An import and export trade management system and method based on artificial intelligence

Abstract

The present invention relates to the technical field of import and export trade, and specifically to an import and export trade management system and method based on artificial intelligence, comprising an inspection process data analysis module, a sampling method difference determination module, an attribute matching package construction module, a selection priority analysis module, a real-time monitoring analysis module and an early warning reminder module; the inspection process data analysis module is used to obtain the inspection process data recorded by the import and export trade management platform; the sampling method difference determination module is used to analyze whether the sampling methods to be analyzed are all the same; the matching attribute package construction module is used to construct an attribute matching package of a first root node and a first sampling method; the selection priority analysis module is used to analyze the selection priority of the root node; the real-time monitoring analysis module is used to obtain the name of the imported and exported goods under the real-time monitoring state, and match it with the attribute matching package; the early warning reminder module is used to transmit an early warning of the sampling method of the imported and exported goods before the inspection process starts.

Description

Technical Field

[0001] The present invention relates to the technical field of import and export trade, and specifically to an import and export trade management system and method based on artificial intelligence. Background Art

[0002] Import and export trade refers to the transaction of goods and services across national borders, which consists of import trade and export trade. Import and export trade can adjust the utilization rate of domestic production factors, adjust the economic structure and increase fiscal revenue. There are four main links in the inspection of import and export commodities in my country, including acceptance of application, sampling, inspection and issuance of certificates. In the sampling link, the existing technology is mostly manual sampling inspection, and there is no clear regulation on the sampling method of import and export goods. Different import and export trade status and different sampling inspection staff may have multiple sampling methods for the same type of goods.

[0003] As a result, in actual applications, there is no unified analysis standard, and there is no prediction of the selection of sampling methods before the inspection stage begins; it is impossible to effectively monitor the digital efficiency of imported and exported goods in the inspection process, and it is impossible to effectively analyze and warn of the possible impact when the sampling method produces abnormalities before the sampling stage. Summary of the invention

[0004] The purpose of the present invention is to provide an import and export trade management system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: an import and export trade management method based on artificial intelligence, the method comprising the following analysis steps:

[0006] Step S1: Obtain the inspection process data recorded in the import and export trade management platform. The inspection process data is a data package of the entire process of goods inspection. The inspection process data refers to the attribute characteristic data recorded before the start of the sampling inspection link to the completion of the inspection process of the imported and exported goods; the attribute characteristics include the name of the imported and exported goods, the location data corresponding to the link, the time data corresponding to the link, the sampling method of the goods and the sampling results; the sampling results include qualified and unqualified;

[0007] Step S2: Taking the name of the imported and exported goods as the root node, extract the sampling methods corresponding to the same root node in different data packets in the inspection process data as the sampling methods to be analyzed. If the sampling methods to be analyzed are the same, the transmission root node is the first root node, the sampling method to be analyzed is the first sampling method, and construct an attribute matching package A of the first root node and the first sampling method, A = {first root node → first sampling method}; the attribute matching package refers to the sampling method corresponding to the imported and exported goods; if there are differences in the sampling methods to be analyzed, the selection priority of the root node is analyzed based on the data packet to which the root node belongs;

[0008] Step S3: Based on the selection priority of the imported and exported goods, the names of the imported and exported goods under real-time monitoring are obtained; when the names of the imported and exported goods exist in the attribute matching package, the first sampling method in the corresponding attribute matching package is transmitted before the inspection process starts for early warning reminders; the purpose of early warning for the imported and exported goods in the attribute matching package is to quickly and intelligently warn and remind the work before the inspection process starts in the cumbersome and complex import and export goods management process, so that relevant personnel can make adaptive preparations for the inspection process and reduce the error rate of the inspection work after the inspection process starts; when the name of the imported and exported goods does not exist in the attribute matching package, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed;

[0009] Step S4: Setting a similarity index threshold, based on the relationship between the similarity index threshold and the similarity index, transmitting an early warning of the sampling method for import and export goods before the inspection process begins.

[0010] Furthermore, in step S2, the selection priority of the root node is analyzed based on the data packets to which the root node belongs, including the following analysis steps:

[0011] The inspection process data corresponding to the root nodes with different sampling methods to be analyzed are used to construct a data package to be analyzed, and the data package to be analyzed is composed of a root node and the inspection process data corresponding to m sampling methods; the root node in the data package to be analyzed is the target root node, and the sampling method and the corresponding inspection process data in the data package to be analyzed are the target sub-data package;

[0012] Get the time data corresponding to the link in the i-th target sub-data packet of the target root node, extract the sampling time T of the i-th target sub-data packet of the target root node to record the sampling link 1i And the interval time T from completing the sampling process to obtaining the sampling results 2i ; Using the formula: t i =T 1i / H1+T 2i / H2, calculate the time efficiency coefficient t of the i-th target sub-data packet of the target root node i ; Where i≤m; H1, H2 represent the average sampling duration and the average interval duration of the m target sub-data packets contained in the target root node; H1=1 / m(∑T 1i ), H2=1 / m(∑T 2i ); The analysis of the time efficiency coefficient reflects that the process consumption time of the same target root node corresponding to different sampling links is different. The shorter the time, the higher the efficiency and more convenient the sampling method. Therefore, in this application, the time impact is used as one of the factors for analyzing the selection and judgment of the sampling method;

[0013] Mark the data packets with unqualified sampling results in the m target sub-data packets of the target root node as abnormal data packets; obtain the total frequency n of the abnormal data packets in the corresponding record data packets of the target root node, and calculate the unqualified rate of the target root node corresponding to the i-th target sub-data packet as r i , r i =(c i / m i )*(c i / n); where m i represents the total number of times the i-th target sub-packet is recorded; c i Indicates the number of times the i-th target sub-packet is recorded as an abnormal packet;

[0014] Get the position data corresponding to the link in the abnormal data packet. When the position of the sampling link is different from the position of the forward link that needs to be returned after the sampling abnormality, the mark feature value is 1. When the position of the sampling link is the same as the position of the forward link that needs to be returned after the sampling abnormality, the mark feature value is 0. The forward link refers to the process link before the sampling link recorded in chronological order according to the inspection process data. Use the formula: g i =0.45*r i +0.55*a, calculate the correlation influence coefficient gi of the i-th target sub-data packet of the target root node; a={0,1}.

[0015] The difference between the position of the sampling link in the abnormal data packet and the position of the forward link that needs to be returned after the sampling abnormality is analyzed. This is because after the sampling fails, it is necessary to return to the previous link such as the packaging and packaging link to replace the unqualified products for re-inspection. Then there will be difficulty in returning to the forward link after the goods fail to meet the standards. When the sampling of goods is carried out during the packaging of the product, if an abnormality occurs here, it can be stopped immediately, and there is no need to move the goods, which reduces the complexity of solving problems when encountering difficulties in import and export trade and saves the import and export process time.

[0016] Furthermore, step S2 also includes the following analysis steps:

[0017] Get the time efficiency coefficient t of the i-th target sub-packet i And the correlation coefficient g i , using the formula:

[0018] V i =k1*t i +k2*g i

[0019] Calculate the priority index V of the i-th target sub-data packet of the target root node i, where k1 and k2 represent the reference coefficient corresponding to the timeliness coefficient and the reference coefficient corresponding to the correlation influence coefficient respectively; k1+k2=1,0 <k1、k2<1;

[0020] The priority indices corresponding to the m target sub-data packets of the target root node are arranged in ascending order, and a selection priority corresponding to the target root node is generated, and the selection priority of the selection priority is the same as the priority arrangement order.

[0021] Furthermore, in step S3, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed, including the following analysis steps:

[0022] Get the m corresponding to the i-th sampling method under the target root node i target sub-packet, extract the mth i The location data corresponding to the target sub-packets are sorted in chronological order to generate a location set b mi ;

[0023] M i The location set b corresponding to the target sub-packets mi Carry out b m1 ∩b m2 ∩b m3 ......∩b mi Operation, get the intersection position, and output the intersection position as the reliable position w of the i-th sampling method i ;

[0024] The intersection operation of the location set in the target sub-data package is performed on the consideration that even if the sampling methods are different, the inspected goods are the same, so there must be the same process links. The analysis of the same process links is limited to the real-time monitoring state to ensure the consistency of the general direction of the imported and exported goods and the historical data analysis of the same goods.

[0025] Obtain the location data f1 of the import and export goods in the real-time monitoring state after they enter the inspection process but before the sampling phase begins, and extract the location data p1 of the first target sub-data package in the selection priority;

[0026] Then calculate the position data f1 and the reliable position w i , the similarity index e{f1→w i ,p1},e{f1→w i ,p1}=(e f +e p ) / (e0+e1); where e f Represents the position data f1 and the reliable position w i The number of identical positions when comparing, e pIt indicates the number of identical positions when comparing position data f1 and position data p1. e0 indicates the existence of a reliable position w i , e1 represents the total number of position data in the first target sub-data packet.

[0027] The difference in the location data of different target sub-packets in the analysis and selection priority is to determine the direction of the links of the real-time monitored import and export goods in a digital way, because the selection of the sampling method can be obtained from the data analysis of each link. For example, if the warehouse capacity is tight, sampling cannot be carried out in the warehouse. It is reflected in the data that when entering the warehouse, it will be notified that there is no need to enter the warehouse, and the location of the link does not need to be changed. Then, when it corresponds to the location data of such a sampling link in the history, similar matching can be performed to achieve an early warning effect on the sampling method;

[0028] Further, step S4 includes the following analysis steps:

[0029] When the similarity index is greater than or equal to the similarity index threshold, the first sampling method warning signal in the priority level is selected when the name of the imported and exported goods in the transmission inspection process data is the same as that in the real-time import and export goods;

[0030] When the similarity index is less than the similarity index threshold, remove the first target sub-packet, obtain the position data in the target sub-packets in the selection priority in turn, and calculate the similarity index; extract the sampling method in the target sub-packet corresponding to the first similarity index greater than or equal to the similarity index threshold after calculation to perform transmission warning.

[0031] Import and export trade management system, including inspection process data analysis module, sampling method difference determination module, attribute matching package construction module, selection priority analysis module, real-time monitoring analysis module and early warning reminder module;

[0032] The inspection process data analysis module is used to obtain the inspection process data recorded in the import and export trade management platform. The inspection process data is a data package containing the entire process of goods inspection. The inspection process data refers to the attribute feature data recorded from the start of the sampling inspection link to the completion of the inspection process of imported and exported goods.

[0033] The sampling method difference determination module is used to take the name of the imported and exported goods as the root node, extract the sampling methods corresponding to the same root node in different data packets in the inspection process data as the sampling methods to be analyzed; and analyze whether the sampling methods to be analyzed are all the same;

[0034] The attribute matching package construction module is used for, when the sampling methods to be analyzed are the same, the transmission root node is the first root node, the sampling method to be analyzed is the first sampling method, and constructing the attribute matching package of the first root node and the first sampling method;

[0035] The selection priority analysis module is used to analyze the selection priority of the root node when the sampling methods to be analyzed are different;

[0036] The real-time monitoring and analysis module is used to obtain the names of import and export goods under real-time monitoring status and match them with the attribute matching package; if the match is successful, the signal is transmitted to the early warning reminder module, and if the match fails, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed;

[0037] The early warning reminder module is used to transmit an early warning of the sampling method of imported and exported goods before the inspection process begins based on the relationship between the similarity index threshold and the similarity index.

[0038] Further, the selection priority analysis module includes a data packet to be analyzed construction unit, a timeliness coefficient calculation unit, a correlation influence coefficient calculation unit, a priority index calculation unit and a selection priority generation unit;

[0039] The to-be-analyzed data packet construction unit is used to construct a data packet formed by a target root node and a target sub-data packet;

[0040] The time efficiency coefficient calculation unit is used to calculate the time efficiency coefficient based on the sampling time and the interval time;

[0041] The correlation influence coefficient calculation unit is used to calculate the correlation influence coefficient based on the unqualified rate and the characteristic value; the unqualified rate is calculated by the data corresponding to the sampling result in the target sub-data package when it is unqualified and the frequency of the target sub-data package record; the characteristic value refers to the characteristic number marked according to the difference between the position of the sampling link and the position of the forward link returned after the sampling is abnormal;

[0042] The priority index calculation unit calculates the priority index based on the timeliness coefficient and the correlation influence coefficient;

[0043] The selection priority generation unit is used to arrange the priority indexes corresponding to the target sub-data packets of the target root node in ascending order, and generate the selection priority corresponding to the target root node.

[0044] Furthermore, the real-time monitoring and analysis module includes a location set generation unit, a reliable location output unit, a location data acquisition unit, and a similarity index calculation unit;

[0045] The position set generating unit is used to extract the position data corresponding to the target sub-data packet and sort them in chronological order to generate a position set;

[0046] The reliable position output unit is used to perform an intersection operation on the position set corresponding to the target sub-data packet to obtain the intersection position, and output the intersection position as the reliable position corresponding to the sampling method;

[0047] The location data acquisition unit is used to obtain the location data of the import and export goods in the real-time monitoring state after they enter the inspection process but before the sampling phase begins, and the location data of the first target sub-data package in the selection priority;

[0048] The similarity index calculation unit is used to calculate the similarity index based on the position data and the reliable position.

[0049] Further, the early warning reminder module includes a threshold comparison unit, a feedback output unit and a transmission early warning unit;

[0050] The threshold comparison unit is used to compare the magnitude relationship between the similarity index and the similarity index threshold, and to transmit signals to the feedback transmission unit and the transmission warning unit;

[0051] The feedback output unit is used to transmit a signal to the real-time monitoring and analysis module to calculate a new similarity index based on the magnitude relationship of the output of the threshold comparison unit, and transmit a signal to the transmission warning unit when the similarity index is greater than or equal to the similarity index threshold;

[0052] The transmission warning unit is used to receive signals and perform transmission warning in a sampling manner.

[0053] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention extracts data of the same type of import and export goods with multiple sampling methods in historical data, analyzes the efficiency of multiple sampling methods, and thus makes a digital selection of sampling methods before sampling for real-time monitoring of import and export goods; at the same time, the present application analyzes the difference between the position of the sampling link in the abnormal data packet and the position of the forward link that needs to be returned after the sampling is abnormal, thereby reducing the complexity of solving problems when encountering difficulties in import and export trade and saving the process time of import and export; the difference in the position data of different target sub-data packets in the analysis and selection priority is to determine the direction of the link of the real-time monitored import and export goods in a digital way; the present application gives the selection priority of the sampling method in consideration of efficiency, convenience and the degree of impact when anomalies occur, and comprehensively recommends and warns the sampling method of import and export goods under real-time monitoring, so that through artificial intelligence calculation analysis, the import and export goods are more controllable and more digital in the inspection process, reducing the uncertainty of human influence. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0055] Figure 1 It is a structural schematic diagram of an import and export trade management system based on artificial intelligence of the present invention;

[0056] Figure 2It is a relationship diagram between a target root node and a target sub-data packet of an import and export trade management method based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] See also Figure 1-Figure 2 The present invention provides a technical solution: an import and export trade management method based on artificial intelligence, the method comprising the following analysis steps:

[0059] Step S1: Obtain the inspection process data recorded by the import and export trade management platform. The inspection process data is a data package of the entire process of goods inspection. The inspection process data refers to the attribute characteristic data recorded before the start of the sampling inspection link to the completion of the inspection process of the imported and exported goods; the attribute characteristics include the name of the imported and exported goods, the location data corresponding to the link, the time data corresponding to the link, the sampling method of the goods and the sampling results; the sampling results include qualified and unqualified; in actual applications, the judgment of qualified and unqualified sampling results is based on the qualified ratio corresponding to the current sampled goods, and there are differences between different goods;

[0060] Step S2: Taking the name of the imported and exported goods as the root node, extract the sampling methods corresponding to the same root node in different data packets in the inspection process data as the sampling methods to be analyzed. If the sampling methods to be analyzed are the same, the transmission root node is the first root node, the sampling method to be analyzed is the first sampling method, and construct an attribute matching package A of the first root node and the first sampling method, A = {first root node → first sampling method}; the attribute matching package refers to the sampling method corresponding to the imported and exported goods; if there are differences in the sampling methods to be analyzed, the selection priority of the root node is analyzed based on the data packet to which the root node belongs;

[0061] Step S3: Based on the selection priority of the imported and exported goods, the names of the imported and exported goods under real-time monitoring are obtained; when the names of the imported and exported goods exist in the attribute matching package, the first sampling method in the corresponding attribute matching package is transmitted before the inspection process starts for early warning reminders; the purpose of early warning for the imported and exported goods in the attribute matching package is to quickly and intelligently warn and remind the work before the inspection process starts in the cumbersome and complex import and export goods management process, so that relevant personnel can make adaptive preparations for the inspection process and reduce the error rate of the inspection work after the inspection process starts; when the name of the imported and exported goods does not exist in the attribute matching package, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed;

[0062] Step S4: Setting a similarity index threshold, based on the relationship between the similarity index threshold and the similarity index, transmitting an early warning of the sampling method for import and export goods before the inspection process begins.

[0063] In step S2, the selection priority of the root node is analyzed based on the data packets to which the root node belongs, and the analysis steps include the following:

[0064] The inspection process data corresponding to the root nodes with different sampling methods to be analyzed are used to construct a data package to be analyzed, and the data package to be analyzed is composed of a root node and inspection process data corresponding to m sampling methods; the root node in the data package to be analyzed is the target root node, and the sampling methods and corresponding inspection process data in the data package to be analyzed are target sub-data packages; each target sub-data package represents process data corresponding to a sampling method;

[0065] Get the time data corresponding to the link in the i-th target sub-data packet of the target root node, extract the sampling time T of the i-th target sub-data packet of the target root node to record the sampling link 1i And the interval time T from completing the sampling process to obtaining the sampling results 2i ; Using the formula: t i =T 1i / H1+T 2i / H2, calculate the time efficiency coefficient t of the i-th target sub-data packet of the target root node i ; Where i≤m; H1, H2 represent the average sampling duration and the average interval duration of the m target sub-data packets contained in the target root node; H1=1 / m(∑T 1i ), H2=1 / m(∑T 2i ); The analysis of the time efficiency coefficient reflects that the process consumption time of the same target root node corresponding to different sampling links is different. The shorter the time, the higher the efficiency and more convenient the sampling method. Therefore, in this application, the time impact is used as one of the factors for analyzing the selection and judgment of the sampling method;

[0066] Mark the data packets with unqualified sampling results in the m target sub-data packets of the target root node as abnormal data packets; obtain the total frequency n of the abnormal data packets in the corresponding record data packets of the target root node, and calculate the unqualified rate of the target root node corresponding to the i-th target sub-data packet as r i , r i =(c i / m i )*(c i / n); where m i represents the total number of times the i-th target sub-packet is recorded; c i Indicates the number of times the i-th target sub-packet is recorded as an abnormal packet;

[0067] Get the position data corresponding to the link in the abnormal data packet. When the position of the sampling link is different from the position of the forward link that needs to be returned after the sampling abnormality, the mark feature value is 1. When the position of the sampling link is the same as the position of the forward link that needs to be returned after the sampling abnormality, the mark feature value is 0. The forward link refers to the process link before the sampling link recorded in chronological order according to the inspection process data. Use the formula: g i =0.45*r i +0.55*a, calculate the correlation influence coefficient gi of the i-th target sub-data packet of the target root node; a={0,1}.

[0068] The difference between the position of the sampling link in the abnormal data packet and the position of the forward link that needs to be returned after the sampling abnormality is analyzed. This is because after the sampling fails, it is necessary to return to the previous link such as the packaging and packaging link to replace the unqualified products for re-inspection. Then there will be difficulty in returning to the forward link after the goods fail to meet the standards. When the sampling of goods is carried out during the packaging of the product, if an abnormality occurs here, it can be stopped immediately, and there is no need to move the goods, which reduces the complexity of solving problems when encountering difficulties in import and export trade and saves the import and export process time.

[0069] Step S2 also includes the following analysis steps:

[0070] Get the time efficiency coefficient t of the i-th target sub-packet i And the correlation coefficient g i , using the formula:

[0071] V i =k1*t i +k2*g i

[0072] Calculate the priority index V of the i-th target sub-data packet of the target root node i, where k1 and k2 represent the reference coefficient corresponding to the timeliness coefficient and the reference coefficient corresponding to the correlation influence coefficient respectively; k1+k2=1,0 <k1、k2<1;

[0073] The priority indexes corresponding to the m target sub-data packages of the target root node are arranged in ascending order, and the selection priority corresponding to the target root node is generated. The selection priority of the selection priority is the same as the priority arrangement order. The smaller the priority index analyzed above, the more stable and reliable the sampling method in the target sub-data package corresponding to the target root node is. Therefore, the same selection priority as the priority arrangement order indicates that the sampling method with the first priority order is the preferred selection method for import and export goods corresponding to the target root node.

[0074] In step S3, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed, including the following analysis steps:

[0075] Get the m corresponding to the i-th sampling method under the target root node i target sub-packet, extract the mth i The location data corresponding to the target sub-packets are sorted in chronological order to generate a location set b mi ;

[0076] M i The location set b corresponding to the target sub-packets mi Carry out b m1 ∩b m2 ∩b m3 ......∩b mi Operation, get the intersection position, and output the intersection position as the reliable position w of the i-th sampling method i ;

[0077] The intersection operation of the location set in the target sub-data package is performed on the consideration that even if the sampling methods are different, the inspected goods are the same, so there must be the same process links. The analysis of the same process links is limited to the real-time monitoring state to ensure the consistency of the general direction of the imported and exported goods and the historical data analysis of the same goods.

[0078] Obtain the location data f1 of the import and export goods in the real-time monitoring state after they enter the inspection process but before the sampling phase begins, and extract the location data p1 of the first target sub-data package in the selection priority;

[0079] Then calculate the position data f1 and the reliable position w i , the similarity index e{f1→w i ,p1},e{f1→w i ,p1}=(e f +e p) / (e0+e1); where e f Represents the position data f1 and the reliable position w i The number of identical positions when comparing, e p It indicates the number of identical positions when comparing position data f1 and position data p1. e0 indicates the existence of a reliable position w i , e1 represents the total number of position data in the first target sub-data packet.

[0080] The difference in the location data of different target sub-packets in the analysis and selection priority is to determine the direction of the links of the real-time monitored import and export goods in a digital way, because the selection of the sampling method can be obtained from the data analysis of each link. For example, if the warehouse capacity is tight, sampling cannot be carried out in the warehouse. It is reflected in the data that when entering the warehouse, it will be notified that there is no need to enter the warehouse, and the location of the link does not need to be changed. Then, when it corresponds to the location data of such a sampling link in the history, similar matching can be performed to achieve an early warning effect on the sampling method;

[0081] In the import and export trade inspection process, different sampling methods correspond to different orders in the entire inspection process, and the positions of different links are also different. For example, when the sampling method of goods is stacking sampling, the goods are often densely stacked in the warehouse, and when the sampling method is bag-switching sampling, this method often performs regular sampling in the process of unloading the goods according to the unloading order. The location in this case is not necessarily the warehouse. Therefore, the position of the same kind of goods in each link of the inspection process represents a certain tendency to choose a sampling method. This application gives the priority of sampling method selection in consideration of efficiency, convenience, and the degree of impact when anomalies occur, and comprehensively recommends and warns of sampling methods for import and export goods under real-time monitoring.

[0082] Step S4 includes the following analysis steps:

[0083] When the similarity index is greater than or equal to the similarity index threshold, the first sampling method warning signal in the priority level is selected when the name of the imported and exported goods in the transmission inspection process data is the same as that in the real-time import and export goods;

[0084] When the similarity index is less than the similarity index threshold, remove the first target sub-packet, obtain the position data in the target sub-packets in the selection priority in turn, and calculate the similarity index; extract the sampling method in the target sub-packet corresponding to the first similarity index greater than or equal to the similarity index threshold after calculation to perform transmission warning.

[0085] As shown in the embodiment: the similarity index is calculated in the same way as above after removing the first target sub-data packet;

[0086] Then the position data p2 of the second target sub-data packet in the selection priority can be extracted;

[0087] Calculate the position data f1 and the reliable position w i , the similarity index e{f1→w i ,p2},e{f1→w i ,p2}=(e f +e p2 ) / (e0+e1); if the similarity index is still not greater than or equal to the similarity index threshold;

[0088] Then continue to extract the position data p3 of the second target sub-packet in the selection priority; and calculate e{f1→w i ,p3}, if e{f1→w i ,p3} is greater than the similarity index threshold, there is no need to continue calculating the similarity index of the position data of the target sub-packet with the fourth selection priority, and the sampling method recorded in the third target sub-packet in the selection priority is transmitted for transmission warning.

[0089] Import and export trade management system, including inspection process data analysis module, sampling method difference determination module, attribute matching package construction module, selection priority analysis module, real-time monitoring analysis module and early warning reminder module;

[0090] The inspection process data analysis module is used to obtain the inspection process data recorded in the import and export trade management platform. The inspection process data is a data package containing the entire process of goods inspection. The inspection process data refers to the attribute feature data recorded from the start of the sampling inspection link to the completion of the inspection process of imported and exported goods.

[0091] The sampling method difference determination module is used to take the name of the imported and exported goods as the root node, extract the sampling methods corresponding to the same root node in different data packets in the inspection process data as the sampling methods to be analyzed; and analyze whether the sampling methods to be analyzed are all the same;

[0092] The attribute matching package construction module is used for, when the sampling methods to be analyzed are the same, the transmission root node is the first root node, the sampling method to be analyzed is the first sampling method, and constructing the attribute matching package of the first root node and the first sampling method;

[0093] The selection priority analysis module is used to analyze the selection priority of the root node when the sampling methods to be analyzed are different;

[0094] The real-time monitoring and analysis module is used to obtain the names of import and export goods under real-time monitoring status and match them with the attribute matching package; if the match is successful, the signal is transmitted to the early warning reminder module, and if the match fails, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed;

[0095] The early warning reminder module is used to transmit an early warning of the sampling method of imported and exported goods before the inspection process begins based on the relationship between the similarity index threshold and the similarity index.

[0096] The selection priority analysis module includes a data packet construction unit to be analyzed, a time efficiency coefficient calculation unit, a correlation influence coefficient calculation unit, a priority index calculation unit and a selection priority generation unit;

[0097] The to-be-analyzed data packet construction unit is used to construct a data packet formed by a target root node and a target sub-data packet;

[0098] The time efficiency coefficient calculation unit is used to calculate the time efficiency coefficient based on the sampling time and the interval time;

[0099] The correlation influence coefficient calculation unit is used to calculate the correlation influence coefficient based on the unqualified rate and the characteristic value; the unqualified rate is calculated by the data corresponding to the sampling result in the target sub-data package when it is unqualified and the frequency of the target sub-data package record; the characteristic value refers to the characteristic number marked according to the difference between the position of the sampling link and the position of the forward link returned after the sampling is abnormal;

[0100] The priority index calculation unit calculates the priority index based on the timeliness coefficient and the correlation influence coefficient;

[0101] The selection priority generation unit is used to arrange the priority indexes corresponding to the target sub-data packets of the target root node in ascending order, and generate the selection priority corresponding to the target root node.

[0102] The real-time monitoring and analysis module includes a location set generation unit, a reliable location output unit, a location data acquisition unit and a similarity index calculation unit;

[0103] The position set generating unit is used to extract the position data corresponding to the target sub-data packet and sort them in chronological order to generate a position set;

[0104] The reliable position output unit is used to perform an intersection operation on the position set corresponding to the target sub-data packet to obtain the intersection position, and output the intersection position as the reliable position corresponding to the sampling method;

[0105] The location data acquisition unit is used to obtain the location data of the import and export goods in the real-time monitoring state after they enter the inspection process but before the sampling phase begins, and the location data of the first target sub-data package in the selection priority;

[0106] The similarity index calculation unit is used to calculate the similarity index based on the position data and the reliable position.

[0107] The early warning module includes a threshold comparison unit, a feedback output unit and a transmission early warning unit;

[0108] The threshold comparison unit is used to compare the magnitude relationship between the similarity index and the similarity index threshold, and to transmit signals to the feedback transmission unit and the transmission warning unit;

[0109] The feedback output unit is used to transmit a signal to the real-time monitoring and analysis module to calculate a new similarity index based on the magnitude relationship of the output of the threshold comparison unit, and transmit a signal to the transmission warning unit when the similarity index is greater than or equal to the similarity index threshold;

[0110] The transmission warning unit is used to receive signals and perform transmission warning in a sampling manner.

[0111] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0112] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An import and export trade management method based on artificial intelligence, characterized in that: The method comprises the following analysis steps: Step S1: Obtain the inspection process data recorded in the import and export trade management platform. The inspection process data is a data package of the entire process of goods inspection. The inspection process data refers to the attribute feature data recorded before the start of the sampling inspection link to the completion of the inspection process of the imported and exported goods; the attribute features include the name of the imported and exported goods, the location data corresponding to the link, the time data corresponding to the link, the sampling method of the goods, and the sampling results; the sampling results include qualified and unqualified; Step S2: Taking the name of the imported and exported goods as the root node, extracting the sampling methods corresponding to the same root node in different data packets in the inspection process data as the sampling methods to be analyzed; if the sampling methods to be analyzed are the same, the transmission root node is the first root node, the sampling method to be analyzed is the first sampling method, and constructing an attribute matching package A of the first root node and the first sampling method, A={first root node→first sampling method}; the attribute matching package refers to recording the sampling methods corresponding to the imported and exported goods; if there are differences in the sampling methods to be analyzed, the selection priority of the root node is analyzed based on the data packet to which the root node belongs; The step S2 analyzes the selection priority of the root node based on the data packets to which the root node belongs, including the following analysis steps: The inspection process data corresponding to the root node with different sampling methods to be analyzed is used to construct a data packet to be analyzed, wherein the data packet to be analyzed is composed of a root node and inspection process data corresponding to m sampling methods; the root node in the data packet to be analyzed is the target root node, and the sampling methods and corresponding inspection process data in the data packet to be analyzed are the target sub-data packet; Get the time data corresponding to the link in the i-th target sub-data packet of the target root node, extract the sampling time T of the i-th target sub-data packet of the target root node to record the sampling link 1i And the interval time T from completing the sampling process to obtaining the sampling results 2i ; Using the formula: t i =T 1i / H1+T 2i / H2, calculate the time efficiency coefficient t of the i-th target sub-data packet of the target root node i ; Where i≤m; H1, H2 represent the average sampling duration and the average interval duration of the m target sub-data packets contained in the target root node; H1=1 / m(∑T 1i ), H2=1 / m(∑T 2i ); Mark the data packets with unqualified sampling results in the m target sub-data packets of the target root node as abnormal data packets; obtain the total frequency n of the abnormal data packets in the corresponding record data packets of the target root node, and calculate the unqualified rate of the target root node corresponding to the i-th target sub-data packet as r i , r i =(c i / m i )*(c i / n); where m i represents the total number of times the i-th target sub-packet is recorded; c i Indicates the number of times the i-th target sub-packet is recorded as an abnormal packet; Get the position data corresponding to the link in the abnormal data packet. When the position of the sampling link is different from the position of the forward link that needs to be returned after the sampling abnormality, the mark feature value is 1. When the position of the sampling link is the same as the position of the forward link that needs to be returned after the sampling abnormality, the mark feature value is 0. The forward link refers to the process link before the sampling link recorded in chronological order according to the inspection process data. Use the formula: g i =0.45*r i +0.55*a, calculate the correlation influence coefficient gi of the i-th target sub-data packet of the target root node; a={0,1}; The step S2 also includes the following analysis steps: Get the time efficiency coefficient t of the i-th target sub-packet i And the correlation coefficient g i , using the formula: V i =k1*t i +k2*g i Calculate the priority index V of the i-th target sub-data packet of the target root node i , where k1 and k2 represent the reference coefficient corresponding to the timeliness coefficient and the reference coefficient corresponding to the correlation influence coefficient respectively; k1+k2=1,0 <k1、k2<1; Arrange the priority indices corresponding to the m target sub-data packets of the target root node in ascending order, and generate a selection priority corresponding to the target root node, wherein the selection priority of the selection priority is the same as the priority arrangement order; Step S3: Based on the selection priority of the imported and exported goods, the names of the imported and exported goods under real-time monitoring are obtained; when the names of the imported and exported goods exist in the attribute matching package, the first sampling method in the corresponding attribute matching package is transmitted before the inspection process starts for early warning reminder; when the names of the imported and exported goods do not exist in the attribute matching package, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed; The step S3 of analyzing the similarity index between the real-time inspection process data and the data in the selection priority includes the following analysis steps: Get the m corresponding to the i-th sampling method under the target root node i target sub-packet, extract the mth i The location data corresponding to the target sub-packets are sorted in chronological order to generate a location set b mi ; M i The location set b corresponding to the target sub-packets mi Carry out b m1 ∩b m2 ∩b m3 ......∩b mi Operation is performed to obtain the intersection position, and the intersection position is output as the reliable position w of the i-th sampling method. i ; Obtain the location data f1 of the import and export goods in the real-time monitoring state after they enter the inspection process but before the sampling phase begins, and extract the location data p1 of the first target sub-data package in the selection priority; Then calculate the position data f1 and the reliable position w i , the similarity index e{f1→w i ,p1},e{f1→w i ,p1}=(e f +e p ) / (e0+e1); where e f Represents the position data f1 and the reliable position w i The number of identical positions when comparing, e p It indicates the number of identical positions when comparing position data f1 and position data p1. e0 indicates the existence of a reliable position w i , e1 represents the total number of position data in the first target sub-data packet; Step S4: Setting a similarity index threshold, based on the relationship between the similarity index threshold and the similarity index, transmitting an early warning of the sampling method for import and export goods before the inspection process begins.

2. The import and export trade management method based on artificial intelligence according to claim 1 is characterized by: The step S4 comprises the following analysis steps: When the similarity index is greater than or equal to the similarity index threshold, the first sampling method warning signal in the priority level is selected when the name of the imported and exported goods in the transmission inspection process data is the same as that in the real-time import and export goods; When the similarity index is less than the similarity index threshold, remove the first target sub-packet, obtain the position data in the target sub-packets in the selection priority in turn, and calculate the similarity index; extract the sampling method in the target sub-packet corresponding to the first similarity index greater than or equal to the similarity index threshold after calculation to perform transmission warning.

3. An import and export trade management system using an import and export trade management method based on artificial intelligence as claimed in any one of claims 1 to 2, characterized in that: It includes inspection process data analysis module, sampling method difference determination module, attribute matching package construction module, selection priority analysis module, real-time monitoring analysis module and early warning reminder module; The inspection process data analysis module is used to obtain the inspection process data recorded by the import and export trade management platform. The inspection process data is a data package containing the entire process of goods inspection. The inspection process data refers to the attribute feature data recorded from before the start of the sampling inspection link to the completion of the inspection process of imported and exported goods; The sampling method difference determination module is used to take the name of the imported and exported goods as the root node, extract the sampling methods corresponding to the same root node in different data packets in the inspection process data as the sampling methods to be analyzed; and analyze whether the sampling methods to be analyzed are all the same; The attribute matching package construction module is used for, when the sampling modes to be analyzed are the same, the transmission root node is the first root node, the sampling mode to be analyzed is the first sampling mode, and constructing the attribute matching package of the first root node and the first sampling mode; The selection priority analysis module is used to analyze the selection priority of the root node when the sampling modes to be analyzed are different; The real-time monitoring and analysis module is used to obtain the names of import and export goods under real-time monitoring status and match them with the attribute matching package; if the match is successful, a signal is transmitted to the early warning reminder module; if the match fails, the similarity index between the real-time inspection process data and the data in the selection priority is analyzed; The early warning reminder module is used to transmit an early warning of the sampling method of imported and exported goods before the inspection process begins based on the relationship between the similarity index threshold and the similarity index.

4. The import and export trade management system according to claim 3 is characterized by: The selection priority analysis module includes a data packet construction unit to be analyzed, a timeliness coefficient calculation unit, a correlation influence coefficient calculation unit, a priority index calculation unit and a selection priority generation unit; The to-be-analyzed data packet construction unit is used to construct a data packet formed by a target root node and a target sub-data packet; The time efficiency coefficient calculation unit is used to calculate the time efficiency coefficient based on the sampling duration and the interval duration; The correlation influence coefficient calculation unit is used to calculate the correlation influence coefficient based on the unqualified rate and the characteristic value; the unqualified rate is calculated by the data corresponding to the sampling result in the target sub-data packet when it is unqualified and the frequency of the target sub-data packet record; the characteristic value refers to the characteristic number marked according to the difference between the position of the sampling link and the position of the forward link returned after the sampling is abnormal; The priority index calculation unit calculates the priority index based on the timeliness coefficient and the association influence coefficient; The selection priority generation unit is used to arrange the priority indexes corresponding to the target sub-data packets of the target root node in ascending order, and generate the selection priority corresponding to the target root node.

5. The import and export trade management system according to claim 4 is characterized by: The real-time monitoring and analysis module includes a location set generation unit, a reliable location output unit, a location data acquisition unit and a similarity index calculation unit; The position set generating unit is used to extract the position data corresponding to the target sub-data packet and sort them in chronological order to generate a position set; The reliable position output unit is used to perform an intersection operation on the position set corresponding to the target sub-data packet to obtain the intersection position, and output the intersection position as the reliable position corresponding to the sampling method; The position data acquisition unit is used to obtain the position data of the import and export goods in the real-time monitoring state after they enter the inspection process but before the sampling phase begins, and the position data of the first target sub-data package in the selection priority; The similarity index calculation unit is used to calculate the similarity index based on the position data and the reliable position.

6. The import and export trade management system according to claim 5, characterized in that: The early warning module includes a threshold comparison unit, a feedback output unit and a transmission early warning unit; The threshold comparison unit is used to compare the magnitude relationship between the similarity index and the similarity index threshold, and to perform signal transmission with the feedback transmission unit and the transmission warning unit; The feedback output unit is used to transmit a signal to the real-time monitoring and analysis module to calculate a new similarity index based on the magnitude relationship of the output of the threshold comparison unit, and transmit a signal to the transmission warning unit when the similarity index is greater than or equal to the similarity index threshold; The transmission warning unit is used to receive signals and perform transmission warning in a sampling manner.

Images