluster is also persistently stored in the database.

[0090] The beneficial effects of this embodiment include: ensuring data accuracy and availability by collecting and preprocessing channel error codes in real time, providing a reliable data foundation for subsequent analysis and decision-making. Using a combination of Kafka and ES to quickly respond to large amounts of data collection and processing needs, improving responsiveness and processing efficiency. The introduction of a Redis cluster enables more efficient management of in-memory data, increasing data processing speed and reducing resource waste. Through data statistics and analysis, a deep understanding of error code patterns and causes is achieved, providing data support for error prevention and optimization.

[0091] In summary, the efficiency and accuracy of intelligent marketing distribution methods are improved through real-time data collection, efficient data processing and in-depth data analysis.

[0092] See also Figure 3 , Figure 3 The following is a flow chart showing the error code identification and classification process in the intelligent marketing distribution process provided by an embodiment of the present application. In one embodiment, S102 includes:

[0093] S301, performing statistics and analysis on error codes according to the channel error information record table, and obtaining a rule table including error code classification rules;

[0094] S302, using a decision tree algorithm or a SVM algorithm, automatically marking the channel error codes returned by each channel according to the error code classification rule to obtain marking information;

[0095] S303: Identify and establish a record relationship between each channel error code and the error code classification type according to the tag information to form channel error code classification information.

[0096] S102 may further include: S304, maintaining and managing the channel error code classification information.

[0097] This approach can be understood as first performing statistics and analysis on the collected error codes based on the channel error information record table. For example, this involves quantitatively analyzing the error code frequency, type, and correlation to identify error code patterns and trends. Error code classification rules are then extracted. These rules can be considered to represent the characteristics and classification criteria of different error codes, providing a basis for subsequent automatic classification.

[0098] Using a decision tree algorithm or a support vector machine (SVM) algorithm, error codes returned by each channel are automatically labeled according to error code classification rules. Machine learning algorithms can be considered to categorize and label error codes based on predefined rules or learned patterns. Labeling information includes the error code category and severity. Based on this labeling information, a record relationship between each channel's error code and its classification type is identified and established. Each error code is assigned a classification label, forming channel error code classification information. This classification information helps the system quickly identify error types and implement appropriate handling strategies during subsequent marketing distribution.

[0099] Furthermore, the channel error code classification information can be maintained and managed by updating classification rules, deleting obsolete error codes, adding new error codes, etc., to ensure the accuracy and timeliness of the channel error code classification information.

[0100] The beneficial effects of this embodiment include: by marking and categorizing error codes, manual intervention can be reduced, and the speed and accuracy of error handling can be improved. Through automated error code statistical analysis, machine learning-driven automatic marking, and systematic error code classification, the error handling capabilities of the intelligent marketing distribution method are improved.

[0101] See also Figure 4 , Figure 4 The following is a flow chart showing the channel recommendation process in the intelligent marketing distribution process provided by an embodiment of the present application. In one embodiment, S103 includes:

[0102] S401, analyzing and mining the SMS post-process data using Hadoop technology to obtain an indicator report, wherein the indicator report is used to provide different dimensions of the SMS marketing task;

[0103] S402: Calculate the scores of each channel in different dimensions of the indicator report according to the business algorithm, and calculate the final channel score based on the weight of each dimension; the dimensions include project click cost, project click rate, project gross profit ratio, price, and project success rate; the weight of each dimension can be determined based on actual business conditions or preset values;

[0104] S403 , sorting the final channel scores, obtaining a recommended channel group and the number of channels in the recommended channel group, and obtaining a salvaged channel group and the number of channels in the salvaged channel group.

[0105] Hadoop technology enables analysis and data mining of SMS post-process data. This is because Hadoop excels at handling large datasets, storing and processing large amounts of data, making it suitable for complex data analysis and mining tasks. This step generates a metrics report containing key performance indicators for SMS marketing tasks across different dimensions, providing data support for subsequent channel scoring and recommendations.

[0106] Then, based on a business algorithm, each channel's score across various dimensions in the indicator report is calculated. Dimensions include project cost per click, project click-through rate, project gross profit margin, price, and project success rate, each reflecting a different aspect of channel performance. The final channel score is then calculated based on the weighting of each dimension. Weightings can be determined based on actual business circumstances, reflecting the contribution of each dimension to business objectives and ensuring the scoring system can flexibly adapt to varying business needs and market environments.

[0107] Finally, all channels are sorted according to the final channel score, and a recommended channel group and the number of channels within it are obtained. The recommended channel group contains the channels with the highest scores and is considered the best choice for SMS marketing tasks. The salvaged channel group and the number of channels within it are also obtained. The channel score can be calculated by multiplying the score of each dimension by its corresponding weight, and then summing the product of the weighted scores for each dimension. If the business algorithm is a weighted scoring algorithm, for example, the weights used in the calculation process for the recommended channel group and the salvaged channel group will be different. Suppose a business algorithm needs to calculate a channel score based on three dimensions: project cost per click, project click-through rate, and project gross profit ratio. The weights for each dimension are assigned based on business importance; for example, cost per click accounts for 30%, click-through rate accounts for 40%, and gross profit ratio accounts for 30%. The final channel score can be calculated using the following formula: Final score = (0.3 × project cost per click score) + (0.4 × click-through rate score) + (0.3 × gross profit ratio score). The business algorithm is also, for example, a machine learning model, which predicts the future performance of each channel based on historical data and scores it accordingly. The channel with the highest predicted recommended channel score is obtained and placed in the recommended channel group, and the channel with the highest predicted salvage channel score is obtained and placed in the salvage channel group.

[0108] The beneficial effects of this embodiment include: accurate channel scoring and recommendation ensure that SMS marketing tasks are sent through the most effective channels, thereby improving marketing efficiency and effectiveness. Marketing resources are rationally allocated based on channel performance scores, prioritizing investment in high-return channels and reducing inefficient expenditures. By considering the weighting of different dimensions, it can adapt to different business goals and market changes, improving business flexibility and adaptability. By selecting the optimal channel for sending SMS messages, ineffective and intrusive marketing messages can be reduced, thereby improving the customer experience.

[0109] In summary, through big data analysis, multi-dimensional scoring and intelligent recommendation, the efficiency and effectiveness of the intelligent marketing distribution system have been improved, resource allocation has been optimized, and business adaptability and customer experience have been enhanced.

[0110] See also Figure 5 , Figure 5 A flow chart of intelligent distribution in the intelligent marketing distribution process provided in an embodiment of the present application is shown.

[0111] In one embodiment, S104 includes:

[0112] S501, obtaining corresponding processing strategies and retransmission strategies according to the error code classification type of each channel error code;

[0113] S502, performing verification pre-processing on the SMS marketing task to be distributed, including filtering the blacklist, frequency, and encryption and decryption;

[0114] S503, sending SMS data through the recommended channel group, wherein the number of numbers sent by each channel in the recommended channel group is calculated based on the channel weight;

[0115] S504: Automatically trigger the corresponding personalized policy mechanism based on the channel error code returned by the failed channel and the channel error code classification information. When the processing policy is retransmission, the SMS data is returned to the channel within the salvage channel group for forwarding. If a channel within the salvage channel group fails to send, determine whether the processing policy for the error code returned by the failed channel is retransmission. If so, continue forwarding using other channels in the salvage channel group; if not, stop forwarding. The processing policies include immediate retransmission, delayed retransmission, no retransmission, immediate retry, delayed retry, and no retry.

[0116] This embodiment obtains the corresponding processing strategy and retransmission strategy according to the error code classification type of each channel error code. This step is based on the error code classification information and rule table established previously to ensure that each error code can correspond to a preset strategy. Then, the SMS marketing task to be distributed is pre-processed for verification. The pre-processing includes filtering the blacklist, frequency control, and encryption and decryption processing to ensure the compliance, security, and effectiveness of SMS sending. SMS data is sent through the recommended channel group, where the number of numbers sent by each channel in the recommended channel group is calculated based on the channel weight. Since the channel score and weight information are used, the effect of SMS distribution can be optimized.

[0117] If a channel fails to send, the corresponding personalized policy mechanism will be automatically triggered based on the returned channel error code and channel error code classification information. For processing policies that require retransmission, the SMS data will be returned to the channel within the retrieving channel group for forwarding.

[0118] If a channel in the salvage channel group fails to send, the system determines whether the error code returned by the failed channel is a retransmission policy. If so, the system continues forwarding using other channels in the salvage channel group. If not, the system stops forwarding.

[0119] The beneficial effects of this embodiment include: intelligent error code processing and personalized policy implementation can improve the success rate of SMS delivery and reduce the number of failed SMS messages. SMS tasks are allocated according to channel weights, optimizing resource utilization and improving marketing efficiency. In summary, this technical solution improves the efficiency of the intelligent marketing distribution system through intelligent error handling, verification preprocessing, optimized resource allocation, and personalized policy mechanisms.

[0120] In one embodiment, S104 further includes: receiving error codes returned by each channel in real time, and saving the error codes to the ES.

[0121] The technical solution of this embodiment can monitor and receive error codes returned from various channels (such as SMS and email) in real time. The results and status of channel operations can be captured instantly through an integrated API. Error codes provide direct feedback on any problems encountered by the channel while executing a send task. Received error codes are stored in ES, which can process and store large amounts of data while providing fast search and analysis capabilities.

[0122] The beneficial effects of this embodiment include: Real-time reception and storage of error codes enables rapid response to transmission issues, improving error handling efficiency. Utilizing ES's search function, in-depth analysis of error code data is possible, identifying error trends and patterns, providing data support for optimizing transmission strategies. Real-time monitoring and processing of error codes enables the system to promptly adjust transmission plans to adapt to changing channel conditions, improving system responsiveness and adaptability. Analyzing error code data can identify which channels are most reliable and which require optimization or replacement, thereby optimizing resource allocation.

[0123] As an example, an intelligent marketing distribution method includes the following steps:

[0124] S1. Collect, analyze, and preprocess error codes returned by the channel, as well as marketing post-process data. Use Kafka + ES to collect error codes generated during SMS delivery in real time, ensuring real-time processing of error information returned by the channel. The error codes and related information are then recorded in log files and databases for subsequent analysis and preprocessing using big data technologies (Hadoop and Spark).

[0125] During the SMS sending process, Kafka+ES technology is used to collect channel error code information, and then asynchronous consumption and storage are used to record the error code and related information in log files and databases. Then, big data technology is used for analysis and preprocessing, which can fully ensure the timeliness and accuracy of data collection and provide guarantees for system decision-making.

[0126] S2. Each channel returns a different error code. Deep machine learning (SVM, decision trees, etc.) is used to process and classify massive amounts of error codes. Collected channel error codes are automatically categorized according to classification rules. The channel classification results are stored and archived. A rule table is established, correlating information between corresponding channels and error code classifications to ensure systematic data traceability. Error code categories in the system-defined rule table include network issues, invalid numbers, and unavailable services.

[0127] By establishing error code classification rules and using deep machine learning to process massive amounts of data, we can improve data processing efficiency and accuracy and achieve efficient and automatic classification of error codes.

[0128] S3. Perform big data analysis and mining on massive amounts of SMS post-process data. A recommendation algorithm is used to calculate a score for each channel. High-quality channels are then recommended and divided into a recommended channel group and a salvaged channel group. The recommended channel group and the salvaged channel group can be considered to have different recommendation dimensions when using the recommendation algorithm. Each channel in the recommended channel group and the salvaged channel group consists of several channels, each with a score. The weight of each channel is determined by calculating the quality score.

[0129] S4. Intelligent marketing distribution is performed based on channel error code classification, recommended channel groups, and retrieval channel groups. According to the recommended channel group, the number of marketing in each channel is calculated by weight for each marketing batch. After the channel distribution fails, the personalized strategy mechanism is automatically triggered based on the identified error code classification. Different processing strategies are adopted for different types of error codes, such as immediate retransmission, delayed retransmission, no retransmission, immediate retry, delayed retry, no retry, etc. The retransmission operation is automatically performed, and the data enters the retrieval channel group channel. Then, based on the maximum number of retrievals, it is decided whether to continue to intelligently distribute to the channels in the retrieval channel group. The maximum number of retrievals is, for example, 10 times, 11 times, 20 times, etc.

[0130] Based on a method that combines channel error code classification with system-recommended channel groups and salvage channel groups, the error code returned by the channel is identified by the error code classification to determine whether to trigger the automatic retransmission mechanism, thereby implementing a set of intelligent marketing distribution engines and combining them with monitoring strategies to ensure the effective operation of the system.

[0131] Its beneficial effects include: 1. The collection, storage, analysis, processing, classification and archiving management of channel error codes establishes a corresponding relationship between disorganized channel error codes and orderly error code classification, providing data support for subsequent intelligent marketing distribution and the entire company's later decision-making. 2. Through the channel groups (recommended channel groups) and salvage channel groups recommended by big data, the reliability and utilization rate of the channels are improved, human subjectivity and uncertainty are reduced, and the limitations of being unable to provide personalized solutions after a single channel fails to send, as well as the insufficient utilization of high-quality customer data, are solved, so as to optimize the use value of customer data and channel marketing as much as possible. 3. The intelligent marketing distribution method based on deep learning error code classification automatically generates personalized marketing strategies through real-time analysis of user behavior data and error code information, so as to maximize user experience and satisfaction and improve user conversion rate. The entire intelligent marketing process reduces unnecessary marketing investment, improves marketing efficiency, and enhances the market competitiveness of enterprises.

[0132] System embodiment.

[0133] See also Figure 6 , Figure 6 A structural diagram of the intelligent marketing distribution system provided in an embodiment of the present application is shown.

[0134] This embodiment provides an intelligent marketing distribution system based on deep learning. Its specific implementation is consistent with the implementation method and technical effects described in the above method embodiment, and will not be repeated here. The intelligent marketing distribution system includes:

[0135] A data collection module, the data collection module is used to obtain channel error codes and SMS post-process data; perform data statistics and analysis on the multiple channel error codes obtained and save them to a database, and save the SMS post-process data to the database; the channel error codes can be collected from the intelligent distribution module;

[0136] An error code identification and classification module, which is used to automatically classify each of the channel error codes through a deep machine learning algorithm, obtain tag information of each of the channel error codes, and associate each of the channel error codes with an error code classification type as channel error code classification information;

[0137] A channel recommendation module, configured to obtain a recommended channel group and the number of channels in the recommended channel group through data analysis and mining based on the SMS post-process data;

[0138] An intelligent distribution module is used to distribute SMS marketing tasks for each marketing batch according to the channel weight of the recommended channel group. If the distribution fails, the corresponding personalized strategy mechanism is automatically triggered according to the channel error code classification information. The personalized strategy mechanism is used to adopt different SMS data retransmission processing strategies for channel error codes of different error code classification types.

[0139] In one embodiment, the channel recommendation module is further configured to obtain a salvage channel group and the number of channels in the salvage channel group through data analysis and mining based on the SMS post-process data in the database;

[0140] The intelligent distribution module is also used to execute the SMS data retransmission processing strategy. If the SMS data enters the channel of the retrieval channel group, it will determine whether to continue to distribute it to the channel in the retrieval channel group based on the actual retrieval times and the maximum preset retrieval times.

[0141] See also Figure 7 , Figure 7 A structural diagram of the data collection module provided in an embodiment of the present application is shown.

[0142] In one embodiment, the data collection module includes:

[0143] A first collection unit, which is used to collect error codes generated in real time through the Kafka processing platform and ES during the SMS data sending process and use them as channel error codes;

[0144] A second collecting unit, the second collecting unit is used to collect SMS post-process data;

[0145] A first preprocessing unit, configured to perform data preprocessing on the channel error code, extracting field information, performing internal data conversion, and removing invalid information, and storing the SMS post-process data in a Redis cluster;

[0146] The data statistics and analysis unit is used to perform data statistics and analysis on the channel error code after data preprocessing and store it in the channel error information record table of the database; it is also used to store the SMS post-process data stored in the Redis cluster in the database.

[0147] See also Figure 8 , Figure 8 A structural diagram of the error code identification and classification module provided in an embodiment of the present application is shown.

[0148] In one embodiment, the error code identification and classification module includes:

[0149] a rule table establishing unit, configured to perform statistics and analysis on error codes according to the channel error information record table, and obtain a rule table including error code classification rules;

[0150] An information marking unit, configured to automatically mark the channel error codes returned by each channel according to the error code classification rule using a decision tree algorithm or a support vector machine (SVM) algorithm to obtain marking information;

[0151] a relationship establishing unit, configured to identify and establish a record relationship between each channel error code and an error code classification type according to the tag information, thereby forming channel error code classification information;

[0152] A record relationship maintenance unit is used to maintain and manage the channel error code classification information.

[0153] See also Figure 9 , Figure 9 A schematic diagram of the structure of the channel recommendation module provided in an embodiment of the present application is shown.

[0154] In one embodiment, the channel recommendation module includes:

[0155] A data analysis and mining unit, which is used to analyze and mine the SMS post-process data using Hadoop technology to obtain an indicator report, which is used to provide different dimensions of SMS marketing tasks;

[0156] A scoring calculation unit is used to calculate the scores of each channel in different dimensions of the indicator report based on a business algorithm, including project click cost, project click rate, project gross profit ratio, price, and project success rate; and calculate the final channel score based on the weight of each dimension; wherein the weight of each dimension is obtained based on actual business conditions;

[0157] The channel group acquisition unit is used to sort the final channel scores, acquire the recommended channel group and the number of channels in the recommended channel group, and acquire the salvaged channel group and the number of channels in the salvaged channel group.

[0158] See also Figure 10 , Figure 10 A structural diagram of the intelligent distribution module provided in an embodiment of the present application is shown.

[0159] In one embodiment, the intelligent distribution module includes:

[0160] a configuration unit, configured to obtain corresponding processing strategies and retransmission strategies according to error code classification types of the channel error codes;

[0161] A second pre-processing unit, which is used to perform verification pre-processing on the SMS marketing task to be distributed, including filtering blacklist, frequency, and encryption and decryption;

[0162] a marketing distribution unit, configured to send SMS data via the recommended channel group, wherein the number of numbers sent by each channel in the recommended channel group is calculated based on a channel weight;

[0163] A retransmission judgment unit is used to automatically trigger a corresponding personalized strategy mechanism based on the channel error code returned by the channel that failed to send and the classification information of the channel error code. The processing strategies include immediate retransmission, delayed retransmission, no retransmission, immediate retry, delayed retry and no retry. When the processing strategy is retransmission, the SMS data is returned to the channel in the retrieval channel group for forwarding. If a channel in the retrieval channel group fails to send, it is determined whether the processing strategy of the error code returned by the failed channel is retransmission. If it is retransmission, the forwarding is continued using other channels in the retrieval channel group. If it is not retransmission, the forwarding is stopped.

[0164] In one embodiment, the intelligent distribution module further includes:

[0165] The information return unit is used to receive the error codes returned by each channel in real time and save the error codes to the ES.

[0166] As an example, an intelligent marketing distribution system is provided, which includes a data collection module, an error code identification and classification module, a channel recommendation module and an intelligent distribution module.

[0167] The data collection module receives and stores information returned by the channel, performs data preprocessing, and finally, through data statistics and analysis, provides a basis for establishing error code classification rules. It also receives and stores SMS post-processing data to build a large database, providing data basis for subsequently obtaining recommended channel groups. Specifically, the data collection module performs the following steps:

[0168] In the first step, a job retrieves error information returned by the channel from ES and performs preprocessing to extract field information, perform internal data conversion, and remove invalid information. Simultaneously, another job receives and processes SMS post-processing data, first storing it in a Redis cluster and then saving it to the database.

[0169] The second step is to perform data statistics and analysis on the pre-processed channel error information, classify and display it, and further eliminate error codes with lower probability.

[0170] The third step is to save the analyzed data to the database and create an original channel error information record table.

[0171] The error code identification and classification module is used to establish an error code classification mechanism after statistical analysis of the error codes of all channels. Then, through machine learning algorithms, it automatically labels all channel error codes, establishes the relationship between different error codes and error code classifications for each channel, and maintains the error code classification table, channel information, channel error codes, and error code classification record relationships and archiving management. Specifically, the error code identification and classification module is used to perform the following steps:

[0172] The first step is to collect statistics and analyze common error codes and establish an error code classification rule table.

[0173] In the second step, the unique error codes returned by each channel are automatically labeled according to the error code classification rules through machine learning (SVM, decision tree, etc.) algorithms.

[0174] The third step is to automatically identify and establish a record relationship between the error codes returned by each channel and the error code classification based on the tag information.

[0175] The fourth step is to maintain the record relationship and archiving management of channel information, channel error codes, error code classification, etc.

[0176] The channel recommendation module is used to analyze channel big data and calculate the scores of different dimensions of each channel (such as project click cost, project click rate, project gross profit ratio, price, project success rate, etc.), and then calculate the final score of each channel through a rule-based algorithm. Finally, it provides a recommended channel group and a return channel group. Specifically, the channel recommendation module is used to perform the following steps:

[0177] The first step is to use Hadoop-related technologies to analyze and mine the channel SMS post-process data and provide an indicator report.

[0178] The second step is to calculate the scores of different dimensions of each channel (such as project click cost, project click rate, project gross profit ratio, price, project success rate and other indicators) based on the business algorithm.

[0179] The third step is to determine the weight of each dimension based on the actual business situation, and finally calculate the final channel score based on the rules.

[0180] The fourth step is to determine the recommended channel group and the number of channels in the channel group, the recommended salvage channel group, and the number of channels in the salvage channel group based on the final channel score ranking and daily business configuration.

[0181] Daily business configuration can be understood as the configuration information obtained every day based on market dynamics, promotional activities, customer feedback, budget constraints and other factors. Daily business configuration includes the number of SMS messages sent, target audience, time window, and budget allocation parameters. The recapture channel group refers to a collection of channels used to process SMS messages that need to be resent after the initial sending fails, and has the ability to handle specific error types. Based on the daily business configuration and through big data recommendation, it is determined which channels should be included in the recapture channel group. For example, if the business configuration of a certain day emphasizes a high delivery rate to a specific area, the recapture channel recommended by big data is the channel that performs well in the area. The number of channels in the recapture channel group (and the recommended channel group) can be determined based on the current load situation. Therefore, by dynamically adjusting the recapture channel group and the number of channels, resources can be allocated more effectively to ensure the best marketing effect within a limited budget.

[0182] The intelligent distribution module is used to perform the following steps:

[0183] The first step is to configure the automatic retransmission strategy and handling strategy for error codes. Different handling strategies are used for different error codes, such as immediate retry, delayed retry, and no retry. You can create a retransmission strategy library to define the corresponding retransmission strategy for each error code category. The retransmission strategy includes parameters such as the number of retries and the interval time.

[0184] The second step is to obtain the sending task and content information regularly, filter the blacklist, frequency, encryption and decryption and other verification preprocessing.

[0185] The third step is to send SMS data through the recommended channel group and calculate the number of numbers that fall on the channel according to the weight of each channel in the recommended channel group.

[0186] For example, a recommended channel group has five channels: Channel A, Channel B, Channel C, Channel D, and Channel E, with channel weights of 10%, 15%, 20%, 25%, and 30%, respectively. If a total of 10,000 numbers are marketed, then theoretically, Channel A would receive 1,000, Channel B 1,500, Channel C 2,000, Channel D 2,500, and Channel E 3,000.

[0187] The fourth step is to make a judgment based on the error code returned by the channel that failed to send. The error code is a feedback provided by the channel to indicate the reason for the failure to send. The error code is judged according to the preset retransmission strategy to see if it meets the (automatic) retransmission conditions. If the automatic retransmission conditions are met, the corresponding retransmission strategy is obtained, and the corresponding SMS data is returned to the channel in the retrieval channel group for forwarding. If a channel in the retrieval channel group still fails to send, it is determined whether the error code returned by the channel meets the automatic retransmission strategy. If so, the message will continue to be forwarded by other channels in the retrieval channel group until there is no other channel to send after polling. If the automatic retransmission strategy is not met, the forwarding will be stopped.

[0188] For example, there are three retrieving channels: Channel X, Channel Y, and Channel Z. Suppose that error code 1 of Channel A in the recommended channel group meets the retransmission policy, while error codes 2 and 3 do not. If Channel A sends a call to a certain number and the return error code is 1, the call will enter the retrieving channel group and, based on the weight, be forwarded to one of Channels X, Y, or Z. Assuming the call lands on Channel X, and error code 4 of Channel X meets the retransmission policy, if Channel X also fails to send the call to the same number with error code 4, the call will be forwarded to Channel Y or Channel Z, and so on. If an intermediate channel succeeds or the return error code does not meet the corresponding channel's retransmission policy, the forwarding process ends.

[0189] The fifth step is to monitor the intelligent marketing distribution strategy in real time, dynamically adjust the strategy or issue email alerts for special scenarios and emergencies. In other words, the execution of key indicators of the intelligent marketing distribution strategy, including SMS sending status, channel performance, user response, etc., is continuously monitored. Marketing strategies are automatically adjusted to adapt to market changes or respond to emergencies. For example, if a channel suddenly performs poorly, the number of SMS sent through this channel will be automatically reduced, or a backup channel will be switched to ensure delivery efficiency and success rate. Special scenarios, such as market emergencies, will automatically trigger the alarm mechanism and send email notifications to system administrators or relevant persons in charge so that they can understand the situation in a timely manner and take necessary manual intervention measures.

[0190] Step 6: Receive and process the error codes returned by each channel in real time, and save the information to ES for use by subsequent data collection, analysis, and statistics modules.

[0191] Electronic device embodiments.

[0192] This embodiment provides an electronic device comprising a memory and a processor. The memory stores a computer program, and the processor is configured to implement the steps of any of the method embodiments when executing the computer program. The specific implementation and technical effects achieved are consistent with those described in the aforementioned method embodiments, and some details are omitted for brevity.

[0193] See also Figure 11 , Figure 11 The figure shows a structural block diagram of an electronic device provided in an embodiment of the present application.

[0194] The electronic device 10 may include, for example, at least one memory 11 , at least one processor 12 , and a bus 13 connecting different platform systems.

[0195] The memory 11 may include (computer) readable media in the form of volatile memory, such as a random access memory (RAM) 111 and / or a cache memory 112 , and may further include a read-only memory (ROM) 113 .

[0196] The memory 11 also stores a computer program, which can be executed by the processor 12 so that the processor 12 implements the steps of any of the above methods.

[0197] The memory 11 may also include a utility 114 having at least one program module 115, such program module 115 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include an implementation of a network environment.

[0198] Accordingly, the processor 12 may execute the aforementioned computer program and may execute the utility 114 .

[0199] The processor 12 may be implemented as one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field programmable gate arrays (FPGAs), or other electronic components.

[0200] Bus 13 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures.

[0201] The electronic device 10 may also communicate with one or more external devices, such as a keyboard, a pointing device, a Bluetooth device, etc., and may also communicate with one or more devices capable of interacting with the electronic device 10, and / or any device that enables the electronic device 10 to communicate with one or more other computing devices (e.g., a router, a modem, etc.). Such communication may be performed via an input / output interface 14. Furthermore, the electronic device 10 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 15. The network adapter 15 may communicate with other modules of the electronic device 10 via the bus 13. It should be understood that, although not shown in the figures, other hardware and / or software modules may be used in conjunction with the electronic device 10 in actual applications, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0202] Computer readable storage medium embodiments.

[0203] The embodiment of the present application also provides a computer-readable storage medium, the specific embodiment of which is consistent with the embodiment described in the above method embodiment and the technical effects achieved, and some contents will not be repeated here.

[0204] The computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, the computer program implements the steps of any of the above methods or implements the functions of any of the above electronic devices.

[0205] The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. In an embodiment of the present application, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by an instruction execution system, device, or device or used in combination with it. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0206] A computer-readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, carrying readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable storage medium may also be any computer-readable medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, or any suitable combination thereof. The program code used to perform the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a standalone software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. Where a remote computing device is involved, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0207] The technical features of the above-described embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above-described embodiments only express several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent application. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present invention, several variations and improvements can be made, which all fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent of the present invention shall be based on the attached claims.

Claims

1. An intelligent marketing distribution system based on deep learning, characterized in that: The intelligent marketing distribution system includes: A data collection module, the data collection module is used to obtain channel error codes and short message post-process data; perform data statistics and analysis on the obtained multiple channel error codes and save them to a database, and save the short message post-process data to the database; An error code identification and classification module, the error code identification and classification module is used to automatically classify each of the channel error codes through a deep machine learning algorithm, obtain tag information of each of the channel error codes; associate each of the channel error codes with an error code classification type and use it as channel error code classification information; A channel recommendation module, the channel recommendation module is used to obtain a recommended channel group and the number of channels in the recommended channel group through data analysis and mining based on the short message post-process data; An intelligent distribution module is used to distribute SMS marketing tasks for each marketing batch according to the channel weight of the recommended channel group. If the distribution fails, the corresponding personalized strategy mechanism is automatically triggered according to the channel error code classification information. The personalized strategy mechanism is used to adopt different SMS data retransmission processing strategies for channel error codes of different error code classification types.

2. The deep learning-based intelligent marketing distribution system according to claim 1, characterized in that: The channel recommendation module is also used to obtain a salvage channel group and the number of channels in the salvage channel group through data analysis and mining based on the short message post-process data in the database; The intelligent distribution module is also used to execute the SMS data retransmission processing strategy. If the SMS data enters the channel of the retrieval channel group, it is determined whether to continue to distribute it to the channel in the retrieval channel group according to the actual retrieval times and the maximum preset retrieval times.

3. The deep learning-based intelligent marketing distribution system according to claim 2, characterized in that: The data collection module includes: A first collection unit, the first collection unit is used to collect error codes generated in real time through the kafka processing platform and ES during the SMS data sending process and use them as channel error codes; A second collection unit, the second collection unit is used to collect short message post-process data; A first preprocessing unit, the first preprocessing unit is used to perform data preprocessing on the channel error code to extract field information, convert internal data, and remove invalid information, and store the short message post-process data in a Redis cluster; The data statistics and analysis unit is used to perform data statistics and analysis on the channel error code after data preprocessing and store it in the channel error information record table of the database; it is also used to store the short message post-process data stored in the Redis cluster in the database.

4. The deep learning-based intelligent marketing distribution system according to claim 3 is characterized in that: The error code identification and classification module includes: A rule table establishment unit, the rule table establishment unit is used to count and analyze the error codes according to the channel error information record table to obtain a rule table including error code classification rules; An information marking unit, the information marking unit is used to automatically mark the channel error codes returned by each channel according to the error code classification rule by using a decision tree algorithm or a SVM algorithm to obtain marking information; A relationship establishing unit, the relationship establishing unit is used to identify and establish a record relationship between each channel error code and the error code classification type according to the tag information to form channel error code classification information; A record relationship maintenance unit, wherein the record relationship maintenance unit is used to maintain and manage the channel error code classification information.

5. The deep learning-based intelligent marketing distribution system according to claim 4 is characterized in that: The channel recommendation module includes: A data analysis and mining unit, wherein the data analysis and mining unit is used to analyze and mine the SMS post-process data through Hadoop technology to obtain an indicator report, wherein the indicator report is used to provide different dimensions of SMS marketing tasks; A scoring calculation unit, which is used to calculate the scores of each channel in different dimensions of the indicator report according to the business algorithm, the dimensions including project click cost, project click rate, project gross profit ratio, price and project success rate; and calculate the final channel score according to the weight of each dimension; wherein the weight of each dimension is obtained according to the actual business situation; A channel group acquisition unit is used to sort the final channel scores, acquire a recommended channel group and the number of channels in the recommended channel group, and acquire a salvaged channel group and the number of channels in the salvaged channel group.

6. The deep learning-based intelligent marketing distribution system according to claim 5, characterized in that: The intelligent distribution module includes: A configuration unit, the configuration unit is used to obtain corresponding processing strategies and retransmission strategies according to the error code classification type of each channel error code; A second preprocessing unit, the second preprocessing unit is used to perform verification preprocessing including filtering blacklist, frequency and encryption and decryption on the SMS marketing task to be distributed; A marketing distribution unit, the marketing distribution unit is used to send SMS data through the recommended channel group, wherein the number of numbers sent by each channel in the recommended channel group is calculated according to the channel weight; A retransmission judgment unit, the retransmission judgment unit is used to automatically trigger a corresponding personalized strategy mechanism according to a channel error code returned by a channel that fails to send and classification information of the channel error code, wherein the processing strategies include immediate retransmission, delayed retransmission, no retransmission, immediate retry, delayed retry and no retry; when the processing strategy is retransmission, the SMS data is returned to the channel in the retrieval channel group for forwarding; if a channel in the retrieval channel group fails to send, it is determined whether the processing strategy of the error code returned by the failed channel is retransmission, if it is retransmission, the forwarding is continued using other channels in the retrieval channel group, and if it is not retransmission, the forwarding is stopped.

7. The deep learning-based intelligent marketing distribution system according to claim 6, characterized in that: The intelligent distribution module also includes: The information return unit is used to receive the error code returned by each channel in real time and save the error code to ES.

8. An intelligent marketing distribution method based on deep learning error code classification, characterized in that: The intelligent marketing distribution method comprises: Obtaining channel error codes and SMS post-process data; performing data statistics and analysis on the obtained multiple channel error codes and saving them to a database, and saving the SMS post-process data to the database; Automatically classify each of the channel error codes through a deep machine learning algorithm to obtain tag information of each of the channel error codes; associate each of the channel error codes with an error code classification type and use the error code classification information as the channel error code; Based on the SMS post-process data, obtaining a recommended channel group and the number of channels in the recommended channel group through data analysis and mining; For each marketing batch, the SMS marketing task is distributed according to the channel weight of the recommended channel group. If the distribution fails, the corresponding personalized strategy mechanism is automatically triggered according to the channel error code classification information. The personalized strategy mechanism is used to adopt different SMS data retransmission processing strategies for channel error codes of different error code classification types.

9. An electronic device, characterized in that: The electronic device comprises a memory and at least one processor, wherein the memory stores a computer program, and the at least one processor is configured to implement the steps of the method of claim 8 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, the steps of the method according to claim 8 are implemented.

Images