An AI big data real-time processing and analysis method

Abstract

The application discloses an AI big data real-time processing and analysis method, relates to the technical field of enterprise order data analysis and processing, and comprises the following steps: collecting current enterprise operating product information, performing model statistics on the enterprise operating products, establishing an enterprise product database, and dividing the enterprise operating products into main products and byproducts; collecting enterprise historical transaction order data, classifying orders according to the main product models in the historical transaction orders, performing byproduct correlation degree analysis on the enterprise operating main products based on the historical transaction order contents under different categories, and performing reasonable degree analysis on subsequent enterprise to-be-transacted orders according to the correlation degree analysis results of the enterprise operating main products and byproducts. The application processes and analyzes enterprise to-be-transacted order data in depth by combining enterprise historical order big data, judges the reasonableness of the to-be-transacted orders, and optimizes the real-time processing and analysis process of enterprise business data by considering the particularity of the purchasing party.

Description

Technical Field

[0001] This invention relates to the field of enterprise order data analysis and processing technology, and more specifically, to an AI big data real-time processing and analysis method. Background Technology

[0002] In today's highly competitive business environment, the accuracy and efficiency of enterprise order data processing play a crucial role in the success or failure of its operations. However, traditional enterprise order data processing methods have many drawbacks and are difficult to meet the needs of enterprises for refined management and improved customer satisfaction.

[0003] Enterprises lack in-depth mining and effective utilization of historical transaction order big data in the order review process. It is difficult to deeply analyze the correlation between order data of the enterprise's main products and by-products, resulting in an inability to accurately identify potential errors or unreasonable aspects of the orders by salespersons, leading to low practicality. At the same time, because different buyers may have unique purchasing habits and business needs, the existing big data real-time processing and analysis processes cannot further analyze and process enterprise order data in conjunction with the characteristics of the buyers, resulting in low functionality. This significantly reduces the smoothness of enterprise business processes, causes serious waste of resources in production, distribution and other links, and low overall operational efficiency.

[0004] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0005] In response to the problems in related technologies, this invention proposes an AI big data real-time processing and analysis method to overcome the aforementioned technical problems existing in the current related technologies.

[0006] Therefore, the specific technical solution adopted by the present invention is as follows:

[0007] A method for real-time AI big data processing and analysis, comprising the following steps:

[0008] S1. Collect information on the products currently operated by the enterprise, conduct model statistics on the products operated by the enterprise, establish a product database for the enterprise, and classify the products operated by the enterprise into main products and by-products;

[0009] S2. Collect historical transaction order data of enterprises, classify orders according to the main product models in the historical transaction orders, and conduct by-product correlation analysis on the main products of enterprises based on the content of historical transaction orders under different categories;

[0010] S3. Based on the correlation analysis results between the enterprise's main products and by-products, conduct a reasonableness analysis on the subsequent pending orders of the enterprise, determine the reasonableness of the pending orders, mark the orders that fail the reasonableness determination as abnormal for the first time, and send the orders that pass the reasonableness determination to the warehousing department for picking and shipping.

[0011] S4. For orders marked as abnormal for the first time that fail the reasonableness judgment, extract the purchaser data of the current order marked as abnormal, and search the company's historical orders in combination with the purchaser data. Analyze the current company's historical orders for the abnormal main product and by-product in the order marked as abnormal, and make a second judgment on the order marked as abnormal.

[0012] In a preferred embodiment, S2 includes the following steps:

[0013] S21. Collect historical transaction order data of enterprises, establish an enterprise order database through MySQL, create files in the enterprise order database based on the main product models of the enterprise's business, and classify the corresponding historical transaction orders into the corresponding files according to the main product models in the historical transaction order data;

[0014] S22. Based on the types of by-products within the archives of different main products, conduct a correlation analysis between the main products and by-products, including the following steps:

[0015] S221. Extract historical transaction order data from different main product files, and calculate the co-occurrence frequency and lift of main products and secondary products in the current main product file:

[0016] ;

[0017] Where A and a represent the main product model and the by-product model, respectively. This represents the co-occurrence frequency of the main product A and the by-product a. This represents the number of times that main product A and by-product a appear simultaneously in historical transaction orders in the current main product file, where N represents the total number of orders.

[0018] ;

[0019] ;

[0020] in, This represents the degree of improvement between the main product A and the by-product a. This represents the frequency of by-product 'a' in the total orders. This represents the number of times by-product 'a' appears in historical transaction orders within the current main product file.

[0021] S222. Based on the co-occurrence frequency and lift of different by-products in the current main product file, calculate the correlation between different by-products and the current main product. The specific algorithm formula is as follows:

[0022] ;

[0023] ;

[0024] ;

[0025] in, This represents the degree of correlation between the main product A and the by-product a. These represent the weighting percentages of co-occurrence frequency and lift, respectively. When the main product and by-product a have a degree of correlation, a related by-product a file is created in each main product file. In each related by-product file, all by-product models with a correlation degree greater than 0 for the current main product are recorded, and the corresponding correlation degree values ​​are recorded.

[0026] In a preferred embodiment, S21 includes the following steps:

[0027] S211. Obtain the current historical transaction order data of the enterprise from the enterprise's internal sales management system, financial system, and order management system, including order number, order date, customer information, main product model, secondary product model, and product quantity;

[0028] S212. Establish an enterprise order database using MySQL, extract all different main product models from the enterprise product database, create a corresponding main product file in the database for each main product model, traverse every historical transaction order in the enterprise order database, and classify the corresponding orders into the corresponding main product files according to the main product model in the order data.

[0029] In a preferred embodiment, step S3 includes the following steps:

[0030] S31. Collect data on subsequent pending orders of enterprises. Based on the enterprise's product classification results in the enterprise's product database, extract the main product and by-product data from the pending orders. Based on the correlation analysis results of the enterprise's main products and by-products, obtain the correlation value between the by-products and main products of the current pending orders.

[0031] In a preferred embodiment, S31 includes the following steps:

[0032] S311. Extract the by-product models from the enterprise's subsequent pending orders, search for the current by-product model in different related by-product files in the enterprise's order database, extract all related main products of the current by-product, and establish a set of related main products. :

[0033] ;

[0034] in, This represents the set of all main product models with positive correlation to the current by-product in the enterprise's order database;

[0035] S312, Based on the product collection of related entities Determine if there is any overlap between the main product and the current pending orders:

[0036] Extract the main products from the current pending orders, establish a set h of main products for the current pending orders, and calculate the set of related main products. The intersection with the set of products h of the current pending orders:

[0037] ;

[0038] In the above formula, This represents the reasonableness of by-products in the current pending orders. When the by-product q in the pending order is related to the main product in the pending order, the reasonableness of all by-products in the current pending order is calculated. When the reasonableness of all by-products is 1, the current pending order is determined to be a reasonable order. The reasonable order is then transmitted into the enterprise's subsequent workflow and the warehousing department will handle the order picking and delivery.

[0039] When there is a by-product in any currently pending order When this occurs, it indicates that the by-products in the pending order are not related to the main product in the pending order, and the current order data is marked as an initial abnormal order.

[0040] In a preferred embodiment, step S4 includes the following steps:

[0041] S41. For the first abnormal order that fails the reasonableness judgment, extract the purchaser data of the current first abnormal order, search the historical orders of the current purchaser in the enterprise, and conduct historical analysis on the abnormal by-products in the current first abnormal order.

[0042] S42. Based on the historical analysis results and the purchaser's historical order history within the enterprise, a second assessment is conducted on the initial abnormal order to determine its final reasonableness.

[0043] In a preferred embodiment, S41 includes the following steps:

[0044] S411. Based on the name of the purchaser in the current initial abnormal order, search the enterprise order database to obtain all historical order data related to the purchaser in the current enterprise order database, establish a historical analysis dataset, and summarize the historical order data related to the purchaser into the historical analysis dataset.

[0045] S412. For the current initial abnormal order, count the by-product models with a reasonableness of 0 within the abnormal order, and retrieve the relevant historical orders from the purchaser containing the corresponding by-products from the historical analysis dataset. Count the main product models contained in the relevant historical orders and establish a historical main product model set. , where n is the model of the by-product with a reasonableness of 0 in the abnormal order.

[0046] In a preferred embodiment, S42 includes the following steps:

[0047] S421. Calculate the corresponding values ​​of by-product models with a reasonableness of 0 in different abnormal orders. The relationship with h:

[0048] ;

[0049] when When the abnormal by-product n in the current initial abnormal order meets the historical situation of the order purchaser, when all abnormal by-products meet the historical situation of the purchaser, the current initial abnormal order is judged as a reasonable order and transmitted into the enterprise's subsequent workflow, and the warehousing department will carry out the picking and delivery.

[0050] If any abnormal by-product exists in the initial abnormal order If the initial abnormal order is determined to be an unreasonable order, it will be returned to the company's salesperson for reconfirmation. After the salesperson confirms the order, it will be transferred to the company's subsequent workflow.

[0051] S422. When there is no historical order data related to the purchaser in the initial abnormal order in the enterprise, the current initial abnormal order is determined to be an unreasonable order. It is returned to the enterprise's salesperson for reconfirmation. After the salesperson confirms, the reasonable order is transmitted to the enterprise's subsequent workflow.

[0052] In a preferred embodiment, S1 includes the following steps:

[0053] S11. Compile current business product data, including product name and product model, and establish a business product database;

[0054] S12. Classify the products in the enterprise product database by type through the administrator account, including main products and secondary products, and create files for each in the enterprise product database, and classify the enterprise's operating products under the corresponding files into the corresponding files.

[0055] The beneficial effects of this invention are as follows:

[0056] 1. This invention analyzes the relationship between main products and by-products based on the main product and by-product orders in the big data of historical transaction orders of an enterprise. Based on the correlation between main products and by-products, it performs real-time rationality analysis on the pending orders within the enterprise, and discovers possible salesperson errors or unreasonable aspects in pending orders in real time. Corrections are made in the early stage of order delivery to prevent erroneous orders from entering the subsequent process, reduce customer complaints and returns, and enhance practicality.

[0057] 2. This invention targets pending orders of enterprises. Based on the correlation between the main products and by-products of the business entity and the historical order data of the current abnormal order purchaser and the enterprise, it performs a secondary analysis on the initial abnormal order. This avoids unreasonable judgment of abnormal orders due to the special characteristics of the purchaser. It provides a comprehensive and in-depth understanding of the enterprise's real-time order status, thereby reducing misjudgments caused by the purchaser's special purchasing habits, business needs and other factors, and ensuring the accuracy of order data processing and judgment.

[0058] 3. This invention uses historical big data to perform rationality analysis on real-time order data of enterprises, which can avoid errors in product input by salespersons due to similar product models. Accurate order judgment helps optimize the enterprise's business processes. Reasonable orders can smoothly enter subsequent production, delivery and other stages, reduce unnecessary process backtracking and adjustments caused by incorrect judgment, improve the smoothness of the entire enterprise's business processes, and enhance the functionality of the method. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a flowchart of an AI big data real-time processing and analysis method according to an embodiment of the present invention. Detailed Implementation

[0061] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0062] According to an embodiment of the present invention, a method for real-time processing and analysis of AI big data is provided.

[0063] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments:

[0064] Example 1:

[0065] like Figure 1 As shown, according to an embodiment of the present invention, an AI big data real-time processing and analysis method includes the following steps:

[0066] S1. Collect information on the products currently operated by the enterprise, conduct model statistics on the products operated by the enterprise, establish a product database for the enterprise, and classify the products operated by the enterprise into main products and by-products;

[0067] S11. Compile current business product data, including product name and product model, and establish a business product database;

[0068] S12. Classify the products in the enterprise product database by type through the administrator account, including main products and secondary products, and create files for each in the enterprise product database, and classify the enterprise's operating products under the corresponding files into the corresponding files.

[0069] It should be noted that the products sold by enterprises are categorized by administrator account. The main products are the equipment that can realize the core functions, while the secondary products are accessories that help to improve the main products to realize the core functions. For example, for a photography equipment sales company, the main products include cameras and camcorders, while the secondary products include lenses and tripods. For a manufacturing sales company, the main products include air pumps, while the secondary products include oil-water separators and airbrushes.

[0070] S2. Collect historical transaction order data of enterprises, classify orders according to the main product models in the historical transaction orders, and conduct by-product correlation analysis on the main products of enterprises based on the content of historical transaction orders under different categories;

[0071] S21. Collect historical transaction order data of enterprises, establish an enterprise order database through MySQL, create files in the enterprise order database based on the main product models of the enterprise's business, and classify the corresponding historical transaction orders into the corresponding files according to the main product models in the historical transaction order data;

[0072] S211. Obtain the current historical transaction order data of the enterprise from the enterprise's internal sales management system, financial system, and order management system, including order number, order date, customer information, main product model, secondary product model, and product quantity;

[0073] S212. Establish an enterprise order database using MySQL, extract all different main product models from the enterprise product database, create a corresponding main product file in the database for each main product model, traverse each historical transaction order data in the enterprise order database, and classify the corresponding orders into the corresponding main product files according to the main product model in the order data.

[0074] S22. Based on the types of by-products within the archives of different main products, conduct a correlation analysis between the main products and by-products, including the following steps:

[0075] S221. Extract historical transaction order data from different main product files, and calculate the co-occurrence frequency and lift of main products and secondary products in the current main product file:

[0076] ;

[0077] Where A and a represent the main product model and the by-product model, respectively. This represents the co-occurrence frequency of the main product A and the by-product a. This represents the number of times that main product A and by-product a appear simultaneously in historical transaction orders in the current main product file, where N represents the total number of orders.

[0078] It should be noted that the output range of co-occurrence frequency is between 0 and 1. The closer the output value is to 1, the higher the frequency of the current by-product and the main product appearing in the order, that is, the closer the correlation. The closer the value is to 0, the lower the frequency of their co-occurrence and the weaker the correlation. In addition, since all historical transaction orders in the current main product file contain the current main product, the co-occurrence frequency, support and confidence values ​​of the current main product and the by-product are the same.

[0079] ;

[0080] ;

[0081] in, This represents the degree of improvement between the main product A and the by-product a. This represents the frequency of by-product 'a' in the total orders. This represents the number of times by-product 'a' appears in historical transaction orders within the current main product file.

[0082] It should be noted that lift is used to determine whether the relationship between the main product and the by-product is accidental. The value of lift is greater than 0. A value greater than 1 indicates a positive correlation between the main product A and the by-product a, meaning that purchasing the main product A increases the likelihood of purchasing the by-product a. When the degree of increase is greater than 1... When the lift is equal to 1, it means there is no correlation between the main product A and the by-product a. When the lift is less than 1, it means there is a negative correlation between the main product A and the by-product a, that is, purchasing the main product A will reduce the likelihood of purchasing the by-product a.

[0083] S222. Based on the co-occurrence frequency and lift of different by-products in the current main product file, calculate the correlation between different by-products and the current main product. The specific algorithm formula is as follows:

[0084] ;

[0085] ;

[0086] ;

[0087] in, This represents the degree of correlation between the main product A and the by-product a. These represent the weighting percentages of co-occurrence frequency and lift, respectively. When the main product and by-product a have a degree of correlation, a related by-product a file is created in each main product file. In each related by-product file, all by-product models with a correlation degree greater than 0 for the current main product are recorded, and the corresponding correlation degree values ​​are recorded.

[0088] It should be noted that the initial calculation range for the lift of the main product and by-product is (0, +∞). The lift is then limited to between -1 and 1 using a conversion formula. The closer to 1, the stronger the positive correlation. A value closer to -1 indicates a stronger negative correlation, which helps to standardize the base numerical range for subsequent correlation calculations. Setting them to 0.6 and 0.4 respectively indicates that the weight of co-occurrence frequency is greater than the weight of lift. Alternatively, adjustments can be made based on the specific circumstances. The values ​​are adjusted to meet different practical usage conditions.

[0089] Example 2:

[0090] S3. Based on the correlation analysis results between the enterprise's main products and by-products, conduct a reasonableness analysis on the subsequent pending orders of the enterprise, determine the reasonableness of the pending orders, mark the orders that fail the reasonableness determination as abnormal for the first time, and send the orders that pass the reasonableness determination to the warehousing department for picking and shipping.

[0091] S31. Statistically analyze the data of subsequent pending orders of enterprises. Based on the enterprise's product classification results in the enterprise's product database, extract the main product and by-product data from the pending orders. Based on the correlation analysis results of the enterprise's main products and by-products, obtain the correlation value between the by-products and main products of the current pending orders.

[0092] S311. Extract the by-product models from the enterprise's subsequent pending orders, search for the current by-product model in different related by-product files in the enterprise's order database, extract all related main products of the current by-product, and establish a set of related main products. :

[0093] ;

[0094] in, This represents the set of all main product models with positive correlation to the current by-product in the enterprise's order database;

[0095] It should be noted that when querying the relationship between a by-product and the main product, it is necessary to use the database query function, such as the SELECT statement in MySQL, to filter out the main product containing relevant records based on the by-product model.

[0096] S312, Based on the product collection of related entities Determine if there is any overlap between the main product and the current pending orders:

[0097] Extract the main products from the current pending orders, establish a set h of main products for the current pending orders, and calculate the set of related main products. The intersection with the set of products h of the current pending orders:

[0098] ;

[0099] In the above formula, This represents the reasonableness of by-products in the current pending orders. When the by-product q in the pending order is related to the main product in the pending order, the reasonableness of all by-products in the current pending order is calculated. When the reasonableness of all by-products is 1, the current pending order is determined to be a reasonable order. The reasonable order is then transmitted into the enterprise's subsequent workflow and the warehousing department will handle the order picking and delivery.

[0100] When there is a by-product in any currently pending order When this occurs, it indicates that the by-products in the pending order are not related to the main product in the pending order, and the current order data is marked as an initial abnormal order.

[0101] It should be noted that when the reasonableness score is 1, it means that there is an intersection between the related main product set of the relevant by-product and the main product set of the current pending order. That is, there is a usage correlation between the related main product and the by-product. When all by-products in the pending order have a usage correlation with a certain main product in the pending order, the current pending order is a reasonable order and can proceed with subsequent order fulfillment and delivery. When the reasonableness score of any by-product is 0, it means that the current by-product may be an abnormal by-product in the current pending order, and there may be problems such as incorrect by-product model. A second verification is required to ensure the reasonableness of subsequent delivery orders and avoid returns and exchanges due to operational errors.

[0102] S4. For orders marked as abnormal for the first time that fail the reasonableness judgment, extract the purchaser data of the current order marked as abnormal, search the enterprise's historical orders in combination with the purchaser data, analyze the current enterprise's historical orders for the abnormal main product and by-product in the order marked as abnormal, and make a second judgment on the order marked as abnormal.

[0103] S41. For the first abnormal order that fails the reasonableness judgment, extract the purchaser data of the current first abnormal order, search the historical orders of the current purchaser in the enterprise, and conduct historical analysis on the abnormal by-products in the current first abnormal order.

[0104] S411. Based on the name of the purchaser in the current initial abnormal order, search the enterprise order database to obtain all historical order data related to the purchaser in the current enterprise order database, establish a historical analysis dataset, and summarize the historical order data related to the purchaser into the historical analysis dataset.

[0105] S412. For the current initial abnormal order, count the by-product models with a reasonableness of 0 within the abnormal order, and retrieve the relevant historical orders from the purchaser containing the corresponding by-products from the historical analysis dataset. Count the main product models contained in the relevant historical orders and establish a historical main product model set. , where n is the model of the by-product with a reasonableness of 0 in the abnormal order.

[0106] S42. Based on the historical analysis results and the purchaser's historical order history within the enterprise, a second assessment is conducted on the initial abnormal order to determine its final reasonableness.

[0107] S421. Calculate the corresponding values ​​of by-product models with a reasonableness of 0 in different abnormal orders. The relationship with h:

[0108] ;

[0109] when When the abnormal by-product n in the current initial abnormal order meets the historical situation of the order purchaser, when all abnormal by-products meet the historical situation of the purchaser, the current initial abnormal order is judged as a reasonable order and transmitted into the enterprise's subsequent workflow, and the warehousing department will carry out the picking and delivery.

[0110] If any abnormal by-product exists in the initial abnormal order If the initial abnormal order is determined to be an unreasonable order, it will be returned to the company's salesperson for reconfirmation. After the salesperson confirms the order, it will be transferred to the company's subsequent workflow.

[0111] It should be noted that, based on the purchaser's historical order data, the rationality of the initial abnormal order can be objectively assessed. For reasonable orders, they can directly proceed to the next process, improving work efficiency. For unreasonable orders, they can be promptly returned to the salesperson for confirmation, reducing the risk of errors in the company's order data.

[0112] S422. When there is no historical order data related to the purchaser in the initial abnormal order in the enterprise, the current initial abnormal order is determined to be an unreasonable order. It is returned to the enterprise's salesperson for reconfirmation. After the salesperson confirms, the reasonable order is transmitted to the enterprise's subsequent workflow.

[0113] It should be noted that by conducting historical analysis of the purchasing party's current order database for initial abnormal orders, it is possible to avoid abnormal correlation between by-products and main products due to low improvement, uncover potential business relationships between the purchasing party and the current enterprise, and achieve real-time analysis and processing of enterprise order data.

[0114] In summary, this invention analyzes the relationship between main products and by-products based on the main product and by-product order data in the enterprise's historical transaction order big data, and performs real-time rationality analysis on pending orders within the enterprise based on the correlation between main products and by-products. This allows for the timely detection of potential salesperson errors or unreasonable aspects in pending orders, enabling corrections before order delivery, preventing erroneous orders from entering the subsequent process, reducing customer complaints and returns, and enhancing practicality.

[0115] By analyzing the correlation between the main products and by-products of a company's pending orders and the company's historical order data, a secondary analysis is conducted on the initial abnormal orders. This avoids unreasonable judgments of abnormal orders due to the special characteristics of the buyers, and provides a comprehensive and in-depth understanding of the company's real-time order situation. This reduces misjudgments caused by the special purchasing habits and business needs of the buyers, ensuring the accuracy of order data processing and judgment.

[0116] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for real-time processing and analysis of AI big data, characterized in that, The method includes the following steps: S1. Collect information on the products currently operated by the enterprise, conduct model statistics on the products operated by the enterprise, establish a product database for the enterprise, and classify the products operated by the enterprise into main products and by-products; S2. Collect historical transaction order data of enterprises, classify orders according to the main product models in the historical transaction orders, and conduct by-product correlation analysis on the main products of enterprises based on the content of historical transaction orders under different categories; S3. Based on the correlation analysis results between the enterprise's main products and by-products, conduct a reasonableness analysis on the subsequent pending orders of the enterprise, determine the reasonableness of the pending orders, mark the orders that fail the reasonableness determination as abnormal for the first time, and send the orders that pass the reasonableness determination to the warehousing department for picking and shipping. S31. Statistically analyze the data of subsequent pending orders of enterprises. Based on the enterprise's product classification results in the enterprise's product database, extract the main product and by-product data from the pending orders. Based on the correlation analysis results of the enterprise's main products and by-products, obtain the correlation value between the by-products and main products of the current pending orders. S311. Extract the by-product models from the enterprise's subsequent pending orders, search for the current by-product model in different related by-product files in the enterprise's order database, extract all related main products of the current by-product, and establish a set of related main products. : ； in, This represents the set of all main product models with positive correlation to the current by-product q in the enterprise order database; S312, Based on the product collection of related entities Determine if there is any overlap between the main product and the current pending orders: Extract the main products from the current pending orders, establish a set h of main products for the current pending orders, and calculate the set of related main products. The intersection with the set of products h of the current pending orders: ； In the above formula, This represents the reasonableness of by-product q in the current pending orders, when When the by-product q in the pending order is related to the main product in the pending order, the reasonableness of all by-products in the current pending order is calculated. When the reasonableness of all by-products is 1, the current pending order is determined to be a reasonable order. The reasonable order is then transmitted into the enterprise's subsequent workflow and the warehousing department will handle the order picking and delivery. When there is a by-product in any currently pending order When this occurs, it means that the by-product q in the pending order is not related to the main product in the pending order, and the current order data is marked as the first abnormal order; S4. For orders marked as abnormal for the first time that fail the reasonableness judgment, extract the purchaser data of the current order marked as abnormal, and search the company's historical orders in combination with the purchaser data. Analyze the current company's historical orders for the abnormal main product and by-product in the order marked as abnormal, and make a second judgment on the order marked as abnormal.

2. The AI ​​big data real-time processing and analysis method according to claim 1, characterized in that, S2 includes the following steps: S21. Collect historical transaction order data of the enterprise, through... Establish an enterprise order database. Based on the product models of the enterprise's main business, create files in the enterprise order database. According to the main product models in the historical transaction order data, classify the corresponding historical transaction orders into the corresponding files. S22. Based on the types of by-products within the archives of different main products, conduct a correlation analysis between the main products and by-products, including the following steps: S221. Extract historical transaction order data from different main product files, and calculate the co-occurrence frequency and lift of main products and secondary products in the current main product file: ； in, These represent the main product model and the by-product model, respectively. This represents the co-occurrence frequency of the main product A and the by-product a. This represents the number of times that main product A and by-product a appear simultaneously in historical transaction orders in the current main product file, where N represents the total number of orders. ； ； in, This represents the degree of improvement between the main product A and the by-product a. This represents the frequency of by-product 'a' in the total orders. This represents the number of times by-product 'a' appears in historical transaction orders within the current main product file. S222. Based on the co-occurrence frequency and lift of different by-products in the current main product file, calculate the correlation between different by-products and the current main product. The specific algorithm formula is as follows: ； ； ； in, This represents the degree of correlation between the main product A and the by-product a. These represent the weighting percentages of co-occurrence frequency and lift, respectively. When the value is greater than 0, it means that there is a correlation between the main product A and the by-product a. A related by-product file is created in each main product file. In each related by-product file, all by-product models with a correlation greater than 0 for the current main product are recorded, and the corresponding correlation value is recorded.

3. The AI ​​big data real-time processing and analysis method according to claim 2, characterized in that, S21 includes the following steps: S211. Obtain the current historical transaction order data of the enterprise from the enterprise's internal sales management system, financial system, and order management system, including order number, order date, customer information, main product model, secondary product model, and product quantity; S212, Through Establish an enterprise order database, extract all different main product models from the enterprise product database, create a corresponding main product file in the database for each main product model, traverse every historical transaction order data in the enterprise order database, and classify the corresponding orders into the corresponding main product files according to the main product model in the order data.

4. The AI ​​big data real-time processing and analysis method according to claim 1, characterized in that, S4 includes the following steps: S41. For the first abnormal order that fails the reasonableness judgment, extract the purchaser data of the current first abnormal order, search the historical orders of the current purchaser in the enterprise, and conduct historical analysis on the abnormal by-products in the current first abnormal order. S42. Based on the historical analysis results and the purchaser's historical order history within the enterprise, a second assessment is conducted on the initial abnormal order to determine its final reasonableness.

5. The AI ​​big data real-time processing and analysis method according to claim 4, characterized in that, S41 includes the following steps: S411. Based on the name of the purchaser in the current initial abnormal order, search the enterprise order database to obtain all historical order data related to the purchaser in the current enterprise order database, establish a historical analysis dataset, and summarize the historical order data related to the purchaser into the historical analysis dataset. S412. For the current initial abnormal order, count the by-product models with a reasonableness of 0 within the abnormal order, and retrieve the relevant historical orders from the purchaser containing the corresponding by-products from the historical analysis dataset. Count the main product models contained in the relevant historical orders and establish a historical main product model set. , where n is the model of the by-product with a reasonableness of 0 in the abnormal order.

6. The AI ​​big data real-time processing and analysis method according to claim 5, characterized in that, S42 includes the following steps: S421. Calculate the correlation between the by-product models with a reasonableness of 0 in different abnormal orders: ； when When all abnormal by-products in the current initial abnormal order are consistent with the historical situation of the order purchaser, the current initial abnormal order is judged as a reasonable order and transmitted into the enterprise's subsequent workflow, and the warehousing department will carry out the picking and delivery. If any abnormal by-product exists in the initial abnormal order If the initial abnormal order is determined to be an unreasonable order, it will be returned to the company's salesperson for reconfirmation. After the salesperson confirms the order, it will be transferred to the company's subsequent workflow. S422. When there is no historical order data related to the purchaser in the initial abnormal order in the enterprise, the current initial abnormal order is determined to be an unreasonable order. It is returned to the enterprise's salesperson for reconfirmation. After the salesperson confirms, the reasonable order is transmitted to the enterprise's subsequent workflow.

7. The AI ​​big data real-time processing and analysis method according to claim 1, characterized in that, S1 includes the following steps: S11. Compile current business product data, including product name and product model, and establish a business product database; S12. Classify the products in the enterprise product database by type through the administrator account, including main products and secondary products, and create files for each in the enterprise product database, and classify the enterprise's operating products under the corresponding files into the corresponding files.

Images