Method and device for predicting purchase quantity

Abstract

The invention discloses a method and device for predicting the purchase volume, and the method comprises the steps: determining the predicted purchase volume of a first commodity according to the current inventory information and the predicted sales volume of the first commodity; determining a first purchasing constraint matched with the current condition of the first commodity from the condition parameters of the plurality of first purchasing constraints; any first purchase constraint is a purchase constraint matched with the manual modification sample determined from the purchase constraints; according to any manual quantity-changing sample characterization, the prediction purchase quantity of the sample is manually changed under the condition of the sample, and then the purchase quantity of the sample is obtained; and according to the behavior parameter of the matched first purchase constraint, changing the predicted purchase quantity, and determining the purchase quantity of the first commodity. By adopting the method, automatic adjustment of the predicted purchase quantity can be realized, and the workload of workers is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of commodity procurement, and in particular to a method and device for predicting procurement volume. Background Art

[0002] In the classic supply chain of suppliers, large warehouses and forward warehouses, procurement refers to the process of large warehouses placing orders with suppliers. If the purchase volume of goods is too large, it is easy to cause goods to pile up. If the purchase volume of goods is too small, it is easy to cause shortages of goods. How to predict the purchase volume of goods is an important part of procurement.

[0003] The existing methods for predicting commodity purchase volume often require adjustments by operations staff and purchasing staff. Due to the large number of commodity types, a large amount of human resources are consumed. Therefore, how to achieve automatic adjustment of the predicted purchase volume is an urgent problem to be solved. Summary of the invention

[0004] The present application provides a method and device for predicting purchase volume, which are used to achieve automatic adjustment of the predicted purchase volume.

[0005] In a first aspect, an embodiment of the present application provides a method for predicting purchase volume, which is executed by a device for predicting purchase volume, and the method includes: determining the predicted purchase volume of a first commodity based on the current inventory information and estimated sales volume of the first commodity; determining a first purchase constraint that matches the current status of the first commodity from the condition parameters of multiple first purchase constraints; any first purchase constraint is a purchase constraint that matches an artificially modified sample determined from each purchase constraint; any artificially modified sample represents the purchase volume of the sample obtained by artificially modifying the predicted purchase volume of the sample under the status of the sample; modifying the predicted purchase volume according to the behavior parameters of the matched first purchase constraint to determine the purchase volume of the first commodity, wherein the behavior parameters are used to express the method of modifying the predicted purchase volume under the condition parameters of the purchase constraint.

[0006] By adopting the above method, the purchase quantity of the goods can be automatically adjusted by predicting the purchase quantity and one of the first purchase constraints among multiple first purchase constraints, thereby reducing the work of modifying the predicted purchase quantity by the operation staff.

[0007] In a possible implementation, the first procurement constraint includes a characterization description and an implementation parameter; the characterization description is a description of the change in the sample under the current situation; the implementation parameters include conditional parameters and behavioral parameters, and the behavioral parameters are used to describe the method of changing the predicted procurement quantity under the conditional parameters that meet the procurement constraint.

[0008] In a possible implementation, the method further includes: if there is a second manually modified sample in the manually modified sample that does not match the characterization description of any of the procurement constraints, setting a new procurement constraint for the second manually modified sample by manual labeling.

[0009] By adopting the above method, the procurement constraints in the procurement constraint library can be improved by adding new procurement constraints, so that the procurement constraints in the procurement constraint library are more comprehensive and accurate.

[0010] In a possible implementation method, any first procurement constraint is a procurement constraint that matches a manual modification sample determined from various procurement constraints, including: for any procurement constraint in the procurement constraint library, determining a first manual modification sample that matches the conditional parameters of the procurement constraint from various manual modification samples; modifying the predicted procurement quantity in the first manual modification sample according to the behavioral parameters of the procurement constraint to obtain modified data under the procurement constraint; determining the effective state of the procurement constraint according to the modified data and the procurement quantity in the first manual modification sample; the effective state includes three types: cannot take effect, can take effect, and should take effect; for each procurement constraint, calculating the precision of the procurement constraint through the proportion of the procurement constraint that cannot take effect and the proportion of the procurement constraint that should take effect; and determining whether the procurement constraint is the first procurement constraint that matches the manual modification sample according to the precision of the procurement constraint.

[0011] By adopting the above method, the first procurement constraint can be screened out from the procurement constraint library to ensure the accuracy of the first procurement constraint.

[0012] In a possible implementation method, the effectiveness status of the procurement constraint is determined based on the procurement quantity in the modified data and the matched first artificially modified sample, including: if the modification tendency of the modified data is the same as the modification tendency of the matched first artificially modified sample, and the modified data does not exceed the procurement quantity in the matched first artificially modified sample, then determining that the procurement constraint should be in an effective state.

[0013] The above method provides a method for determining the effective state of a procurement constraint. According to the above method, it is more accurate and convenient to determine whether the procurement constraint is in the effective state.

[0014] In a possible implementation method, the method also includes: if the change tendency of the modified data is not the same as the change tendency of the matched first artificial modified sample, and the modified data does not exceed the purchase quantity in the matched first artificial modified sample, then the minimum purchase quantity of the commodity is calculated according to the out-of-stock fitting sales volume and out-of-stock rate of the commodity in the artificial modified sample; the maximum purchase quantity of the commodity is calculated according to the saleable period, preparation days and out-of-stock fitting sales volume of the commodity; and the effective status of the procurement constraint is determined according to the relationship between the predicted purchase quantity and the minimum purchase quantity and the relationship between the modified data and the maximum purchase quantity.

[0015] In a possible implementation method, any first procurement constraint is a procurement constraint that matches the manually modified sample determined from various procurement constraints, including: determining a third manually modified sample that is manually marked to meet the procurement constraint characterization description from various manually modified samples; determining a fourth manually modified sample that meets the procurement constraint behavior parameters from the third manually modified sample; calculating the recall rate of the procurement constraint based on the number of the third manually modified samples and the number of the fourth manually modified samples; and determining whether the procurement constraint is the first procurement constraint that matches the manually modified sample based on the recall rate of the procurement constraint.

[0016] In a possible implementation method, if the recall rate or precision rate of the procurement constraint does not meet the preset requirements, the condition parameters and behavior parameters of the procurement constraint are modified.

[0017] In a possible implementation method, the current inventory information includes in-stock inventory and in-transit inventory; the estimated sales volume includes the estimated sales volume from the current time to the target purchase arrival date.

[0018] In a possible implementation method, determining a first procurement constraint that matches the current status of the first commodity from condition parameters of multiple first procurement constraints includes: determining one or more first procurement constraints that match the current status of the first commodity from condition parameters of multiple first procurement constraints; and determining the first procurement constraint that matches the current status of the first commodity based on the accuracy of the matched one or more first procurement constraints.

[0019] In a second aspect, an embodiment of the present application provides a device for predicting purchase quantity, the device comprising: a determination module for determining the predicted purchase quantity of the first commodity based on the current inventory information and estimated sales volume of the first commodity; the determination module is also used to determine a first purchase constraint that matches the current status of the first commodity from the condition parameters of multiple first purchase constraints; any first purchase constraint is a purchase constraint determined from various purchase constraints that matches a manually modified sample; any manually modified sample represents the purchase quantity of the sample obtained by manually modifying the predicted purchase quantity of the sample under the condition in which the sample is located; a modification module is used to modify the predicted purchase quantity according to the behavior parameters of the matched first purchase constraint to determine the purchase quantity of the first commodity, wherein the behavior parameters are used to express a method for modifying the predicted purchase quantity under the condition parameters of the purchase constraint.

[0020] In a possible implementation, the device also includes a setting module, and the setting module is used to manually set a new procurement constraint for the second artificially modified sample if there is a second artificially modified sample in the artificially modified sample that does not match the characterization description of any of the procurement constraints.

[0021] In a possible implementation method, the determination module is also used to, for any procurement constraint in the procurement constraint library, determine a first manual modification sample that matches the conditional parameters of the procurement constraint from each manual modification sample; the modification module is also used to modify the predicted procurement quantity in the first manual modification sample according to the behavioral parameters of the procurement constraint to obtain modified data under the procurement constraint; the determination module is also used to determine the effectiveness status of the procurement constraint based on the modified data and the procurement quantity in the first manual modification sample; the device also includes a calculation module, which is used to, for each procurement constraint, calculate the accuracy of the procurement constraint through the proportion of the procurement constraint that cannot be effective and the proportion of the procurement constraint that should be effective; the determination module is also used to determine whether the procurement constraint is the first procurement constraint that matches the manual modification sample based on the accuracy of the procurement constraint.

[0022] In one possible implementation, the determination module is specifically used to determine that the procurement constraint should be effective if the modification tendency of the modified data is the same as the modification tendency of the matched first artificial modification sample, and the modified data does not exceed the procurement quantity in the matched first artificial modification sample.

[0023] In a possible implementation, the calculation module is further used to calculate the minimum purchase quantity of the commodity based on the out-of-stock fitting sales volume and out-of-stock rate of the commodity in the manually modified sample; and to calculate the maximum purchase quantity of the commodity based on the saleable period, inventory days and out-of-stock fitting sales volume of the commodity; the determination module is further used to determine the effective status of the procurement constraint based on the relationship between the predicted procurement quantity and the minimum procurement quantity and the relationship between the modified data and the maximum procurement quantity.

[0024] In one possible implementation, the determination module is specifically used to determine, from the various manually modified samples, a third manually modified sample that is manually marked to satisfy the procurement constraint characterization description; to determine, from the third manually modified sample, a fourth manually modified sample that satisfies the procurement constraint behavior parameters; the calculation module is also used to calculate the recall rate of the procurement constraint based on the number of the third manually modified samples and the number of the fourth manually modified samples; the determination module is specifically used to determine, based on the recall rate of the procurement constraint, whether the procurement constraint is the first procurement constraint that matches the manually modified sample.

[0025] In a possible implementation, the device further includes a modification module, and the modification module is used to modify the condition parameters and behavior parameters of the procurement constraint if the recall rate or the precision rate of the procurement constraint does not meet the preset requirements.

[0026] In one possible implementation, the determination module is specifically used to determine one or more first procurement constraints that match the current status of the first commodity from the condition parameters of multiple first procurement constraints; and determine the first procurement constraint that matches the current status of the first commodity based on the accuracy of the matched one or more first procurement constraints.

[0027] In a third aspect, an embodiment of the present application provides a device for predicting purchase volume, the device comprising a memory and a processor, the memory being used to store computer programs or instructions; the processor being used to call the computer programs or instructions stored in the memory to execute a method in any possible implementation of the first aspect.

[0028] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which instructions are stored. When a computer reads and executes the instructions, the computer executes a method in any possible implementation of the first aspect.

[0029] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein instructions are stored in the computer program product. When a computer reads and executes the instructions, the computer executes a method in any possible implementation of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0031] Figure 1 A schematic diagram of four steps from planned purchase to arrival of a commodity provided in an embodiment of the present application;

[0032] Figure 2 A flowchart corresponding to a method for predicting purchase volume provided in an embodiment of the present application;

[0033] Figure 3 A flowchart corresponding to a method for determining a first purchasing constraint provided in an embodiment of the present application;

[0034] Figure 4 A schematic diagram of an interface for matching manually modified samples according to a representation of procurement constraints is provided for an embodiment of the present application;

[0035] Figure 5 A schematic diagram of internal modules of a device 5000 for predicting purchase quantity provided in an embodiment of the present application;

[0036] Figure 6 A schematic diagram of the structure of a purchase quantity prediction device 6000 provided in an embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0038] Supplier-large warehouse-forward warehouse-customer is a classic supply chain system. In this supply chain system, procurement refers to the process of large warehouses placing orders with suppliers. The purchase quantity of goods is usually determined by the staff based on their own experience, or predicted based on a preset algorithm. However, the prediction of the algorithm is often not accurate enough and not flexible enough, requiring operators to further modify the predicted quantity of the algorithm. After that, the purchasing staff can also further modify the purchase quantity modified by the operators. For example, the operators can increase the purchase quantity of the most recently promoted goods, and the purchasing staff can change the quantity of goods that are less than a box to a whole box. In this case, the goods need to go through four links from planned procurement to the arrival of the goods. Figure 1 A schematic diagram of the four links from planned purchase to arrival of a commodity provided for an embodiment of the present application includes: algorithm-recommended quantity, operational quantity change, purchase quantity change and order placement, and actual arrival. Among them, the algorithm-recommended quantity can be determined based on the inventory in stock, inventory in transit, estimated sales and stocking target. The inventory in stock can represent the inventory of the large warehouse and the inventory of the forward warehouse; the inventory in transit can represent the inventory in transit; the estimated sales represent the estimated sales of the commodity for each day from the current moment to the target date of purchase arrival; the stocking target can represent the estimated sales and safety stock of the target sales day, and the safety stock is used to ensure the uncertainty of commodity sales. Operational quantity change generally depends on information such as the historical out-of-stock rate, loss rate, estimated accuracy and future activity plans of the commodity. Purchase quantity change and order placement generally consider information such as the minimum order quantity, order unit and supplier stability of the commodity. Actual arrival means that the commodity arrives at the large warehouse after the order is placed.

[0039] For the two links of operational quantity change and procurement quantity change and order placement, human considerations account for a large factor, and the basis for the quantity change is inaccurate. Different operators may have different quantity change results. In other words, the final purchase quantity of the goods obtained by the above method is highly random, and due to the wide variety of goods, if each product undergoes operational quantity change and procurement quantity change, the human resources consumed will be huge.

[0040] Based on this, an embodiment of the present application provides a method for predicting purchase volume, which is used to realize automatic adjustment of the predicted purchase volume and reduce the workload of operators in modifying the predicted purchase volume. Figure 2 A flowchart corresponding to a method for predicting purchase volume provided in an embodiment of the present application is provided. The flowchart can be executed by a device for predicting purchase volume, such as Figure 2 As shown, the process includes the following steps:

[0041] Step 201: The device for predicting the purchase quantity determines the predicted purchase quantity of the first commodity according to the current inventory information and the estimated sales volume of the first commodity.

[0042] Exemplarily, the current inventory information of the first commodity includes in-stock inventory, in-transit inventory and safety inventory. The in-stock inventory can be the inventory of the commodity in the large warehouse and the inventory in the forward warehouse. The in-transit inventory can be the inventory of the commodity in transit. The number of large warehouses and forward warehouses is not limited in the embodiments of this application. The safety inventory is the inventory used to ensure the uncertainty of commodity sales. When the actual sales volume of the commodity suddenly increases, the safety inventory of the commodity can be used. The estimated sales volume is the estimated sales volume of the first commodity from the current moment to the target date of procurement arrival. The estimated sales volume of the commodity is also calculated according to the algorithm. This application does not limit the algorithm for calculating the estimated sales volume. The current inventory information of the first commodity and the estimated sales volume on the target date are input into the preset procurement volume prediction algorithm to obtain the predicted procurement volume of the first commodity.

[0043] In step 202, the device for predicting the purchase quantity determines a first purchase constraint that matches the current status of the first commodity from the condition parameters of multiple first purchase constraints. Any first purchase constraint is a purchase constraint that matches the manually modified sample determined from various purchase constraints. Any manually modified sample represents the purchase quantity of the sample obtained by manually modifying the predicted purchase quantity of the sample under the status of the sample.

[0044] Exemplarily, a procurement constraint includes a condition parameter and a behavior parameter, wherein the condition parameter indicates the condition for hitting the procurement constraint, and the behavior parameter indicates the way to handle the predicted procurement volume after hitting the procurement constraint. For example, for the procurement constraint "If the out-of-stock rate of a product is lower than 200% of the target out-of-stock rate, increase the quantity by 20%", the out-of-stock rate of a product lower than 200% of the target out-of-stock rate is the condition parameter of the procurement constraint, and the increase in quantity by 20% is the behavior parameter of the procurement constraint. It is understandable that the procurement constraint is merely an example, and the embodiments of the present application do not limit the content of the procurement constraint.

[0045] The purchasing constraints also include a characterization description of the purchasing constraints, which is a description of the change in quantity of the sample under the current situation. The characterization description does not include specific parameters. The characterization descriptions provided in the embodiments of the present application include but are not limited to the following: when the out-of-stock rate of the product is high, increase the purchase quantity; when the damage rate of the product is high, reduce the purchase quantity; when the predicted purchase quantity has been increased in the previous few times, continue to increase the quantity; when the predicted purchase quantity has been reduced in the previous few times, continue to reduce the quantity; when the predicted purchase quantity of the product is higher than the estimated sales volume of the product, reduce the purchase quantity; when the predicted purchase quantity of the product is not an integer, round it to 10 or round it to the box gauge. There are also some characterization descriptions in special scenarios, such as the processing method for new products or the processing method for active products. In addition, there are some characterization descriptions of invalid purchasing constraints. This part of the purchasing constraints aggravates the out-of-stock or loss of the product. The statistics of invalid purchasing constraints can feed back the invalid purchasing constraints to the operators and purchasing personnel so that the operators and purchasing personnel can standardize their own change operations.

[0046] The artificially modified sample represents the purchase quantity of the sample obtained by artificially modifying the predicted purchase quantity of the sample under the situation in which the sample is located, that is, any artificially modified sample includes the predicted purchase quantity of the commodity predicted by the purchase quantity prediction algorithm, and the purchase quantity and commodity status information obtained after the predicted purchase quantity is manually modified. For example, a manually modified sample can be "a commodity in a large warehouse on a certain date, the saleable period is 3 days, the predicted purchase quantity of the purchase quantity prediction algorithm is 947, the operational modification is 900, the average out-of-stock rate in the first seven days is 10%, the average loss rate is 2%, the average out-of-stock rate in the first three days is 7%, the average loss rate is 3%, there are reduction operations for 3 consecutive days, the stocking target is 1180, of which the estimated sales volume is 800, the safety inventory is 380, and the actual sales volume after out-of-stock compensation is 613". It can be understood that this artificially modified sample is only an example, and this application does not limit the specific content of the artificially modified sample. A normal manually modified sample data may include the current status of the product, such as the out-of-stock rate and loss rate in the above example, as well as the predicted purchase quantity of the product predicted by the purchase quantity prediction algorithm, the modification of the predicted purchase quantity by the operation staff, and the further modification of the data modified by the operation staff by the purchase staff to obtain the final purchase quantity of the product, and the actual sales effect or sales volume after actual out-of-stock compensation.

[0047] It should be noted that after determining the purchasing constraints, it can be used as a reference processing method for the quantity change processing method of the previous link, that is, the present application can determine the purchasing constraints that meet the quantity change habits of the operating personnel and the purchasing constraints that meet the quantity change habits of the purchasing personnel, and then adjust the purchasing quantity predicted by the purchasing quantity prediction algorithm based on the purchasing constraints of the operation change quantity, and adjust the quantity change of the operating personnel based on the purchasing constraints of the purchase change quantity. After multiple iterations, the purchase quantity predicted by the purchase quantity prediction algorithm can be adjusted to obtain a purchase quantity that is more suitable for the product, and then the workload of the operating personnel and the purchasing personnel for the commodity change quantity is gradually reduced, so that the operating personnel and the purchasing personnel do not need to change the quantity of the commodity in the end. Therefore, in order to determine the purchasing constraints that meet the operation change quantity, it is necessary to screen out samples in which the purchasing personnel have almost no change quantity, so that the final sales of the commodity can depend on the purchase quantity after the operation personnel change quantity; similarly, in order to determine the purchasing constraints that meet the purchase change quantity, it is necessary to screen out samples in which the operating personnel have almost no change quantity, so that the final sales of the commodity can depend on the purchase quantity after the purchase change quantity. The method for determining procurement constraints that meet procurement quantity changes in the embodiment of the present application is the same as the method for determining procurement constraints that meet operational quantity changes, and is described here using the example of determining procurement constraints that meet operational quantity changes.

[0048] Specifically, any first purchasing constraint included in step 202 is a purchasing constraint that is determined from among various purchasing constraints and matches the artificially modified sample. Figure 3 A flowchart corresponding to a method for determining a first procurement constraint provided in an embodiment of the present application.

[0049] Step 301: label the manually modified samples based on the representations of the procurement constraints.

[0050] For any procurement constraint in the procurement constraint library, determine the manual quantity change sample that matches the representation description of each procurement constraint from each manual quantity change sample. The representation description of one procurement constraint can match multiple manual quantity change samples, and one manual quantity change sample can also match multiple procurement constraint representation descriptions. The representation descriptions of multiple procurement constraints can be the same. For example, one procurement constraint is "If the predicted procurement volume of the procurement volume prediction algorithm is higher than the estimated sales volume by 10% to 20%, the predicted procurement volume is reduced by 5%", and the other procurement constraint is "If the predicted procurement volume of the procurement volume prediction algorithm is higher than the estimated sales volume by 20% to 30%, the predicted procurement volume is reduced by 10%". The representation descriptions of these two procurement constraints are the same, and both can be "If the predicted procurement volume of the procurement volume prediction algorithm is higher than the estimated sales volume, the predicted procurement volume is reduced". Among them, matching the manual quantity change samples according to the representation description of the procurement constraint can be performed by the system to automatically match the manual quantity change samples first, and then the procurement constraints matched by the system can be manually modified or determined. Specifically, the embodiment of the present application provides a schematic diagram of an interface for matching manually modified samples according to a representation of a procurement constraint, such as Figure 4 As shown, a manual quantity modification sample will pop up on the system interface one by one, and the system marking box will automatically pop up the procurement constraints marked by the system for the popped-up manual quantity modification sample. The staff confirms or modifies the procurement constraints marked by the system based on their own experience. If the procurement constraints automatically marked by the system are correct, the staff clicks Confirm, and then clicks OK, and the next manual quantity modification sample will automatically pop up on the system; if the staff needs to modify the procurement constraints automatically marked by the system, the staff clicks Modify, modifies the procurement constraints in the manual marking box, and then clicks OK. Figure 4 It is only a schematic diagram of the interface for marking and describing the manually modified samples in the embodiment of the present application, and the embodiment of the present application is not limited to the specific schematic diagram of the interface.

[0051] In order to enable the staff to more accurately mark the manually modified samples, the system can optionally provide statistical analysis of the manually modified samples. Specifically, it is stipulated that samples with a change amplitude less than α1 and a change rate less than β1 are artificial reductions, and samples with a change amplitude greater than α2 and a change rate greater than β2 are artificial reductions. The range of α1 and α2 is [- predicted purchase volume, +∞), the range of β1 and β2 is [-1, +∞), and the others are unmodified. The unmodified samples include samples with no changes and samples with slight changes. The change amplitude is the purchase volume after manual change minus the predicted purchase volume, and the change rate is the ratio of the change amplitude to the predicted purchase volume. Based on the samples with manual additions, samples with manual reductions and samples without changes, the proportion of each type of samples in the total samples is calculated. Optionally, the proportion of the categories to which the goods belong can also be counted, such as the proportion of fresh food, the proportion of baked goods, and the proportion of meat and poultry in all the samples of quantity changes; the proportion of the warehouses to which the goods belong can also be counted, such as the proportion of the first warehouse, the proportion of the second warehouse, and the proportion of the major warehouses in all the samples of manual quantity changes. Optionally, the proportion of the samples of top product quantity changes and the proportion of the samples of non-top product quantity changes can also be counted. Optionally, the distribution performance of the quantity change amplitude and the quantity change rate can also be counted. Taking the distribution performance of the quantity change amplitude as an example, the proportion of samples with a quantity change amplitude of 0 to 10, the proportion of samples with a quantity change amplitude of 10 to 50, the proportion of samples with a quantity change amplitude of 50 to 200, and the proportion of samples with a quantity change amplitude of more than 200 can be counted in the case of manual quantity addition. Optionally, some special cases can also be counted, such as the proportion of samples with a predicted purchase volume less than a certain proportion of the estimated sales volume or the proportion of samples with a predicted purchase volume greater than a certain proportion of the estimated sales volume. Other statistical analyses may also be performed on the actual modified samples according to actual conditions. The embodiment of the present application does not specifically limit the statistical analysis method of the modified samples.

[0052] In addition to statistical analysis of samples after quantity changes, statistical analysis can also be performed on the predicted purchase volume in the manually changed samples and the purchase volume after manual changes to confirm the tendency of quantity changes. For example, the tendency of quantity changes based on average daily sales volume, the saleable period, and the ratio of predicted purchase volume to predicted sales volume can be used to confirm the tendency of quantity changes under each basis. Specifically, the average daily sales volume is used as an example to illustrate that the predicted purchase volume and the value after manual changes are statistically analyzed for commodities with different average daily sales volumes. The tendency of quantity changes based on the saleable period and the ratio of predicted purchase volume to predicted sales volume will not be elaborated on here.

[0053] Based on the comparative statistical analysis before and after manual modification and the analysis of the proportion of samples after modification, the staff can more clearly mark each manually modified sample with accurate procurement constraints.

[0054] Step 302: If there is a second manually modified sample in the manually modified sample that does not match the representation description of any procurement constraint in each procurement constraint, a new procurement constraint is manually set for the second manually modified sample.

[0055] In this case, the interface diagram for matching the manually modified samples according to the representation of the procurement constraints can also include a new button, and the newly added procurement constraints also include representations, conditional parameters, and behavioral parameters, and the newly added procurement constraints are also added to the procurement constraint library. The samples annotated with the manually modified samples will be automatically written into the database, and the annotated manually modified samples can be written into the database every several samples or every fixed time length. Of course, the procurement constraints that originally existed in the procurement constraint library were also newly created by the previous staff.

[0056] Step 303: modify the first artificial modification sample based on the behavior parameters of the procurement constraint.

[0057] For any procurement constraint in the procurement constraint library, determine the first manual quantity modification sample that matches the conditional parameters of each procurement constraint from each manual quantity modification sample. A procurement constraint can match multiple first manual quantity modification samples, and a manual quantity modification sample can also match different procurement constraints. Exemplarily, if one of the procurement constraints in the procurement constraint library is "If the predicted procurement volume of the procurement volume prediction algorithm is 10% to 20% higher than the estimated sales volume, then the predicted procurement volume is reduced by 5%", and the predicted procurement volume of the procurement volume prediction algorithm in the above-mentioned manual quantity modification sample is 947, and the estimated sales volume is 800, then the procurement constraint matches the manual quantity modification sample. According to the behavior parameters of the procurement constraint, the predicted procurement volume in the first manual quantity modification sample is modified to obtain the modified data under the procurement constraint. Using the above example, the behavior parameters of the procurement constraint are to reduce the predicted procurement volume by 5%, then the modified data under the procurement constraint is 947-947*5%=899.65. That is to say, after the artificially modified sample in the above example is modified by the purchase constraint, the modified sample is "the predicted purchase quantity of the purchase quantity prediction algorithm is 947, and the data after the purchase constraint modification is 899.65". Then this modified sample matches the first artificially modified sample through the purchase constraint.

[0058] Step 304, calculating the accuracy of the procurement constraints.

[0059] Exemplarily, there are 1000 manually modified samples. Since one manually modified sample can match one or more procurement constraints, one manually modified sample forms a modified sample after passing through a procurement constraint. Therefore, 1000 manually modified samples can form more than 1000 modified samples after matching the procurement constraints in the procurement constraint library. Assume that 2000 modified samples are formed. Therefore, each of the 2000 modified samples corresponds to a procurement constraint.

[0060] According to the modified data and the purchase quantity in the first artificial modified sample, it is determined whether the purchase constraint is effective, that is, according to the sample after the purchase constraint is modified and the first artificial modified sample, it is determined whether the purchase constraint is effective in the modified sample. It should be noted that a first artificial modified sample can match different purchase constraints, and according to different behavior parameters of the purchase constraints, modified samples corresponding to multiple first artificial modified samples can be obtained.

[0061] The method of determining whether the purchase constraint is effective according to the purchase quantity in the first artificially modified sample and the modified data after the quantity change is the same as the purchase quantity in the first artificially modified sample, and the modified data does not exceed the purchase quantity in the first artificially modified sample, then the purchase constraint is in the effective state. Specifically, the change tendency refers to whether the purchase quantity is increased or decreased relative to the predicted purchase quantity. For example, the predicted purchase quantity of the purchase quantity prediction algorithm is 170, and the data after the purchase constraint change is 180. Therefore, the change tendency of the modified data is increased. The first artificially modified sample matched by the modified sample is artificially changed to 200, and the change tendency of the artificially modified sample is also increased. If the change tendency of the modified data is the same as the change tendency of the matched artificially modified sample, and the modified data 180 does not exceed the purchase quantity 200 in the artificially modified sample, then the purchase constraint corresponding to the modified sample is in the effective state on the modified sample.

[0062] If the change tendency of the modified data is not the same as the change tendency of the first artificial modified sample, and the modified data does not exceed the purchase volume in the first artificial modified sample, that is, if the change tendency of the modified data is different from the change tendency of the first artificial modified sample, or the modified data exceeds the purchase volume in the first artificial modified sample, then the minimum purchase volume of the product is calculated according to the out-of-stock fitting sales volume and out-of-stock rate of the product in the artificial modified sample. Specifically, the out-of-stock fitting sales volume is the sales volume value assuming that the product is not out of stock, and the out-of-stock rate includes the allowable out-of-stock rate and the actual out-of-stock rate of the product, and the allowable out-of-stock rate is the maximum value of the allowable out-of-stock rate of the product. For example, to calculate the minimum purchase volume of the product, if the allowable out-of-stock rate of the product is 2% and the actual out-of-stock rate of the product is 1%, then the minimum purchase volume of the product is obtained according to the allowable out-of-stock rate of the product; if the allowable out-of-stock rate of the product is 2% and the actual out-of-stock rate of the product is 3%, then the minimum purchase volume of the product is obtained according to the allowable out-of-stock rate and the out-of-stock fitting sales volume. The maximum purchase quantity of the product is calculated based on the saleable period, stocking days and out-of-stock fitting sales of the product. Specifically, the saleable period is the shelf life of the product, which is assumed to be 7 days, and the stocking days are the days covered by the purchase of the product, which is assumed to be 5 days. The maximum purchase quantity of the product is calculated based on the minimum value of the saleable period and the stocking days, which is 5 days, and the out-of-stock fitting sales of the product. The effective state of the purchase constraint is determined based on the relationship between the predicted purchase quantity and the minimum purchase quantity and the relationship between the modified data and the maximum purchase quantity. Specifically, the effective state of the purchase constraint is determined based on the relationship between the predicted purchase quantity and the minimum purchase quantity and the relationship between the modified data and the maximum purchase quantity. It can be determined by the following Table 1.

[0063] Table 1

[0064]

[0065] Among them, a is the minimum purchase quantity, b is the maximum purchase quantity, x is the predicted purchase quantity of the purchase quantity prediction algorithm, and y is the data after the predicted purchase quantity is modified according to the matching purchase constraints. The last row of the table is not modified, which means that y is changed very little compared to x. Take a row in Table 1 as an example. If the data after the purchase constraint modification is an increase compared to the predicted purchase quantity, and the predicted purchase quantity is less than the minimum purchase quantity, and the modified data is less than or equal to the maximum purchase quantity, then the purchase constraint should be effective on the modified sample.

[0066] For each procurement constraint, the effective state of the procurement constraint is used to determine whether the procurement constraint is the first procurement constraint matched by the artificially modified sample. According to the above Table 1, the effective state of each procurement constraint is determined. The precision of each procurement constraint is determined by the effective state of each procurement constraint. Specifically, it can be seen from the above Table 1 that each procurement constraint can have three effective states, including a state that can be effective, a state that should be effective, and a state that cannot be effective. The precision of the procurement constraint is determined based on the proportion of the ineffective state and the proportion of the effective state of the procurement constraint. If the proportion of the ineffective state of the procurement constraint is lower than the first preset threshold and the proportion of the effective state is higher than the second preset threshold, the precision of the procurement constraint meets the requirements. For example, there is a procurement constraint in the procurement constraint library, which is procurement constraint 1. Among the 2000 modified samples, 50 samples hit procurement constraint 1. Among the 50 modified samples, the states of procurement constraint 1 that should be effective, can be effective, and cannot be effective are 40, 7, and 2 respectively. Therefore, the proportion of procurement constraint 1 that should be effective is 4 / 5, and the proportion of ineffective state is 1 / 25. If the first preset threshold is 0.1 and the second preset threshold is 0.7, the precision of procurement constraint 1 meets the set requirements. In addition, the recall rate of procurement constraint 1 needs to be calculated. If both the precision and recall rate of procurement constraint 1 meet the set requirements, procurement constraint 1 is the first procurement constraint that matches the manually modified sample.

[0067] Step 305, calculating the recall rate of the procurement constraints.

[0068] Determine from each manually modified sample a third manually modified sample whose manual marking satisfies the procurement constraint characterization description; determine from the third manually modified sample a fourth manually modified sample that satisfies the procurement constraint behavior parameters; and calculate the recall rate of the procurement constraint based on the number of the third manually modified samples and the number of the fourth manually modified samples.

[0069] For example, if there are 32 third manually modified samples that are manually marked and meet the characterization description of procurement constraint 1 among the manually modified samples, if the manually modified samples are matched according to the condition parameters and behavior parameters of procurement constraint 1, and 2 fourth manually modified samples are found to match procurement constraint 1, then the recall rate of procurement constraint 1 is 1 / 16. If the recall rate of the procurement constraint is higher than the preset recall rate threshold, the recall rate of the procurement constraint meets the set requirements.

[0070] If there is no manually marked procurement constraint 2 in each manually modified sample, it is assumed that the recall rate of procurement constraint 2 meets the set requirements.

[0071] Step 306, determining whether the precision and recall of the procurement constraint meet the set requirements, if so, determining the procurement constraint as the first procurement constraint.

[0072] According to the precision rate and recall rate of each procurement constraint, determine whether it is the first procurement constraint that matches the manually modified sample. If the precision rate and recall rate of the procurement constraint meet the set requirements, the procurement constraint can be the first procurement constraint. According to the behavior parameters of the first procurement constraint, the predicted procurement quantity is modified to determine the procurement quantity of the first commodity.

[0073] Step 307: If the precision and recall of the procurement constraint do not meet the set requirements, the condition parameters and behavior parameters of the procurement constraint are modified.

[0074] If the recall rate or precision rate of the procurement constraint does not meet the preset requirements, the condition parameters and behavior parameters of the procurement constraint are modified based on the staff's experience or parameter search method, and the parameter search method can be a grid search. For example, one of the specific procurement constraints is "If the actual out-of-stock rate of the commodity on the previous day exceeds 200% of the target out-of-stock rate of the commodity, then increase the quantity by 50%" can be modified to "If the actual out-of-stock rate of the commodity on the previous day exceeds 150% of the target out-of-stock rate of the commodity, then increase the quantity by 20%." The condition parameters and behavior parameters of the procurement constraint are modified so that the precision and recall rate of the procurement constraint meet the set requirements.

[0075] Step 308: If the condition parameters and behavior parameters of the purchase constraint are modified for more than a certain number of times, and the accuracy and recall rates still do not meet the set requirements, the purchase constraint is deleted.

[0076] Specifically, if the condition parameters and behavior parameters of a procurement constraint are modified more than a certain number of times and still do not meet the set requirements of accuracy and recall, the procurement constraint will be deleted from the procurement constraint library.

[0077] Based on the calculation of the precision and recall rate of the procurement constraints in the procurement constraint library and the modification of the condition parameters and behavior parameters of the procurement constraints, multiple first procurement constraints are determined, and the multiple first procurement constraints are used to modify the predicted procurement quantity of the goods to obtain the procurement quantity of the goods. It should be noted that the multiple first procurement constraints determined after a round of iteration are not necessarily suitable for all goods. For example, there are 1,000 goods in the warehouse. For each commodity, the current status and predicted procurement quantity of each commodity are determined. According to the condition parameters of multiple first procurement constraints, 1,000 samples are matched. Only 500 samples match the first procurement constraint, and 500 samples do not have the corresponding first procurement constraint. At this time, it is necessary to re-acquire a part of the manually modified samples, and mark the re-acquired manually modified samples according to the representation description of multiple first procurement constraints. In this process, the staff can create new procurement constraints. In this way, after multiple rounds of new procurement constraints, the precision and recall of procurement constraints are calculated to determine multiple first procurement constraints, which can make the first procurement constraints more and more accurate and comprehensive.

[0078] A first procurement constraint matching the current status of the first commodity is determined from the conditional parameters of multiple first procurement constraints. The current status of the first commodity includes but is not limited to the first commodity in a large warehouse within a specified date range and related data of the first commodity, such as the out-of-stock rate of the first commodity in the past few days, the estimated sales volume of the first commodity in the next few days, the commodity attributes of the first commodity, and whether the first commodity is a top product. The current status information of the first commodity can be obtained from the database. The number of first procurement constraints matching the current status of the first commodity determined from the conditional parameters of multiple first procurement constraints can be one or more. For example, if the number of first procurement constraints matching the current status of the first commodity is three, then one of the procurement constraints with the highest proportion is determined based on the proportion of the effective status of the three first procurement constraints.

[0079] Step 203: The device for predicting the purchase quantity modifies the predicted purchase quantity according to the behavior parameter of the matched first purchase constraint to determine the purchase quantity of the first commodity.

[0080] Specifically, the predicted purchase quantity of the purchase quantity prediction algorithm is modified according to the determined behavior parameter of the first purchase constraint to determine the purchase quantity of the first commodity. For example, the predicted purchase quantity of the first commodity is 1147, and the predicted purchase quantity is modified according to the behavior parameter of the first purchase constraint to determine that the purchase quantity of the first commodity is 1300.

[0081] The predicted purchase quantity output by the purchase quantity prediction algorithm combined with the post-placed first purchase constraint can determine the final purchase quantity of the product. When the first purchase constraint changes, the purchase quantity of the product can be flexibly changed. If the first purchase constraint is integrated into the purchase quantity prediction algorithm to modify the purchase quantity prediction algorithm, when the first purchase constraint changes, the purchase quantity prediction algorithm also needs to be modified, increasing the workload of determining the purchase quantity of the product. This will reduce the amount of changes to the predicted purchase quantity by the operator. Eventually, as the first purchase constraint becomes more and more accurate, the operator will no longer need to modify the predicted purchase quantity.

[0082] Figure 5 A schematic diagram of the internal modules of a device 5000 for predicting purchase volume provided in an embodiment of the present application is shown in FIG. Figure 5 As shown, the device may include: a determination module 501, a modification module 502, a setting module 503, a calculation module 504, and a modification module 505.

[0083] Among them, the determination module 501 is used to determine the predicted purchase quantity of the first commodity based on the current inventory information and estimated sales volume of the first commodity; and to determine the first purchase constraint that matches the current status of the first commodity from the condition parameters of multiple first purchase constraints; any first purchase constraint is a purchase constraint that matches the manually modified sample determined from various purchase constraints; any manually modified sample represents the purchase quantity of the sample obtained by manually modifying the predicted purchase quantity of the sample under the condition of the sample; the modification module 502 is used to modify the predicted purchase quantity according to the behavior parameters of the matched first purchase constraint to determine the purchase quantity of the first commodity, wherein the behavior parameters are used to express the method of modifying the predicted purchase quantity under the condition parameters of the purchase constraint.

[0084] In a possible implementation, the device also includes a setting module 503, and the setting module 503 is used to manually set a new procurement constraint for the second artificially modified sample if there is a second artificially modified sample in the artificially modified sample that does not match the representation description of any of the procurement constraints.

[0085] In a possible implementation method, the determination module 501 is also used to, for any procurement constraint in the procurement constraint library, determine a first manual modification sample that matches the conditional parameters of the procurement constraint from each manual modification sample; the modification module 502 is also used to modify the predicted procurement quantity in the first manual modification sample according to the behavioral parameters of the procurement constraint to obtain modified data under the procurement constraint; the determination module 501 is also used to determine the effective state of the procurement constraint according to the modified data and the procurement quantity in the first manual modification sample; the device also includes a calculation module 504, and the calculation module 504 is used to, for each procurement constraint, calculate the accuracy of the procurement constraint through the proportion of the procurement constraint that cannot be effective and the proportion that should be effective; the determination module 501 is also used to determine whether the procurement constraint is the first procurement constraint that matches the manual modification sample according to the accuracy of the procurement constraint.

[0086] In one possible implementation, the determination module 501 is specifically used to determine that the procurement constraint should be effective if the modification tendency of the modified data is the same as the modification tendency of the matched first artificial modification sample, and the modified data does not exceed the procurement quantity in the matched first artificial modification sample.

[0087] In a possible implementation, the calculation module 504 is also used to calculate the minimum purchase quantity of the commodity based on the out-of-stock fitting sales volume and out-of-stock rate of the commodity in the manually modified sample; and to calculate the maximum purchase quantity of the commodity based on the saleable period, inventory days and out-of-stock fitting sales volume of the commodity; the determination module 501 is also used to determine the effective status of the procurement constraint based on the relationship between the predicted procurement quantity and the minimum procurement quantity and the relationship between the modified data and the maximum procurement quantity.

[0088] In a possible implementation, the determination module 501 is specifically used to determine, from the various manually modified samples, a third manually modified sample that is manually marked to satisfy the procurement constraint characterization description; to determine, from the third manually modified sample, a fourth manually modified sample that satisfies the procurement constraint behavior parameters; the calculation module is also used to calculate the recall rate of the procurement constraint based on the number of the third manually modified samples and the number of the fourth manually modified samples; the determination module is specifically used to determine, based on the recall rate of the procurement constraint, whether the procurement constraint is the first procurement constraint that matches the manually modified sample.

[0089] In a possible implementation, the device further includes a modification module 505, and the modification module 505 is used to modify the condition parameters and behavior parameters of the procurement constraint if the recall rate or the precision rate of the procurement constraint does not meet the preset requirements.

[0090] In a possible implementation, the determination module 501 is specifically used to determine one or more first procurement constraints that match the current status of the first commodity from the condition parameters of multiple first procurement constraints; and determine the first procurement constraint that matches the current status of the first commodity based on the accuracy of the matched one or more first procurement constraints.

[0091] Figure 6 Schematic diagram of a device 6000 for predicting purchase volume provided in an embodiment of the present application. Figure 6 As shown, it includes at least one processor 601 and a memory 602 connected to the at least one processor 601. The specific connection medium between the processor 601 and the memory 602 is not limited in the embodiment of the present application. Figure 6 For example, the processor 601 and the memory 602 are connected via a bus. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0092] In the embodiment of the present invention, the memory 602 stores instructions that can be executed by at least one processor 601. The at least one processor 601 can implement the steps of the above-mentioned method for predicting the purchase quantity by executing the instructions stored in the memory 602.

[0093] The processor 601 is the control center of the computer device, and can use various interfaces and lines to connect various parts of the computer device, and perform resource settings by running or executing instructions stored in the memory 602 and calling data stored in the memory 602. Optionally, the processor 601 may include one or more processing units, and the processor 601 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 601. In some embodiments, the processor 601 and the memory 602 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on independent chips.

[0094] Processor 601 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware processor for execution, or can be executed by a combination of hardware and software modules in the processor.

[0095] The memory 602 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 602 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (Random Access Memory, RAM), a static random access memory (Static Random Access Memory, SRAM), a programmable read-only memory (Programmable Read Only Memory, PROM), a read-only memory (Read Only Memory, ROM), an electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), a magnetic memory, a disk, an optical disk, etc. The memory 602 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 602 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0096] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0097] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0098] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0099] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0100] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for predicting purchase volume, characterized in that: The method comprises: Determining a predicted purchase quantity of the first commodity based on current inventory information and estimated sales volume of the first commodity; A first purchasing constraint matching the current status of the first commodity is determined from the condition parameters of the plurality of first purchasing constraints; any first purchasing constraint is a purchasing constraint matching the manually modified quantity sample determined from the purchasing constraints; any manually modified quantity sample represents the purchase quantity of the sample obtained by manually modifying the predicted purchase quantity of the sample under the status of the sample; The predicted purchase quantity is modified according to the behavior parameter of the matched first purchase constraint to determine the purchase quantity of the first commodity.

2. The method according to claim 1, characterized in that The first procurement constraint includes a characterization description and a realization parameter; The characterization description is a description of the change in the sample under the current condition; The implementation parameters include condition parameters and behavior parameters. The behavior parameters are used to express the method of modifying the predicted purchase quantity under the condition parameters that meet the purchase constraints.

3. The method according to claim 1, characterized in that The method further comprises: If there is a second manually modified sample in the manually modified samples that does not match the representation description of any procurement constraint in the procurement constraints, a new procurement constraint is set for the second manually modified sample by manual marking.

4. The method according to claim 1, characterized in that: Any first purchasing constraint is a purchasing constraint that matches the artificially modified sample determined from various purchasing constraints, including: For any procurement constraint in the procurement constraint library, determine a first manual quantity modification sample that matches the condition parameter of the procurement constraint from each manual quantity modification sample; Modifying the predicted purchase quantity in the first manually modified sample according to the behavior parameter of the purchase constraint to obtain modified data under the purchase constraint; Determining the effective state of the procurement constraint according to the modified data and the procurement quantity in the matched first manually modified sample; The effective status includes three types: cannot be effective, can be effective, and should be effective; for each procurement constraint, the accuracy of the procurement constraint is calculated by the proportion of the procurement constraint that cannot be effective and the proportion of the procurement constraint that should be effective; According to the accuracy of the procurement constraint, it is determined whether the procurement constraint is a first procurement constraint that matches the artificially modified sample.

5. The method according to claim 4, characterized in that Determining the effective state of the procurement constraint according to the modified data and the procurement quantity in the matched first manually modified sample includes: If the modification tendency of the modified data is the same as the modification tendency of the matched first artificial modification sample, and the modified data does not exceed the purchase amount in the matched first artificial modification sample, it is determined that the purchase constraint should be effective.

6. The method according to claim 5, characterized in that The method further comprises: If the change tendency of the modified data is not the same as the change tendency of the matched first artificially modified sample, and the modified data does not exceed the purchase quantity in the matched first artificially modified sample, then the minimum purchase quantity of the commodity is calculated according to the out-of-stock fitting sales volume and out-of-stock rate of the commodity in the artificially modified sample; Calculate the maximum purchase quantity of the product based on the saleable period, stocking days and out-of-stock fitting sales volume of the product; The effective state of the procurement constraint is determined based on the relationship between the predicted procurement volume and the minimum procurement volume and the relationship between the modified data and the maximum procurement volume.

7. The method according to claim 1, characterized in that Any first purchasing constraint is a purchasing constraint that matches the artificially modified sample determined from various purchasing constraints, including: Determine, from each manually modified sample, a third manually modified sample whose manual marking satisfies the procurement constraint characterization description; Determining a fourth artificially modified sample satisfying the procurement constraint behavior parameter from the third artificially modified sample; Calculating the recall rate of the procurement constraint according to the number of the third manually modified samples and the number of the fourth manually modified samples; According to the recall rate of the procurement constraint, it is determined whether the procurement constraint is a first procurement constraint that matches the manually modified sample.

8. The method according to claim 4 or 7, characterized in that: The method further comprises: If the recall rate or precision rate of the procurement constraint does not meet the preset requirements, the condition parameters and behavior parameters of the procurement constraint are modified.

9. The method according to claim 1, characterized in that: The current inventory information includes inventory in stock and inventory in transit; the estimated sales volume includes the estimated sales volume from the current time to the target date of procurement arrival.

10. The method according to claim 1, characterized in that Determining a first procurement constraint that matches the current status of the first commodity from among the condition parameters of the plurality of first procurement constraints includes: Determine one or more first procurement constraints that match the current status of the first commodity from the condition parameters of multiple first procurement constraints; and determine the first procurement constraint that matches the current status of the first commodity based on the accuracy of the matched one or more first procurement constraints.

11. A device for predicting purchase volume, characterized in that: include: A determination module, configured to determine a predicted purchase quantity of the first commodity based on current inventory information and estimated sales volume of the first commodity; The determination module is further used to determine, from the condition parameters of the plurality of first procurement constraints, a first procurement constraint that matches the current status of the first commodity; any first procurement constraint is a procurement constraint that matches the manually modified quantity sample determined from the procurement constraints; any manually modified quantity sample represents the procurement quantity of the sample obtained by manually modifying the predicted procurement quantity of the sample under the status of the sample; A quantity modification module is used to modify the predicted purchase quantity according to the behavioral parameters of the matched first purchase constraint to determine the purchase quantity of the first commodity, wherein the behavioral parameters are used to express a method for modifying the predicted purchase quantity under the condition parameters of the purchase constraint.

12. A device for predicting purchase volume, characterized in that: include: Memory, used to store computer programs or instructions; A processor, configured to call a computer program or instruction stored in the memory to execute the method according to any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when a computer reads and executes the instructions, the computer executes the method according to any one of claims 1 to 10.

14. A computer program product, characterized in that The computer program product stores instructions, and when a computer reads and executes the instructions, the computer executes the method according to any one of claims 1 to 10.

Images