Supply chain risk early warning method and device based on artificial intelligence

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a supply chain risk early warning method and device based on artificial intelligence, and the method comprises the steps: collecting the sales history of a merchant according to the artificial intelligence, and obtaining the early-warning commodities of the merchant according to the sales history; the method comprises the following steps: generating a commodity association binary tree according to early-warning commodities of a merchant, processing according to a sales history collected by artificial intelligence to obtain a sales cycle and a sales region section of the merchant, obtaining a sales expectation of the merchant, and generating a self-adaptive early-warning method of the merchant according to the sales expectation and the commodity association binary tree; and carrying out supply chain early warning on the merchant according to the self-adaptive early warning method of the merchant. According to the invention, differentiated artificial intelligence risk management strategies can be generated according to actual deviation conditions of different supply chain data, and the reliability of sales data is fully ensured.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based supply chain risk early warning method and device. Background Art

[0002] The rapid development of artificial intelligence technology has made supply chain supervision more reliable. However, at the same time, due to the large amount of supply chain data in different dimensions, once update errors and update delays occur, the accuracy of the supply chain analysis results cannot be guaranteed. Therefore, how to identify and manage artificial intelligence risks in the supply chain has become a technical problem that needs to be solved urgently.

[0003] For commodity sales companies, sales delays may occur due to problems with the supply chain data of supply chain digital devices or untimely updates. Therefore, if differentiated artificial intelligence risk management strategies cannot be generated based on the actual deviations of different supply chain data, the reliability of sales data cannot be guaranteed. Summary of the invention

[0004] To achieve the above objectives, this application provides the following technical solutions: According to a first aspect of the present invention, the present invention claims protection for a supply chain risk early warning method based on artificial intelligence, comprising: Collect the sales history of merchants based on artificial intelligence, and obtain the merchants' early warning products based on the sales history; Generate a binary tree of product associations based on the merchant's warnable products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectations, and generate an adaptive warning method for the merchant based on the sales expectations and the binary tree of product associations; Provide supply chain warnings to merchants based on their adaptive warning methods.

[0005] Furthermore, the sales history of merchants is collected based on artificial intelligence, and the merchant's warnable products obtained based on the sales history include: Based on artificial intelligence, the sales history of multiple branches and online stores opened by a single merchant is collected; the products currently placed by the merchant in the sales channel are obtained as the current candidate products; the placement area of ​​the current candidate products is used as the candidate placement area; The sales history related to the current candidate product is obtained from the sales history collected by artificial intelligence, and used as the candidate sales history; the area with the largest population density in the candidate sales history is obtained as the candidate potential area; The distance between the candidate launch area and the candidate potential area is used as the warning effective distance; The sales history collected by artificial intelligence that is beyond the effective warning distance from the area where the sales history is collected by artificial intelligence is filtered out, and the products in the other sales history are all treated as products that can be warned.

[0006] Furthermore, a product association binary tree is generated based on the merchant's warnable products, the merchant's sales cycle and sales area segment are obtained based on the sales history collected by artificial intelligence, and the merchant's sales expectation is obtained. The merchant's adaptive warning method is generated based on the sales expectation and the product association binary tree, including: Use the mind map processing method to process the merchant's alertable products, and generate a product association binary tree based on the processing log; Process each product that can be warned in the sales history collected by artificial intelligence, and obtain the sales cycle and sales area segment of each product that can be warned based on the processing log; use the sales expectation algorithm to obtain the sales expectation of the merchant based on the sales cycle and sales area segment of each product that can be warned; An adaptive early warning method for merchants is generated based on the merchant's sales expectations and the commodity association binary tree.

[0007] Furthermore, the mind mapping processing method includes: Generate a binary tree of T rows × Y columns as an association processing binary tree, wherein the top row of the association processing binary tree is filled with all the products that can be warned from left to right except the leftmost layer, and the leftmost side of the association processing binary tree is filled with all the regional leaf nodes where all the products that can be warned were sold in the sales history except the first layer from top to bottom, and the regional leaf nodes in the previous layer are earlier than the regional leaf nodes in the next layer; According to artificial intelligence, the area span with the most used area segments in the mind map association curve is obtained, and the area spans are sequentially used as the associated area span 1 to the associated area span L from small to large, wherein the associated area span L is less than or equal to the warning effective distance; For any alertable product in the associated processing binary tree, multiple regional leaf nodes corresponding to the alertable product in the associated processing binary tree are sequentially used as product regional leaf node 1 to product regional leaf node M from top to bottom; the distances between all two adjacent product regional leaf nodes from product regional leaf node 1 to product regional leaf node M are obtained and sequentially used as associated distance span 1 to associated distance span M-1.

[0008] Furthermore, the mind map processing method also includes: When M-1 is not less than L, all associated distance ranges from the associated distance span 1 to the associated distance span M-1 are obtained, and the associated distance range is a range consisting of the associated distance span M1+1 to the associated distance span M1+L, where M1 is a positive integer less than or equal to M-1-L and not less than 0; For any correlation distance range, the correlation coefficient of the correlation distance range is obtained using the range processing algorithm. The range processing algorithm is: , where F is the correlation coefficient, Gi is the i-th correlation area span from correlation area span 1 to correlation area span L, Hi is the i-th correlation distance span in the correlation distance range, Hs is the sum of all correlation area spans, and Hz is the sum of all correlation distance spans corresponding to the correlation distance range; Obtain the correlation coefficients corresponding to all correlation distance ranges, and use the maximum value of all correlation coefficients as the correlation value of the product that can be warned; When M-1 is less than L, the values ​​of associated distance span 1 to associated distance span M-1 are subtracted from associated area span 1 to associated area span M-1 respectively as associated differences WC1 to associated differences WCM-1. When all associated differences are not less than 0, the associated value of the product that can be warned is set to 0; when any associated difference is less than 0, the associated value of the product that can be warned is set to δ; Obtain the correlation values ​​of all the products that can be warned, and take the average of all the correlation values ​​whose correlation values ​​are numbers as the correlation mean; Generate a binary tree of R layers × U columns as the commodity association binary tree, where the top row of the commodity association binary tree is filled with the association value at the rightmost layer, and the leftmost layer of the commodity association binary tree is filled with all the warning commodities with an association value of δ from top to bottom except the first layer, and all the warning commodities are filled in from large to small according to the association value under the warning commodity with an association value of δ, and the association value of each warning commodity is filled in the corresponding layer in the binary tree.

[0009] Furthermore, each of the warning commodities in the sales history collected by the artificial intelligence is processed, and the sales cycle and sales area segment of each warning commodity are obtained according to the processing log; the sales expectation algorithm is used to obtain the sales expectation of the merchant according to the sales cycle and sales area segment of each warning commodity, including: The total sales area segments of all the warnable products in the sales history collected by artificial intelligence are obtained as SalesAll; for any warnable product, the sales area segment of the merchant for the warnable product is taken as Pz, and the value of Pz divided by SalesAll is taken as the sales cycle of the warnable product; the sales area segments and sales cycles of all warnable products are obtained, and the sales expectation algorithm is used to obtain the sales expectation of the merchant. The sales expectation algorithm includes: , where M is the sales expectation, Pzj is the j-th sales area segment, PLj is the j-th sales cycle, and N is the number of products that can be warned.

[0010] Furthermore, the adaptive early warning method includes: The value of the merchant's sales expectation rounded up is used as the standard sales number; when the merchant sells goods, all the warnable goods in the goods-related binary tree are pushed to the merchant. When the push area segment of all the warnable goods in the goods-related binary tree is equal to the standard sales number, all the warnable goods with a correlation value of 0 are removed from the goods-related binary tree; when the push area segment of all the warnable goods in the goods-related binary tree exceeds L, only the warnable goods with a correlation value of δ and a correlation value exceeding the correlation mean are screened out in the goods-related binary tree; when the total area of ​​the warnable goods in the push-related binary tree exceeds the warning effective distance, only the warnable goods with a correlation value of δ are screened out in the goods-related binary tree; When there are only alertable products with an association value of δ in the product association binary tree, the effective warning distance is obtained again and the product association binary tree and sales expectations are obtained again based on the mind map processing method; after obtaining the new product association binary tree and new sales expectations, the use object of the adaptive early warning method is transferred from the product association binary tree of only alertable products with an association value of δ to the new product association binary tree, and at the same time, it is pushed according to the new sales expectations within the adaptive early warning method.

[0011] According to the second aspect of the present invention, the present invention claims protection for a supply chain risk early warning device based on artificial intelligence, which is applicable to the above-mentioned supply chain risk early warning method based on artificial intelligence, including an early warning collection unit, an adaptive method generation unit and a supply chain early warning unit: The early warning collection unit is used to collect the sales history of merchants based on artificial intelligence, and obtain the merchants' early warning products based on the sales history; The method generation unit is used to generate a product association binary tree based on the merchant's warning products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectation, and generate the merchant's adaptive warning method based on the sales expectation and the product association binary tree; The supply chain early warning unit is used to provide supply chain early warning to merchants based on the merchant's adaptive early warning method.

[0012] The present application relates to the field of artificial intelligence technology, and in particular to a supply chain risk early warning method and device based on artificial intelligence, which collects the sales history of merchants based on artificial intelligence, and obtains the merchants' early warning commodities based on the sales history; generates a commodity-related binary tree based on the merchants' early warning commodities, obtains the merchants' sales cycle and sales area segments based on the sales history collected by artificial intelligence, and obtains the merchants' sales expectations, generates the merchants' adaptive early warning method based on the sales expectations and the commodity-related binary tree; and performs supply chain early warning on merchants based on the merchants' adaptive early warning method. The present invention can generate differentiated artificial intelligence risk management strategies according to the actual deviation of different supply chain data, and fully ensure the reliability of sales data. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A structural module diagram of a supply chain risk early warning device based on artificial intelligence as claimed in an embodiment of the present application; Figure 2 A workflow diagram of an artificial intelligence-based supply chain risk early warning method for which protection is requested in an embodiment of the present application. DETAILED DESCRIPTION

[0014] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field without creative work based on the embodiments in the present application are within the scope of protection of this application.

[0015] The terms "first", "second" and "third" in this application are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" can explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of this application are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or equipment.

[0016] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0017] Example 1, please refer to Figure 1As shown, in the first aspect, the present application provides a supply chain risk early warning device based on artificial intelligence, including an early warning collection unit, an adaptive method generation unit and a supply chain early warning unit: The early warning collection unit is used to collect the sales history of merchants based on artificial intelligence, and obtain the merchants' early warning products based on the sales history; The early warning collection unit is configured with a strategy for obtaining products that can be warned. The strategy for obtaining products that can be warned includes: Based on artificial intelligence, the sales history of multiple branches and online stores opened by a single merchant is collected; the products currently placed by the merchant in the sales channel are obtained as the current candidate products; the placement area of ​​the current candidate products is used as the candidate placement area; The sales history related to the current candidate product is obtained from the sales history collected by artificial intelligence, and used as the candidate sales history; the area with the largest population density in the candidate sales history is obtained as the candidate potential area; In the specific implementation process, when the warning effective distance is short, the current candidate product can be replaced and the warning effective distance can be obtained again, so that the processing data when processing according to the warning effective distance is more candidate, and the judgment of whether the warning effective distance is short can be set according to the actual situation; The distance between the candidate launch area and the candidate potential area is used as the warning effective distance; The sales history collected by artificial intelligence that is beyond the effective warning distance from the area where the sales history is collected by artificial intelligence is filtered out, and the products in the other sales history are all treated as products that can be warned.

[0018] The method generation unit is used to generate a product association binary tree based on the merchant's warning products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectation, and generate the merchant's adaptive warning method based on the sales expectation and the product association binary tree; The method generation unit is configured with a mind map processing strategy, a sales expectation obtaining strategy, and an early warning method generation strategy; The mind map processing strategy is used to process the merchant's alertable products using the mind map processing method, and generate a product association binary tree based on the processing log; the mind map processing method includes: Generate a binary tree of T rows × Y columns as an association processing binary tree, wherein the top row of the association processing binary tree is filled with all the products that can be warned from left to right except the leftmost layer, and the leftmost side of the association processing binary tree is filled with all the regional leaf nodes where all the products that can be warned were sold in the sales history except the first layer from top to bottom, and the regional leaf nodes in the previous layer are earlier than the regional leaf nodes in the next layer; According to artificial intelligence, the area span with the most used area segments in the mind map association curve is obtained, and the area spans are sequentially used as the associated area span 1 to the associated area span L from small to large, wherein the associated area span L is less than or equal to the warning effective distance; In specific implementation, specific settings can be made based on the area span with the most used area segments in the mind map curve obtained by artificial intelligence in actual situations; For any alertable product in the associated processing binary tree, multiple regional leaf nodes corresponding to the alertable product in the associated processing binary tree are sequentially used as product regional leaf node 1 to product regional leaf node M from top to bottom; the distances between all two adjacent product regional leaf nodes from product regional leaf node 1 to product regional leaf node M are obtained and sequentially used as associated distance span 1 to associated distance span M-1.

[0019] Mind mapping also includes: When M-1 is not less than L, all associated distance ranges from the associated distance span 1 to the associated distance span M-1 are obtained, and the associated distance range is a range consisting of the associated distance span M1+1 to the associated distance span M1+L, where M1 is a positive integer less than or equal to M-1-L and not less than 0; For any correlation distance range, the correlation coefficient of the correlation distance range is obtained using the range processing algorithm. The range processing algorithm is: , where F is the correlation coefficient, Gi is the i-th correlation area span from correlation area span 1 to correlation area span L, Hi is the i-th correlation distance span in the correlation distance range, Hs is the sum of all correlation area spans, and Hz is the sum of all correlation distance spans corresponding to the correlation distance range; Obtain the correlation coefficients corresponding to all correlation distance ranges, and use the maximum value of all correlation coefficients as the correlation value of the product that can be warned; When M-1 is less than L, the values ​​of associated distance span 1 to associated distance span M-1 are subtracted from associated area span 1 to associated area span M-1 respectively as associated differences WC1 to associated differences WCM-1. When all associated differences are not less than 0, the associated value of the product that can be warned is set to 0; when any associated difference is less than 0, the associated value of the product that can be warned is set to δ; In the specific implementation process, when all the correlation differences are not less than 0, it means that the merchant has a high sales cycle for the warnable product, so the correlation value should be set to 0 to reduce the priority of the warning, thereby increasing the priority of other warnable products that should be warned; Obtain the correlation values ​​of all the products that can be warned, and take the average of all the correlation values ​​whose correlation values ​​are numbers as the correlation mean; Generate a binary tree of R layers × U columns as the commodity association binary tree, where the top row of the commodity association binary tree is filled with the association value at the rightmost layer, and the leftmost layer of the commodity association binary tree is filled with all the warning commodities with an association value of δ from top to bottom except the first layer, and all the warning commodities are filled in from large to small according to the association value under the warning commodity with an association value of δ, and the association value of each warning commodity is filled in the corresponding layer in the binary tree.

[0020] The sales expectation acquisition strategy is used to process each product that can be warned in the sales history collected by artificial intelligence, and obtain the sales cycle and sales area segment of each product that can be warned based on the processing log; the sales expectation algorithm is used to obtain the sales expectation of the merchant based on the sales cycle and sales area segment of each product that can be warned; the sales expectation acquisition strategy includes: The total sales area segments of all the warnable products in the sales history collected by artificial intelligence are obtained as SalesAll; for any warnable product, the sales area segment of the merchant for the warnable product is taken as Pz, and the value of Pz divided by SalesAll is taken as the sales cycle of the warnable product; the sales area segments and sales cycles of all warnable products are obtained, and the sales expectation algorithm is used to obtain the sales expectation of the merchant. The sales expectation algorithm includes: , where M is the sales expectation, Pzj is the j-th sales area segment, PLj is the j-th sales cycle, and N is the number of products that can be warned.

[0021] The early warning method generation strategy is used to generate the merchant's adaptive early warning method based on the merchant's sales expectations and the commodity association binary tree. The adaptive early warning method is: The merchant's sales forecast is rounded up to an integer as the standard sales number; When a merchant sells a product, all the products with warnings in the product-related binary tree are pushed to the merchant. When the push area segment of all the products with warnings in the product-related binary tree is equal to the standard sales number, all the products with warnings whose association value is 0 are removed from the product-related binary tree. When the push area segments of all the alertable products in the product association binary tree exceed L, only the alertable products with correlation values ​​of δ and correlation values ​​exceeding the correlation mean are screened out in the product association binary tree; when the total area of ​​the alertable products in the push association binary tree exceeds the effective warning distance, only the alertable products with correlation values ​​of δ are screened out in the product association binary tree; When there are only alertable products with an association value of δ in the product association binary tree, the effective warning distance is obtained again and the product association binary tree and sales expectations are obtained again based on the mind map processing method; after obtaining the new product association binary tree and new sales expectations, the use object of the adaptive early warning method is transferred from the product association binary tree of only alertable products with an association value of δ to the new product association binary tree, and at the same time, it is pushed according to the new sales expectations within the adaptive early warning method.

[0022] The supply chain early warning unit is used to provide supply chain early warning to merchants based on the merchant's adaptive early warning method.

[0023] Example 2, please refer to Figure 2 As shown, in the second aspect, the present invention provides a supply chain risk warning method based on artificial intelligence, including: step S1, collecting the sales history of the merchant based on artificial intelligence, and obtaining the merchant's warning products based on the sales history.

[0024] Step S1 includes the following sub-steps: Step S101, collecting the sales history of multiple branches opened by a single merchant and online based on artificial intelligence; obtaining the commodities currently placed by the merchant in the sales channel as the current candidate commodities; and taking the placement area of ​​the current candidate commodities as the candidate placement area; The sales history related to the current candidate product is obtained from the sales history collected by artificial intelligence, and used as the candidate sales history; the area with the largest population density in the candidate sales history is obtained as the candidate potential area; The distance between the candidate launch area and the candidate potential area is used as the warning effective distance; Step S102, filtering out the sales history collected by artificial intelligence that is beyond the effective warning distance from the area where the sales history is collected by artificial intelligence, and treating all other commodities in the sales history as commodities that can be warned.

[0025] Step S2, generate a product-related binary tree based on the merchant's alertable products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectations, and generate the merchant's adaptive early warning method based on the sales expectations and the product-related binary tree.

[0026] Step S2 includes the following sub-steps: Step S201, use the mind map processing method to process the merchant's warning products, and generate a product association binary tree based on the processing log.

[0027] The mind map processing method includes: step S2011, generating a binary tree of T rows×Y columns as an association processing binary tree, wherein all the products that can be warned are sequentially filled in from left to right in the top row of the association processing binary tree except the leftmost layer, and all the regional leaf nodes where all the products that can be warned were sold in the sales history are sequentially filled in from top to bottom in the leftmost layer of the association processing binary tree except the first layer, and the regional leaf nodes in the previous layer are earlier than the regional leaf nodes in the next layer; Step S2012, obtaining the area span with the most used area segments in the mind map association curve according to artificial intelligence, and sequentially using the area spans from small to large as the associated area span 1 to the associated area span L, wherein the associated area span L is less than or equal to the warning effective distance; Step S2013, for any alertable product in the associated processing binary tree, multiple area leaf nodes corresponding to the alertable product in the associated processing binary tree are sequentially used as product area leaf node 1 to product area leaf node M from top to bottom; the distances between all two adjacent product area leaf nodes from product area leaf node 1 to product area leaf node M are obtained, and used as the associated distance span 1 to the associated distance span M-1, respectively.

[0028] Mind mapping also includes: Step S2014, when M-1 is not less than L, all associated distance ranges from the associated distance span 1 to the associated distance span M-1 are obtained, and the associated distance range is a range consisting of the associated distance span M1+1 to the associated distance span M1+L, wherein M1 is a positive integer less than or equal to M-1-L and not less than 0; Step S2015: for any associated distance range, a range processing algorithm is used to obtain the associated coefficient of the associated distance range. The range processing algorithm is: , where F is the correlation coefficient, Gi is the i-th correlation area span from correlation area span 1 to correlation area span L, Hi is the i-th correlation distance span in the correlation distance range, Hs is the sum of all correlation area spans, and Hz is the sum of all correlation distance spans corresponding to the correlation distance range; Step S2016, obtaining the correlation coefficients corresponding to all correlation distance ranges, and taking the maximum value of all correlation coefficients as the correlation value of the product that can be warned; Step S2017, when M-1 is less than L, the values ​​of the associated distance span 1 to the associated distance span M-1 are respectively subtracted from the associated area span 1 to the associated area span M-1 as the associated differences WC1 to the associated differences WCM-1, and when all the associated differences are not less than 0, the associated value of the product that can be warned is set to 0; when any of the associated differences is less than 0, the associated value of the product that can be warned is set to δ; Step S2018, obtaining the association values ​​of all the products that can be warned, and taking the average of all the association values ​​whose association values ​​are numbers as the association mean; Step S2019, generate a binary tree of R layers × U columns as a product association binary tree, wherein the rightmost layer of the top row of the product association binary tree is filled with the association value, and the leftmost layer of the product association binary tree is filled with all the warning products with an association value of δ from top to bottom except the first layer, and all the warning products are filled in from large to small according to the association value below the warning product with an association value of δ, and the association value of each warning product is filled in the corresponding level in the binary tree.

[0029] Step S202, process each of the warning commodities in the sales history collected by artificial intelligence, and obtain the sales cycle and sales area segment of each of the warning commodities based on the processing log; use the sales expectation algorithm to obtain the sales expectation of the merchant based on the sales cycle and sales area segment of each of the warning commodities.

[0030] Step S202 includes: obtaining the total sales area segments of all the warnable commodities in the sales history collected by the artificial intelligence as SalesAll; for any warnable commodity, taking the sales area segment of the merchant for the warnable commodity as Pz, and dividing Pz by the value of SalesAll as the sales cycle of the warnable commodity; obtaining the sales area segments and sales cycles of all the warnable commodities, and using the sales expectation algorithm to obtain the sales expectation of the merchant, the sales expectation algorithm including: , where M is the sales expectation, Pzj is the j-th sales area segment, PLj is the j-th sales cycle, and N is the number of products that can be warned.

[0031] Step S203, generating a merchant's adaptive early warning method based on the merchant's sales expectations and the commodity association binary tree.

[0032] The adaptive early warning method includes: step S2031, taking the value of the merchant's sales expectation rounded up as the standard sales number; when the merchant sells the goods, all the early warning goods in the goods-related binary tree are pushed to the merchant, and when the push area segment of all the early warning goods in the goods-related binary tree is equal to the standard sales number, all the early warning goods with a correlation value of 0 are removed from the goods-related binary tree; when the push area segment of all the early warning goods in the goods-related binary tree exceeds L, only the early warning goods with a correlation value of δ and a correlation value exceeding the correlation mean are screened out in the goods-related binary tree; when the total area of ​​the early warning goods in the push-related binary tree exceeds the early warning effective distance, only the early warning goods with a correlation value of δ are screened out in the goods-related binary tree; Step S2032, when there are only alertable products with an association value of δ in the product association binary tree, the warning effective distance is obtained again and the product association binary tree and sales expectations are obtained again based on the mind map processing method; after obtaining the new product association binary tree and the new sales expectations, the use object of the adaptive early warning method is transferred from the product association binary tree of only alertable products with an association value of δ to the new product association binary tree, and at the same time, it is pushed according to the new sales expectations within the adaptive early warning method.

[0033] Step S3, providing supply chain warning to merchants according to the merchant's adaptive warning method.

[0034] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0035] In the several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0036] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a branch functional unit. The above is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the content of the specification and drawings of this application, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present application.

[0037] The specific implementation methods of the invention are described in detail above, but they are only examples, and the present application is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application, and therefore, the equalization, modification, and improvement made without departing from the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A supply chain risk early warning method based on artificial intelligence, characterized in that: include: Collect the sales history of merchants based on artificial intelligence, and obtain the merchants' early warning products based on the sales history; Generate a binary tree of product associations based on the merchant's warnable products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectations, and generate an adaptive warning method for the merchant based on the sales expectations and the binary tree of product associations; Provide supply chain warnings to merchants based on their adaptive warning methods.

2. The supply chain risk early warning method based on artificial intelligence as claimed in claim 1, characterized in that: The sales history of merchants is collected based on artificial intelligence, and the merchants' warning products based on the sales history include: Based on artificial intelligence, the sales history of multiple branches and online stores opened by a single merchant is collected; the products currently placed by the merchant in the sales channel are obtained as the current candidate products; the placement area of ​​the current candidate products is used as the candidate placement area; The sales history related to the current candidate product is obtained from the sales history collected by artificial intelligence, and used as the candidate sales history; the area with the largest population density in the candidate sales history is obtained as the candidate potential area; The distance between the candidate launch area and the candidate potential area is used as the warning effective distance; The sales history collected by artificial intelligence that is beyond the effective warning distance from the area where the sales history is collected by artificial intelligence is filtered out, and the products in the other sales history are all treated as products that can be warned.

3. The supply chain risk early warning method based on artificial intelligence as claimed in claim 2, characterized in that: Generate a binary tree of product associations based on the merchant's warnable products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectations. Generate a merchant's adaptive warning method based on sales expectations and the binary tree of product associations, including: Use the mind map processing method to process the merchant's alertable products, and generate a product association binary tree based on the processing log; Process each product that can be warned in the sales history collected by artificial intelligence, and obtain the sales cycle and sales area segment of each product that can be warned based on the processing log; use the sales expectation algorithm to obtain the sales expectation of the merchant based on the sales cycle and sales area segment of each product that can be warned; An adaptive early warning method for merchants is generated based on the merchant's sales expectations and the commodity association binary tree.

4. The supply chain risk early warning method based on artificial intelligence as claimed in claim 3, characterized in that: The mind map processing method includes: Generate a binary tree of T rows × Y columns as an association processing binary tree, wherein the top row of the association processing binary tree is filled with all the products that can be warned from left to right except the leftmost layer, and the leftmost side of the association processing binary tree is filled with all the regional leaf nodes where all the products that can be warned were sold in the sales history except the first layer from top to bottom, and the regional leaf nodes in the previous layer are earlier than the regional leaf nodes in the next layer; According to artificial intelligence, the area span with the most used area segments in the mind map association curve is obtained, and the area spans are sequentially used as the associated area span 1 to the associated area span L from small to large, wherein the associated area span L is less than or equal to the warning effective distance; For any alertable product in the associated processing binary tree, multiple regional leaf nodes corresponding to the alertable product in the associated processing binary tree are sequentially used as product regional leaf node 1 to product regional leaf node M from top to bottom; the distances between all two adjacent product regional leaf nodes from product regional leaf node 1 to product regional leaf node M are obtained and sequentially used as associated distance span 1 to associated distance span M-1.

5. The supply chain risk early warning method based on artificial intelligence as claimed in claim 4, characterized in that: The mind map processing method also includes: When M-1 is not less than L, all associated distance ranges from the associated distance span 1 to the associated distance span M-1 are obtained, and the associated distance range is a range consisting of the associated distance span M1+1 to the associated distance span M1+L, where M1 is a positive integer less than or equal to M-1-L and not less than 0; For any correlation distance range, the correlation coefficient of the correlation distance range is obtained using the range processing algorithm. The range processing algorithm is: , where F is the correlation coefficient, Gi is the i-th correlation area span from correlation area span 1 to correlation area span L, Hi is the i-th correlation distance span in the correlation distance range, Hs is the sum of all correlation area spans, and Hz is the sum of all correlation distance spans corresponding to the correlation distance range; Obtain the correlation coefficients corresponding to all correlation distance ranges, and use the maximum value of all correlation coefficients as the correlation value of the product that can be warned; When M-1 is less than L, the values ​​of associated distance span 1 to associated distance span M-1 are subtracted from associated area span 1 to associated area span M-1 respectively as associated differences WC1 to associated differences WCM-1. When all associated differences are not less than 0, the associated value of the product that can be warned is set to 0; when any associated difference is less than 0, the associated value of the product that can be warned is set to δ; Obtain the correlation values ​​of all the products that can be warned, and take the average of all the correlation values ​​whose correlation values ​​are numbers as the correlation mean; Generate a binary tree of R layers × U columns as the commodity association binary tree, where the top row of the commodity association binary tree is filled with the association value at the rightmost layer, and the leftmost layer of the commodity association binary tree is filled with all the warning commodities with an association value of δ from top to bottom except the first layer, and all the warning commodities are filled in from large to small according to the association value under the warning commodity with an association value of δ, and the association value of each warning commodity is filled in the corresponding layer in the binary tree.

6. The supply chain risk early warning method based on artificial intelligence as claimed in claim 5, characterized in that: Process each product that can be warned in the sales history collected by artificial intelligence, and obtain the sales cycle and sales area segment of each product that can be warned based on the processing log; use the sales expectation algorithm based on the sales cycle and sales area segment of each product that can be warned to obtain the sales expectations of the merchant, including: The total sales area segments of all the warnable products in the sales history collected by artificial intelligence are obtained as SalesAll; for any warnable product, the sales area segment of the merchant for the warnable product is taken as Pz, and the value of Pz divided by SalesAll is taken as the sales cycle of the warnable product; the sales area segments and sales cycles of all warnable products are obtained, and the sales expectation algorithm is used to obtain the sales expectation of the merchant. The sales expectation algorithm includes: , where M is the sales expectation, Pzj is the j-th sales area segment, PLj is the j-th sales cycle, and N is the number of products that can be warned.

7. The supply chain risk early warning method based on artificial intelligence as claimed in claim 6, characterized in that: Adaptive early warning methods include: The value of the merchant's sales expectation rounded up is used as the standard sales number; when the merchant sells goods, all the warnable goods in the goods-related binary tree are pushed to the merchant. When the push area segment of all the warnable goods in the goods-related binary tree is equal to the standard sales number, all the warnable goods with a correlation value of 0 are removed from the goods-related binary tree; when the push area segment of all the warnable goods in the goods-related binary tree exceeds L, only the warnable goods with a correlation value of δ and a correlation value exceeding the correlation mean are screened out in the goods-related binary tree; when the total area of ​​the warnable goods in the push-related binary tree exceeds the warning effective distance, only the warnable goods with a correlation value of δ are screened out in the goods-related binary tree; When there are only alertable products with an association value of δ in the product association binary tree, the effective warning distance is obtained again and the product association binary tree and sales expectations are obtained again based on the mind map processing method; after obtaining the new product association binary tree and new sales expectations, the use object of the adaptive early warning method is transferred from the product association binary tree of only alertable products with an association value of δ to the new product association binary tree, and at the same time, it is pushed according to the new sales expectations within the adaptive early warning method.

8. A supply chain risk early warning device based on artificial intelligence, applicable to the supply chain risk early warning method based on artificial intelligence according to any one of claims 1 to 7, characterized in that: It includes early warning collection unit, adaptive method generation unit and supply chain early warning unit: The early warning collection unit is used to collect the sales history of merchants based on artificial intelligence, and obtain the merchants' early warning products based on the sales history; The method generation unit is used to generate a product association binary tree based on the merchant's warning products, obtain the merchant's sales cycle and sales area segment based on the sales history collected by artificial intelligence, and obtain the merchant's sales expectation, and generate the merchant's adaptive warning method based on the sales expectation and the product association binary tree; The supply chain early warning unit is used to provide supply chain early warning to merchants based on the merchant's adaptive early warning method.

Images