Logistics permeability detection method and device

Abstract

The invention discloses a logistics permeability detection method and device, and relates to the technical field of computers. A specific embodiment of the method comprises the following steps: inputting a generated permeability time sequence and attribute characteristics of a customer into a permeability calculation model, outputting the permeability and a first permeability interval of a corresponding first logistics, and constructing a training sample of the permeability calculation model by adopting a training sample mask reconstruction mode; in response to a received query request about a permeability detection result, querying a target customer and the permeability and a first permeability interval of the first logistics corresponding to the target customer; and for each target customer, detecting the permeability of the target customer according to the permeability of the first logistics and the first permeability interval. According to the implementation mode, accurate and efficient detection of the logistics permeability is realized, and accurate positioning of customers with order increase potential or reduction risk is facilitated, so that the value of stock customers is further excavated and improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and device for detecting logistics permeability. Background Art

[0002] At present, the behaviors of existing customers in the logistics industry are diverse throughout their life cycle. Mining and improving the value of existing customers is the goal of refined operations. A key link in improving the value of existing customers is to accurately identify the logistics penetration rate of this logistics for each existing customer. The logistics penetration rate of the embodiment of the present invention is the proportion of the number of orders issued by customers using designated logistics to the total number of orders issued by the customer. High-quality logistics penetration rates help to accurately locate existing customers with the potential for increasing orders and the risk of reducing orders. By testing the logistics penetration rate, it can effectively guide account managers to locate customers in advance and select customers with the maximum return rate for targeted visits and maintenance. The existing determination and detection of logistics penetration rates usually adopts manual estimation and detection methods, which are estimated based on indicators such as the customer's historical penetration rate trend and the current order volume, customer feedback information, and then the estimated logistics penetration rate is tested.

[0003] In the process of implementing the present invention, the inventors found that the prior art has the following problems:

[0004] Manual estimation and detection methods are highly dependent on historical experience, and the calculation method is relatively crude, especially for customers who place orders intermittently, for example, only place orders for a few months in a year. When the order placement pattern is unknown or there is no order placement pattern, the error of estimating the logistics penetration rate based on the historical logistics penetration rate will be large. In addition, it is difficult for manual estimation and detection to fully consider the impact of various factors on the logistics penetration rate, resulting in a large deviation in the estimated logistics penetration rate. Correspondingly, it also affects the detection results of the logistics penetration rate, and ultimately leads to misjudgment. It cannot effectively guide account managers to locate customers with the potential for increasing orders or customers at risk of reducing orders, thereby affecting the formulation of customer strategies. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a method and device for detecting logistics permeability, which realizes accurate and efficient detection of logistics permeability, and trains a permeability calculation model using masked reconstructed training samples. Based on the logistics permeability and the first permeability interval output by the permeability calculation model, the logistics permeability of the target customer is queried and detected, which not only improves the calculation and detection efficiency of the logistics permeability, but also improves the accuracy of the calculation and detection of the logistics permeability, which helps account managers to accurately locate customers with the potential for increasing orders or customers at risk of reducing orders, and then formulate targeted and effective customer strategies to further tap and enhance the value of existing customers.

[0006] To achieve the above object, according to one aspect of an embodiment of the present invention, a method for detecting logistics permeability is provided, comprising:

[0007] Generate a time series of the penetration rate of the first logistics according to the first waybill data of the customer using the first logistics and the second waybill data of the customer using the second logistics, and obtain the attribute characteristics of the customer;

[0008] Input the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and the first permeability interval of the first logistics corresponding to the customer, wherein the training samples of the permeability calculation model are constructed by using a training sample mask reconstruction method;

[0009] In response to receiving a query request about a permeability test result, the query condition is parsed from the query request, and the target customers meeting the query condition and their corresponding permeability and first permeability range of the first logistics are queried; for each target customer, a permeability test is performed on the target customer based on the permeability and first permeability range of the first logistics corresponding to the target customer.

[0010] Optionally, the permeability calculation model is trained in the following manner: a plurality of training customers are determined from a customer base, and a training permeability time series and training attribute characteristics of the first logistics corresponding to each of the training customers are determined; according to a preset mask rule, a specified number of time points are determined from each of the training permeability time series, and the permeability data corresponding to the time points in the training permeability time series are masked to obtain a mask sequence, and the masked permeability data is used as a sample label; according to the mask sequence, the training attribute characteristics, and the sample label, a training sample of the permeability calculation model is constructed; using the training sample, the permeability calculation model is trained by the minimum mean square error loss function and gradient descent to obtain a trained permeability calculation model.

[0011] Optionally, determining a specified number of time points from each of the training penetration rate time series, including: judging whether the training penetration rate time series needs to be enhanced based on the length of the training penetration rate time series of each of the training customers and a preset length threshold; if the training penetration rate time series needs to be enhanced, enhancing the training penetration rate time series by sequence replication and splicing; determining a specified number of time points from each of the training penetration rate time series or the enhanced training penetration rate time series.

[0012] Optionally, the permeability calculation model adopts the coding layer structure in the deep learning model based on the self-attention mechanism; the permeability time series and the customer attribute characteristics are input into the pre-trained permeability calculation model, thereby outputting the permeability and the first permeability range of the first logistics corresponding to the customer, including: inputting the permeability time series and the attribute characteristics into the coding layer in the deep learning model based on the self-attention mechanism, and obtaining the intermediate vector matrix output by the coding layer; performing permeability mapping on the intermediate vector matrix to obtain the permeability and the first permeability range of the first logistics corresponding to the customer.

[0013] Optionally, a permeability test is performed on the target customer based on the permeability and the first permeability interval of the first logistics corresponding to the target customer, including: probability mapping the permeability and the first permeability interval according to the query condition to obtain the mapped permeability and the mapped first permeability interval; and obtaining the permeability test result of the target customer by calculating the first deviation between the mapped permeability and the mapped first permeability interval.

[0014] Optionally, when a configured second permeability interval is received, the method further includes: calculating a second deviation between the mapped permeability and the second permeability interval to obtain a second permeability detection result of the target customer; defining a weight coefficient corresponding to the permeability detection result, and a second weight coefficient corresponding to the second permeability detection result, and calculating the final permeability detection result of the target customer based on the weight coefficient and the permeability detection result, and the second weight coefficient and the second permeability detection result.

[0015] Optionally, the attribute features are constructed by extracting and fusing features of the customer's historical logistics waybill information, customer portrait, and the customer's feedback information on the first logistics.

[0016] According to a second aspect of an embodiment of the present invention, there is provided a device for detecting a flow permeability, comprising:

[0017] A logistics data extraction module, for generating a time series of the penetration rate of the first logistics and obtaining the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics and the second waybill data of the customer using the second logistics;

[0018] a permeability calculation module, for inputting the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and first permeability interval of the first logistics corresponding to the customer, wherein the training samples of the permeability calculation model are constructed by adopting a training sample mask reconstruction method;

[0019] A permeability detection module is used to respond to a query request regarding a permeability detection result, parse the query conditions from the query request, and query the target customers meeting the query conditions and their corresponding permeability and first permeability range of the first logistics; for each target customer, perform a permeability detection on the target customer based on the permeability and first permeability range of the first logistics corresponding to the target customer.

[0020] According to a third aspect of an embodiment of the present invention, there is provided an electronic detection device for logistics permeability, comprising:

[0021] one or more processors;

[0022] a storage device for storing one or more programs,

[0023] When the one or more programs are executed by the one or more processors, the one or more processors implement the method provided in the first aspect of the embodiment of the present invention.

[0024] According to a fourth aspect of an embodiment of the present invention, a computer-readable medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method provided by the first aspect of the embodiment of the present invention is implemented.

[0025] An embodiment of the invention has the following advantages or beneficial effects: by generating a permeability time series of the first logistics and obtaining the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics to issue the waybill and the second waybill data of the customer using the second logistics to issue the waybill; inputting the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and first permeability interval of the first logistics corresponding to the customer, and the training samples of the permeability calculation model are constructed by using the training sample mask reconstruction method; in response to receiving a query request about the permeability detection result, parsing the query condition from the query request, and querying the target customer meeting the query condition and the permeability and first permeability interval of the first logistics corresponding to the target customer; For each target customer, a technical solution is provided to detect the penetration rate of the target customer according to the first logistics penetration rate and the first penetration rate interval corresponding to the target customer, thereby realizing an accurate and efficient logistics penetration rate detection method. The permeability calculation model is trained using masked reconstructed training samples, and the logistics penetration rate of the target customer is queried and detected based on the logistics penetration rate and the first penetration rate interval output by the permeability calculation model. This not only improves the calculation and detection efficiency of the logistics penetration rate, but also improves the accuracy of the calculation and detection of the logistics penetration rate, which helps account managers to accurately locate customers with the potential for increasing orders or customers at risk of reducing orders, and then formulate targeted and effective customer strategies to further tap and enhance the value of existing customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention.

[0027] Figure 1 is a schematic diagram of the main process of a method for detecting logistics permeability according to an embodiment of the present invention;

[0028] Figure 2 is a schematic diagram of the construction process of the permeability calculation model of an embodiment of the present invention;

[0029] Figure 3 Schematic diagram of the principle of permeability detection according to an embodiment of the present invention;

[0030] Figure 4 1 is a schematic diagram of the system principle of permeability detection according to an embodiment of the present invention;

[0031] Figure 5 is a schematic diagram of main modules of a device for detecting logistics permeability according to an embodiment of the present invention;

[0032] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0033] Figure 7It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0034] It should be noted that in the technical solution of the present invention, the collection / collection, update, analysis, use, transmission, storage and other aspects of user personal information involved are in compliance with the provisions of relevant laws and regulations, are used for legal and reasonable purposes, are not shared, leaked or sold outside of these legal uses, and are subject to supervision and management by national regulatory authorities. Necessary measures should be taken for user personal information to selectively block the use or access to personal information data to prevent illegal access to such personal information data, ensure that persons who have access to personal information data comply with the provisions of relevant laws and regulations, and ensure the security of user personal information. In addition, once these user personal information data are no longer needed, the risks should be minimized by limiting or even prohibiting data collection and / or deleting data.

[0035] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0036] At present, manual estimation and detection methods are highly dependent on historical experience, and the calculation method is relatively crude, especially for customers who place orders intermittently, for example, only place orders for a few months in a year. When their order placement rules are unknown or there is no order placement rule, the error of estimating the logistics penetration rate by relying on historical logistics penetration rates will be large. In addition, it is difficult for manual estimation and detection to fully consider the impact of various factors on the logistics penetration rate, resulting in a large deviation in the estimated logistics penetration rate. Correspondingly, it also affects the detection results of the logistics penetration rate, and ultimately leads to misjudgment. It cannot effectively guide account managers to locate customers with the potential for increasing orders or customers at risk of reducing orders, thereby affecting the formulation of customer strategies.

[0037] In order to solve the above-mentioned problems existing in the prior art, the present invention proposes a method for detecting logistics permeability, which uses a permeability calculation model to obtain the permeability and first permeability interval of the first logistics corresponding to the customer. The training samples of the permeability calculation model are constructed by using the training sample mask reconstruction method, and when a query request for the permeability detection result is received, the permeability detection is performed according to the first logistics permeability and first permeability interval of the target customer queried, thereby realizing an accurate and efficient logistics permeability detection method, which uses the masked reconstructed training samples to train the permeability calculation model, and based on the logistics permeability and first permeability interval output by the permeability calculation model, the logistics permeability of the target customer is queried and detected, which not only improves the calculation and detection efficiency of the logistics permeability, but also improves the accuracy of the calculation and detection of the logistics permeability, which helps account managers to accurately locate customers with the potential for increasing orders or customers with the risk of reducing orders, and then formulate targeted and effective customer strategies to further explore and enhance the value of existing customers.

[0038] In the introduction to the embodiments of the present invention, the terms and their meanings are as follows:

[0039] Transformer: It is a deep learning model based on the self-attention mechanism. Because of its suitability for parallel computing and the complexity of its own model, it is superior to the previously popular recurrent neural network in terms of accuracy and performance.

[0040] Figure 1 Schematic diagram of the main process of the method for detecting the logistics permeability according to an embodiment of the present invention. Figure 1 As shown, the method for detecting the logistics permeability of the embodiment of the present invention includes the following steps S101 to S103.

[0041] Step S101: Generate a time series of the penetration rate of the first logistics and obtain the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics and the second waybill data of the customer using the second logistics.

[0042] Specifically, for logistics business, the logistics penetration rate of each customer not only represents the current operating status of the logistics business, but also contains clues such as the customer's satisfaction with the logistics business and the intention to continue to cooperate with the logistics business. It can be understood that there are many logistics companies in the logistics field. In the embodiment of the present invention, the business of this logistics company is regarded as the first logistics, and the business of other logistics companies except this logistics company is regarded as the second logistics. Of course, if the purpose of calculating the logistics penetration rate is to specify several logistics companies among other logistics companies, the business of the specified logistics company can also be regarded as the second logistics.

[0043] The embodiment of the present invention utilizes the learning ability of the deep learning model to construct the input data of the model. Based on the business characteristics of the logistics penetration rate, the real-time waybill data of each customer is collected through the online database, specifically including the first waybill data for issuing the order using the first logistics and the second waybill data for issuing the order using the second logistics. The sum of the first waybill data and the second waybill data is used as the total waybill data of the customer. For each customer, by calculating the ratio of the first waybill data at each time point to the total waybill data of the customer, a time series of the penetration rate of the first logistics of each customer is constructed and generated. In addition, considering that the logistics penetration rate of each customer at a future time point will be affected by various factors, such as the customer's business status, business scale, and the customer's level in the logistics business, the embodiment of the present invention takes the customer as the dimension, gathers all the characteristic data related to the logistics penetration rate together, and constructs the attribute characteristics of each customer.

[0044] According to one embodiment of the present invention, the attribute features are constructed by extracting and fusing features of the customer's historical logistics waybill information, customer portrait, and the customer's feedback information on the first logistics.

[0045] Specifically, the embodiment of the present invention collects each customer's historical logistics waybill information, customer portraits, customer feedback information on the first logistics received by the customer manager through online and offline visits to customers, and customer claims information from the offline data warehouse, and extracts and fuses the collected data information according to the customer dimension to obtain the attribute characteristics of each customer. It can be understood that the attribute characteristics of the customer not only include specific numerical features, but also interval features. For example, the feature extraction of the customer's feedback information is usually quantified as a feature value of an intention interval, and accordingly, there will be corresponding interval features in the attribute characteristics.

[0046] Step S102, input the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and the first permeability range of the first logistics corresponding to the customer. The training samples of the permeability calculation model are constructed by using training sample mask reconstruction.

[0047] Specifically, the permeability time series and attribute features of each customer constructed above are input into a pre-trained permeability calculation model, and the training samples of the permeability calculation model are mainly constructed by mask reconstruction of training samples. Based on the trained permeability calculation model, the permeability calculation model is used to predict the logistics permeability of each customer at a specified time point or time period in the future, and considering that the input attribute features of the permeability calculation model also include interval features, the permeability calculation model also has a floating interval for outputting the logistics permeability at the predicted time point or time period, which is the first permeability interval of the embodiment of the present invention.

[0048] According to one embodiment of the present invention, the permeability calculation model is trained in the following manner: a plurality of training customers are determined from a customer base, and a training permeability time series and training attribute characteristics of the first logistics corresponding to each of the training customers are determined; according to a preset mask rule, a specified number of time points are determined from each of the training permeability time series, and the permeability data corresponding to the time points in the training permeability time series are masked to obtain a mask sequence, and the masked permeability data is used as a sample label; according to the mask sequence, the training attribute characteristics, and the sample label, a training sample of the permeability calculation model is constructed; using the training sample, the permeability calculation model is trained by the minimum mean square error loss function and gradient descent to obtain a trained permeability calculation model.

[0049] Specifically, the training samples of the permeability calculation model of the embodiment of the present invention are constructed by mask reconstruction. First, multiple training customers are determined from the customer base. It can be understood that in order to ensure the quality of the samples, comprehensive screening can be performed from various categories, regions and other dimensions according to the characteristics of the customer's business and the region to which the customer belongs, and representative training customers can be selected. Then, the historical waybill data, customer portraits, customer feedback information on the first logistics, and customer claims information of each training customer are collected to construct the training permeability time series and training attribute characteristics of the first logistics corresponding to each training customer.

[0050] Furthermore, considering the accuracy and efficiency of mask reconstruction, the embodiment of the present invention sets the mask rules in advance: the masked time points are the time points in the sequence, not the time points at both ends of the sequence, and the specified number of masked time points cannot be continuous, which ensures that the time points before and after the masked time points are not masked. According to the set masking rules, a specified number of time points are selected from the training permeability time series of each customer as masked time points, and accordingly, the permeability data corresponding to these masked time points in the training permeability time series are shielded to obtain a masked sequence, and the shielded permeability data are used as sample labels for the corresponding masked time points. Then, according to the masked sequence, training attribute characteristics, and corresponding sample labels of each customer, the training samples of the permeability calculation model are constructed. In this way, the training samples are passed through the permeability calculation model. The permeability calculation model predicts the permeability data at the time point of the reconstruction mask through the permeability data at the time points before and after each mask. The permeability calculation model is trained by minimizing the mean square error between the predicted and reconstructed permeability data and the true value of the corresponding sample label and the gradient descent training method. When the training converges, the trained permeability calculation model is obtained.

[0051] According to another embodiment of the present invention, a specified number of time points are determined from each of the training penetration rate time series, including: judging whether the training penetration rate time series needs to be enhanced based on the length of the training penetration rate time series of each of the training customers and a preset length threshold; when the training penetration rate time series needs to be enhanced, enhancing the training penetration rate time series by sequence replication and splicing; determining a specified number of time points from each of the training penetration rate time series or the enhanced training penetration rate time series.

[0052] Specifically, in the process of determining a specified number of time points from each of the training permeability time series, it is considered that the lengths of the training permeability time series of different customers may be different. For customers who have just settled in or do not place orders frequently, their training permeability time series will be shorter. In order to ensure the quality of the training samples, first determine whether the length of each customer's training permeability time series meets the preset length threshold. If not, the training permeability time series needs to be enhanced. The training permeability time series to be enhanced is copied and spliced ​​a limited number of times, and the specific number of copies is determined according to the difference between the length of the training permeability time series and the length threshold. Then, according to the preset masking rules, a specified number of time points are determined from the training permeability time series or the enhanced training permeability time series.

[0053] It can be understood that in the process of training the permeability calculation model using the training sample set, the training set, validation set and test set can be determined according to a preset ratio, such as 7:1:2. The training set is used to train the permeability calculation model by gradient descent, and the parameters of the model are updated. The validation set and the test set are then used to optimize and evaluate the model after the parameter update, so as to train the permeability calculation model. The specific evaluation calculation formula of the permeability calculation model of the embodiment of the present invention is: Weighted mean absolute percentage error = sum(abs(predicted value - actual value)) / sum(abs(actual value)); Mean absolute error = avg(sum(abs(predicted value - actual value))); Mean absolute percentage error = avg(abs(predicted value - actual value) / abs(actual value)).

[0054] Figure 2 It is a schematic diagram of the construction process of the permeability calculation model of the embodiment of the present invention. First, the feature extraction network of the permeability calculation model is determined to establish the permeability calculation model; then, the training attribute features of feature fusion, as well as the mask sequence and sample labels are constructed in parallel to obtain the training samples; then, the permeability model is trained in multiple steps by gradient descent using the training set in the training samples to update the parameters of the permeability calculation model; then, the model is evaluated and tuned in combination with the test set and the validation set, and finally a permeability calculation model that can be used to calculate the first logistics permeability is obtained.

[0055] According to another embodiment of the present invention, the permeability calculation model adopts a coding layer structure in a deep learning model based on a self-attention mechanism; the permeability time series and the customer attribute characteristics are input into a pre-trained permeability calculation model to output the permeability and the first permeability range of the first logistics corresponding to the customer, including: inputting the permeability time series and the attribute characteristics into the coding layer in the deep learning model based on the self-attention mechanism to obtain an intermediate vector matrix output by the coding layer; performing permeability mapping on the intermediate vector matrix to obtain the permeability and the first permeability range of the first logistics corresponding to the customer.

[0056] Specifically, the embodiment of the present invention applies the deep learning model transformer based on the self-attention mechanism to the prediction of time series, making full use of the non-directional characteristics of the transformer. Compared with the traditional long short-term memory and other time series models that need to comprehensively utilize context information, the model adopted by the embodiment of the present invention greatly improves the accuracy of mask reconstruction. In fact, the embodiment of the present invention only needs to use the encoder part of the encoder layer in the transformer model to realize the prediction calculation of the logistics permeability of each customer, and the permeability of the future time point to be calculated is used as the mask mask to mask out the part, and the permeability calculation model is used to obtain the reconstructed permeability data of each time point to be calculated. The specific implementation method is to input the permeability time series and attribute features of each customer into the encoding layer in the deep learning model based on the self-attention mechanism, obtain the intermediate vector matrix output by the encoding layer, and then input the intermediate vector matrix into the fully connected layer, and perform permeability mapping through the activation function therein to obtain the permeability and first permeability interval of the first logistics corresponding to each customer, and store the permeability and first permeability interval of the first logistics corresponding to each customer in the database.

[0057] It can be understood that the permeability of the first logistics output by the permeability calculation model in the embodiment of the present invention can be the permeability at a certain point in the future, or the permeability at a certain time period in the future, and supports the calculation of the permeability of the first logistics at a specified time granularity (day, week, month, quarter, year) and the first permeability interval.

[0058] Step S103, in response to receiving a query request about a permeability test result, parsing the query condition from the query request, querying the target customers meeting the query condition and their corresponding permeability and first permeability range of the first logistics; for each target customer, performing a permeability test on the target customer based on the permeability and first permeability range of the first logistics corresponding to the target customer.

[0059] Specifically, based on the permeability and first permeability interval of the first logistics corresponding to each customer output by the above-mentioned permeability calculation model, the permeability detection system can periodically retrieve the permeability of the first logistics and the corresponding first permeability interval from the database to detect the permeability of the first logistics. It can be understood that in order to better meet the demands of the account manager, the embodiment of the present invention also supports conditional query of permeability. The account manager can enter the query conditions through the visual page of the front end, such as entering the customer ID and the calculation time point. Then the customer with the customer ID is the target customer. Accordingly, the first logistics permeability and the first permeability interval corresponding to the customer ID at the determined calculation time point are queried from the database, and then the first permeability interval is used as a scale to detect whether the first logistics permeability is normal, and the permeability detection is performed on the target customer.

[0060] According to one embodiment of the present invention, a permeability test is performed on the target customer based on the permeability and the first permeability interval of the first logistics corresponding to the target customer, including: performing probability mapping on the permeability and the first permeability interval according to the query condition to obtain the mapped permeability and the mapped first permeability interval; and obtaining the permeability test result of the target customer by calculating the first deviation between the mapped permeability and the mapped first permeability interval.

[0061] Specifically, if the query condition is one dimension, such as the customer ID mentioned above, the queried permeability and the first permeability interval are directly used as the mapped permeability and the mapped first permeability interval, and the permeability detection result is obtained by directly calculating the data difference between the queried permeability and the first permeability interval, also called the first deviation. For example, the permeability is 0.4 and the first permeability interval is 0.5~0.7, then the first deviation is -0.1, that is, the permeability detection result is -0.1, indicating that the target customer is an abnormally declining customer with a reduced amount.

[0062] Understandably, account managers sometimes need to detect the first logistics penetration rate of a certain type of customers. In this case, multiple customers with the same attributes need to be aggregated, such as fresh food customers with strict logistics timeliness requirements and VIP customer levels. In this case, the query conditions are customer business categories and customer levels, and then the penetration rate and first penetration rate interval of each target customer that meets the query conditions are queried. Then, according to the number of target customers aggregated by the query conditions, the penetration rate of each target customer is probabilistically normalized to obtain the mapped penetration rate. At the same time, the first penetration rate interval with the smallest span in the first penetration rate interval of each target customer is selected as the mapped first penetration rate interval, and the deviation between the mapped penetration rate and the mapped first penetration rate interval is used as the penetration rate detection result.

[0063] Of course, you can also first calculate the first deviation between the permeability and the first permeability interval; perform probability mapping on the first deviation according to the query conditions to obtain the permeability detection results of the target customers. For example, calculate the first deviation between the permeability of each target customer and the first permeability interval to obtain multiple first deviations, and then perform probability mapping of mean normalization on each first deviation according to the number of target customers aggregated by the query conditions to finally obtain the permeability detection results. The specific detection method is selected according to business needs.

[0064] According to another embodiment of the present invention, when a configured second permeability interval is received, the method further includes: calculating a second deviation between the mapped permeability and the second permeability interval to obtain a second permeability detection result of the target customer; defining a weight coefficient corresponding to the permeability detection result, and a second weight coefficient corresponding to the second permeability detection result, and calculating the final permeability detection result of the target customer based on the weight coefficient and the permeability detection result, and the second weight coefficient and the second permeability detection result.

[0065] Specifically, considering that the first waybill data of customers using the first logistics to place orders will also be affected by factors such as holidays, promotion days of trading activities, or discount service days provided by the first logistics, for example, the first waybill data during the big promotion period will definitely be higher than the usual data. Therefore, when the set calculation time includes holidays, promotion days, and discount service days, in order to avoid misjudgment, the business side can configure the second penetration rate range in advance, which can be elastically increased or decreased based on the corresponding first penetration rate range, or can configure the corresponding specific second penetration rate range for each customer.

[0066] Furthermore, based on the above-mapped permeability, a second deviation between the mapped permeability and the second permeability interval is calculated to obtain a second permeability detection result. The above permeability detection result and the second permeability detection result are then comprehensively considered, and according to business needs, corresponding weight coefficients are defined for the permeability detection results, and corresponding second weight systems are defined for the second permeability detection results, and the final permeability detection result of the target customer is calculated by weighted addition of the two permeability detection results.

[0067] The embodiment of the present invention takes into account various factors that affect the logistics permeability detection through two permeability detections, and provides a more reasonable detection basis for the logistics permeability detection through the second permeability interval, thereby improving the accuracy of permeability detection.

[0068] Furthermore, in addition to detecting the penetration rate of customers corresponding to the first logistics according to the query conditions of the account manager, the embodiment of the present invention also supports the account manager to classify the detected abnormal customers through the front-end visualization page, which can be divided into potential customers with abnormally increased penetration rate and lost customers with abnormally decreased penetration rate. The system follows the preset principle of maximizing the return rate, provides timely service guarantees for customers who increase orders and tap potential, and makes timely visits and problem recovery for customers who reduce orders and recover losses, thereby improving the ability to understand customers and accurately formulate customer strategies.

[0069] The embodiment of the present invention also experimentally verifies the calculation and detection method of the permeability. Through the propensity score matching algorithm, the differences between customers are eliminated, the customer data is diverted, and the customers are divided into an experimental group and a control group. First, the results of the permeability calculation are verified, as shown in Table 1. The experimental group uses the permeability calculation model of the present invention to calculate the permeability of the customer's first logistics, and the control group uses a solution to manually estimate the permeability of the customer's first logistics based on historical experience. During the test period in the fourth quarter of 2023, the monthly permeability calculation percentage error of 12,928 customers in the experimental group was 21.25%; the calculation percentage error of 1,436 customers in the control group was 39.17%, which was reduced to 49.22% in the experimental group compared with the control group.

[0070] The results of the permeability test were verified again, as shown in Table 2. The experimental group adopted the permeability test scheme of the present invention, and the control group adopted the permeability test scheme based on manual experience. During the test period in the fourth quarter of 2023, 12,928 customers in the experimental group accurately detected 10,714 monthly permeability abnormal nodes, with an accuracy rate of 82.87%; 937 of the 1,436 customers in the control group were accurately detected, with an accuracy rate of 65.25%, an increase of 27.00% in the experimental group compared with the control group.

[0071] Figure 3 It is a schematic diagram of the principle of permeability detection of an embodiment of the present invention. The first waybill data and the second waybill data are obtained from the waybill data unit of the first logistics of online data and the waybill data unit of the second logistics to generate the permeability time series of the first logistics, and then the data related to the calculation of the permeability of the first logistics is collected from the customer-related log data unit (customer log data and customer relationship management CRM log data) of the online data, the business data unit in the offline data warehouse, the customer portrait unit, the customer visit feedback data unit, and the data unit of the second logistics to perform feature extraction and feature fusion to obtain attribute features. The permeability time series and attribute features of the first logistics are input into the permeability calculation model through the feature unit, and the permeability calculation model adopts the encoding layer structure in the deep learning model based on the self-attention mechanism.

[0072] The permeability time series and attribute features are passed through the encoding layer in the deep learning model based on the self-attention mechanism to obtain the intermediate vector matrix output by the encoding layer, which realizes the mask reconstruction of the permeability at the time point to be calculated, and then the permeability and first permeability range of the first logistics corresponding to the customer are obtained by performing permeability mapping on the intermediate vector matrix. When receiving a query request for the permeability detection result, the query condition is parsed from the query request, and the target customer and the permeability and first permeability range of the first logistics corresponding to the query condition are queried. First, the permeability detection is performed on the target customer according to the permeability and first permeability range of the first logistics corresponding to the target customer to obtain the permeability detection result, and then the target customer is subjected to a secondary permeability detection in combination with the configured second permeability range to obtain the second permeability detection result. The probability mapping of the final permeability detection result is obtained by combining the permeability detection result and the second permeability detection result, and the permeability and first permeability range of the queried target customer are displayed in a graphical form on the front-end page through the unified interface service so that the querying party account manager can trace the data.

[0073] Figure 4 : This is a schematic diagram of the system principle of permeability detection in an embodiment of the present invention. The first waybill data of the customer using the first logistics to issue the order and the second waybill data using the second logistics to issue the order in the real-time data are collected to generate the permeability time series of the first logistics, and then construct attribute features from the operating data, customer feedback information, and customer portrait information in the offline data warehouse. The permeability time series and attribute features of the first logistics are input into the permeability calculation model obtained through the establishment of the permeability calculation model, model training and model evaluation, and the permeability and first permeability interval of the first logistics corresponding to the customer output by the model are obtained.

[0074] When a query request about the permeability test result is received, a permeability test is first performed on the target customer based on the permeability of the first logistics and the first permeability interval corresponding to the target customer to map the abnormal degree of the permeability; then, combined with the second permeability interval configured due to permeability influencing factors such as operating policies and holiday promotions, a secondary permeability test is performed on the permeability using the second permeability interval, and the discriminant test based on the first and second permeability intervals is completed based on the comprehensive permeability test results and the second permeability test results to obtain the final permeability test result, thereby detecting customers with the potential for increased orders or customers at risk of reduced orders.

[0075] Finally, based on the principle of maximizing return on investment (ROI), the system screens the customers that need to be visited and determines the specific customer strategy. The system generates a visit script for visiting customers according to the customer strategy and the visit script generation rules configured by the platform, and displays it to the account manager through the front-end page.

[0076] Figure 5Schematic diagram of the main modules of the device for detecting the logistics permeability according to an embodiment of the present invention. Figure 5 As shown, the logistics permeability detection device 500 mainly includes a logistics data extraction module 501, a permeability calculation module 502 and a permeability detection module 503.

[0077] The logistics data extraction module 501 is used to generate a time series of the penetration rate of the first logistics and obtain the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics and the second waybill data of the customer using the second logistics;

[0078] The permeability calculation module 502 is used to input the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, so as to output the permeability and first permeability interval of the first logistics corresponding to the customer, and the training samples of the permeability calculation model are constructed by using a training sample mask reconstruction method;

[0079] The permeability detection module 503 is used to respond to a query request regarding a permeability detection result, parse the query conditions from the query request, and query the target customers meeting the query conditions and their corresponding permeability and first permeability range of the first logistics; for each target customer, perform a permeability detection on the target customer based on the permeability and first permeability range of the first logistics corresponding to the target customer.

[0080] According to one embodiment of the present invention, the permeability calculation model is trained in the following manner: a plurality of training customers are determined from a customer base, and a training permeability time series and training attribute characteristics of the first logistics corresponding to each of the training customers are determined; according to a preset mask rule, a specified number of time points are determined from each of the training permeability time series, and the permeability data corresponding to the time points in the training permeability time series are masked to obtain a mask sequence, and the masked permeability data is used as a sample label; according to the mask sequence, the training attribute characteristics, and the sample label, a training sample of the permeability calculation model is constructed; using the training sample, the permeability calculation model is trained by the minimum mean square error loss function and gradient descent to obtain a trained permeability calculation model.

[0081] According to another embodiment of the present invention, the logistics permeability detection device 500 also includes a permeability calculation model training module (not shown in the figure), which is used to: determine whether the training permeability time series needs to be enhanced based on the length of the training permeability time series of each training customer and a preset length threshold; if the training permeability time series needs to be enhanced, enhance the training permeability time series by sequence replication and splicing; determine a specified number of time points from each of the training permeability time series or the enhanced training permeability time series.

[0082] According to another embodiment of the present invention, the permeability calculation model adopts the coding layer structure in the deep learning model based on the self-attention mechanism; the permeability calculation module 502 is also used to: input the permeability time series and the attribute features into the coding layer in the deep learning model based on the self-attention mechanism to obtain the intermediate vector matrix output by the coding layer; perform permeability mapping on the intermediate vector matrix to obtain the permeability and first permeability range of the first logistics corresponding to the customer.

[0083] According to another embodiment of the present invention, the permeability detection module 503 is further used to: perform probability mapping on the permeability and the first permeability interval according to the query condition to obtain the mapped permeability and the mapped first permeability interval; and obtain the permeability detection result of the target customer by calculating the first deviation between the mapped permeability and the mapped first permeability interval.

[0084] According to another embodiment of the present invention, when a configured second permeability interval is received, the logistics permeability detection device 500 also includes a permeability secondary detection module (not shown in the figure), which is used to: calculate the second deviation between the mapped permeability and the second permeability interval to obtain the second permeability detection result of the target customer; define a weight coefficient corresponding to the permeability detection result, and a second weight coefficient corresponding to the second permeability detection result, and calculate the final permeability detection result of the target customer based on the weight coefficient and the permeability detection result, and the second weight coefficient and the second permeability detection result.

[0085] According to yet another embodiment of the present invention, the attribute features are constructed by extracting and fusing features of the customer's historical logistics waybill information, customer portrait, and the customer's feedback information on the first logistics.

[0086] Figure 6 is an exemplary system architecture diagram to which embodiments of the present invention can be applied.

[0087] like Figure 6As shown, system architecture 600 may include terminal devices 601, 602, 603, network 604 and server 605. Network 604 is used to provide a medium for communication links between terminal devices 601, 602, 603 and server 605. Network 604 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0088] Users can use terminal devices 601, 602, 603 to interact with server 605 through network 604 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 601, 602, 603, such as penetration rate detection applications, etc. (only as an example).

[0089] The terminal devices 601 , 602 , and 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers, etc.

[0090] The server 605 may be a server that provides various services, such as a backend management server that supports the logistics penetration rate of users using the terminal devices 601, 602, and 603 (for example only). The backend management server may generate a time series of the penetration rate of the first logistics and obtain the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics to issue the order and the second waybill data of the customer using the second logistics to issue the order; the permeability time series and the customer attribute characteristics are input into a pre-trained permeability calculation model, thereby outputting the permeability and first permeability interval of the first logistics corresponding to the customer, and the training samples of the permeability calculation model are constructed by using the training sample mask reconstruction method; in response to receiving a query request about the permeability detection result, the query condition is parsed from the query request, and the target customer that meets the query condition and the permeability and first permeability interval of the first logistics corresponding to the target customer are queried; for each target customer, the permeability detection and other processing are performed on the target customer based on the permeability and first permeability interval of the first logistics corresponding to the target customer, and the processing result (such as the detection result, etc. - for example only) is fed back to the terminal device.

[0091] It should be noted that the logistics permeability detection method provided in the embodiment of the present invention is generally executed by the server 605 , and accordingly, the logistics permeability detection device is generally set in the server 605 .

[0092] It should be understood that Figure 6 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0093] Reference below Figure 7 , which shows a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. Figure 7 The terminal device or server shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0094] like Figure 7 As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0095] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, a mouse, etc.; an output section 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, a modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage section 708 as needed.

[0096] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the central processing unit (CPU) 701, the above-mentioned functions defined in the system of the present invention are executed.

[0097] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination thereof.

[0098] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0099] The units involved in the embodiments of the present invention may be implemented by software or hardware. The units described may also be set in a processor, for example, it may be described as: a processor includes: a logistics data extraction module, a permeability calculation module and a permeability detection module.

[0100] Among them, the names of these modules do not constitute a limitation of the modules themselves in certain cases. For example, the logistics data extraction module can also be described as "a module used to generate a time series of the penetration rate of the first logistics and obtain the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics to issue the waybill and the second waybill data of the customer using the second logistics to issue the waybill."

[0101] On the other hand, the present invention also provides a computer-readable medium, which may be included in the device described in the embodiment; or may exist independently and not be assembled into the device. The computer-readable medium carries one or more programs, and when the one or more programs are executed by a device, the device includes: generating a permeability time series of the first logistics according to the first waybill data of the customer using the first logistics to issue the order and the second waybill data of the customer using the second logistics to issue the order, and obtaining the attribute characteristics of the customer; inputting the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and first permeability interval of the first logistics corresponding to the customer, and the training samples of the permeability calculation model are constructed by using the training sample mask reconstruction method; in response to receiving a query request about the permeability detection result, parsing the query condition from the query request, and querying the target customer that meets the query condition and the permeability and first permeability interval of the first logistics corresponding to the target customer; for each target customer, performing permeability detection on the target customer according to the permeability and first permeability interval of the first logistics corresponding to the target customer.

[0102] The technical solution according to the embodiment of the present invention has the following advantages or beneficial effects: by generating a permeability time series of the first logistics and obtaining the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics to issue the waybill and the second waybill data of the customer using the second logistics to issue the waybill; inputting the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and the first permeability range of the first logistics corresponding to the customer, and the training samples of the permeability calculation model are constructed by using the training sample mask reconstruction method; in response to receiving a query request about the permeability detection result, parsing the query condition from the query request, and querying the target customer who meets the query condition and the permeability and the first permeability range of the first logistics corresponding to the customer. For each target customer, a technical solution for permeability detection of the target customer is implemented according to the first logistics permeability and the first permeability interval corresponding to the target customer, thereby realizing an accurate and efficient logistics permeability detection method. The permeability calculation model is trained using masked reconstructed training samples, and the logistics permeability of the target customer is queried and detected based on the logistics permeability and the first permeability interval output by the permeability calculation model. This not only improves the calculation and detection efficiency of the logistics permeability, but also improves the accuracy of the calculation and detection of the logistics permeability, which helps account managers to accurately locate customers with the potential for increasing orders or customers with the risk of reducing orders, and then formulate targeted and effective customer strategies to further tap and enhance the value of existing customers.

[0103] The specific implementation methods described herein do not constitute limitations on the scope of protection of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention shall be included in the scope of protection of the present invention.

Claims

1. A method for detecting logistics permeability, characterized in that: include: Generate a time series of the penetration rate of the first logistics according to the first waybill data of the customer using the first logistics and the second waybill data of the customer using the second logistics, and obtain the attribute characteristics of the customer; Input the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and the first permeability interval of the first logistics corresponding to the customer, wherein the training samples of the permeability calculation model are constructed by using a training sample mask reconstruction method; In response to receiving a query request about a permeability detection result, parsing a query condition from the query request, and querying a target customer meeting the query condition and a permeability and a first permeability interval of the first logistics corresponding thereto; For each target customer, a permeability test is performed on the target customer according to the permeability and the first permeability interval of the first logistics corresponding to the target customer.

2. The method according to claim 1, characterized in that The permeability calculation model is trained in the following way: Determine a plurality of training customers from the customer group, and determine a training penetration rate time series and training attribute characteristics of the first logistics corresponding to each of the training customers; According to a preset masking rule, a specified number of time points are determined from each of the training permeability time series, the permeability data corresponding to the time points in the training permeability time series are masked to obtain a masked sequence, and the masked permeability data is used as a sample label; Constructing a training sample of the permeability calculation model according to the mask sequence, the training attribute feature, and the sample label; The permeability calculation model is trained using the training samples through a minimum mean square error loss function and a gradient descent method to obtain a trained permeability calculation model.

3. The method according to claim 2, characterized in that Determining a specified number of time points from each of the training permeability time series includes: According to the length of the training penetration rate time series of each training customer and a preset length threshold, determining whether the training penetration rate time series needs to be enhanced; When the training permeability time series needs to be enhanced, the training permeability time series is enhanced by sequence replication and splicing; A specified number of time points are determined from each of the training permeability time series or the enhanced training permeability time series.

4. The method according to claim 1, characterized in that The permeability calculation model adopts the encoding layer structure in the deep learning model based on the self-attention mechanism; Inputting the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and first permeability interval of the first logistics corresponding to the customer, including: Inputting the permeability time series and the attribute features into the encoding layer in the deep learning model based on the self-attention mechanism to obtain an intermediate vector matrix output by the encoding layer; The intermediate vector matrix is ​​subjected to permeability mapping to obtain the permeability and the first permeability interval of the first logistics corresponding to the customer.

5. The method according to claim 1, characterized in that According to the permeability rate and the first permeability rate interval of the first logistics corresponding to the target customer, performing permeability detection on the target customer includes: According to the query condition, probability mapping is performed on the permeability and the first permeability interval to obtain a mapped permeability and a mapped first permeability interval; The permeability detection result of the target customer is obtained by calculating the first deviation between the mapped permeability and the mapped first permeability interval.

6. The method according to claim 5, characterized in that Upon receiving the configured second permeability interval, the method further comprises: Calculating a second deviation between the mapped permeability and the second permeability interval to obtain a second permeability detection result of the target customer; Define a weight coefficient corresponding to the permeability detection result and a second weight coefficient corresponding to the second permeability detection result, and calculate the final permeability detection result of the target customer based on the weight coefficient and the permeability detection result, and the second weight coefficient and the second permeability detection result.

7. The method according to claim 1, characterized in that The attribute features are constructed by extracting and fusing features of the customer's historical logistics waybill information, customer portrait, and the customer's feedback information on the first logistics.

8. A device for detecting logistics permeability, characterized in that: include: A logistics data extraction module, for generating a time series of the penetration rate of the first logistics and obtaining the attribute characteristics of the customer based on the first waybill data of the customer using the first logistics and the second waybill data of the customer using the second logistics; a permeability calculation module, for inputting the permeability time series and the customer attribute characteristics into a pre-trained permeability calculation model, thereby outputting the permeability and first permeability interval of the first logistics corresponding to the customer, wherein the training samples of the permeability calculation model are constructed by adopting a training sample mask reconstruction method; a permeability detection module, configured to, in response to receiving a query request regarding a permeability detection result, parse a query condition from the query request, and query the target customers meeting the query condition and the permeability and first permeability interval of the first logistics corresponding thereto; For each target customer, a permeability test is performed on the target customer according to the permeability and the first permeability interval of the first logistics corresponding to the target customer.

9. A mobile electronic device terminal, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images