CN117236850A - Control method, device, equipment and storage medium for replenishment data - Google Patents
Control method, device, equipment and storage medium for replenishment data Download PDFInfo
- Publication number
- CN117236850A CN117236850A CN202311245803.9A CN202311245803A CN117236850A CN 117236850 A CN117236850 A CN 117236850A CN 202311245803 A CN202311245803 A CN 202311245803A CN 117236850 A CN117236850 A CN 117236850A
- Authority
- CN
- China
- Prior art keywords
- data
- value
- target node
- predicted
- node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 74
- 238000004364 calculation method Methods 0.000 claims abstract description 16
- 238000004422 calculation algorithm Methods 0.000 claims description 26
- 238000006243 chemical reaction Methods 0.000 claims description 15
- 238000007781 pre-processing Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 12
- 238000009499 grossing Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 5
- 238000000513 principal component analysis Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 5
- 230000005855 radiation Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 230000003068 static effect Effects 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a control method, a device, equipment and a storage medium of replenishment data, which comprise the following steps: determining a plurality of hot spot models of the target object, and determining a first predicted sales value of the target node according to the hot spot model historical sales data; clustering the node portrait data of the nodes with the same target node type to obtain a plurality of same cluster nodes and determining a second prediction sales value according to the historical sales data; determining a value of a repurchase parameter of the user according to the user portrait data and determining a third predicted sales value according to the value; and carrying out weighted calculation on the first predicted pin value, the second predicted pin value and the third predicted pin value, and determining the final predicted pin value of the target node so as to carry out online replenishment data control. The application considers the buyback parameter values of the same cluster nodes in the nodes with the same target node type and the users around the target node, thereby being capable of dispatching the replenishment data more accurately compared with the prior art by only considering the historical factors or other partial factors.
Description
Technical Field
The present application relates to the field of integrated learning, and in particular, to a method, an apparatus, a device, and a storage medium for controlling replenishment data.
Background
Offline stores remain an important channel of retail sales. In contrast to online retail, the pain point of offline store operation is the scheduling of restocking data.
In the prior art, the following methods are used for dispatching the replenishment data: firstly, dispatching replenishment data according to store history sales volume; secondly, dispatching the replenishment data according to the lower limit threshold value of the stock; and thirdly, dispatching the replenishment data according to the warehouse, store and product inventory replenishment history.
The existing method for scheduling the replenishment data only considers historical factors or other partial factors, and cannot schedule the replenishment data accurately, so that on one hand, smooth sales of products in shops is not enough, on the other hand, excessive sales of common products are caused, stagnation is generated, the problems of large overall stock quantity, higher stock age, unreasonable stock structure and the like are caused, and the management risk and management cost are greatly increased.
Disclosure of Invention
The application provides a control method, a device, equipment and a storage medium of replenishment data, which are used for solving the technical problem that the replenishment data cannot be accurately scheduled only by considering historical factors or other partial factors in the prior art.
In a first aspect, the present application provides a method for controlling restocking data, the method comprising:
User portrait data of a user in a preset area range of a target node are obtained from a resource server, a plurality of hot spot models of a target object are determined according to the user portrait data, and a first predicted sales value of the target node is determined according to historical sales data of the hot spot models;
acquiring node portrait data of nodes with the same type as the target node from the resource server, clustering the node portrait data to obtain a plurality of same clustering nodes, and determining a second predicted sales value of the target node according to the history sales data of the same clustering nodes;
determining a repurchase parameter value of a user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value;
weighting the first predicted pin value, the second predicted pin value and the third predicted pin value to determine a final predicted pin value of the target node;
and performing online replenishment data control according to the final predicted sales value of the target node.
In an preferable technical solution of the control method for replenishment data, the determining the first predicted sales value of the target node according to the historical sales data of the hot spot model includes:
In the historical sales data of the hot spot model, a historical sales time sequence is formed in a fixed period, the historical sales time sequence is processed through an exponential smoothing formula, and the first predicted sales value is calculated, wherein the exponential smoothing formula is as follows:
F t+1 ＝AX t +(1-A)F t
wherein F is t+1 Predicting the pin value, X, for a t+1 cycle t For t period of actual pin value, F t And predicting a pin value for a t period, wherein A is a weight coefficient.
In the above preferred technical solution of the control method for replenishment data, the clustering the node portrait data to obtain a plurality of co-clustered nodes includes:
preprocessing the node image data to obtain preprocessed node image data;
and processing the preprocessing node portrait data through a mean shift algorithm improved by a Gaussian kernel function, and determining the co-clustered nodes.
In the above-mentioned preferred technical scheme of the control method of replenishment data, the preprocessing the node portrait data to obtain preprocessed node portrait data includes:
performing feature value conversion processing on the node portrait data to obtain feature value conversion data, wherein the feature value conversion includes: normalization processing and/or regularization processing;
And carrying out principal component analysis on the eigenvalue conversion data to obtain the preprocessing node portrait data.
In an preferable technical solution of the control method for replenishment data, the determining the second predicted sales value of the target node according to the historical sales data of the same clustered nodes includes:
multiplying the historical sales volume data of the same cluster node by an experience coefficient to obtain predicted sales volume data;
calculating the sales volume ratio of the target object of each model in the hot spot model in the historical sales volume data of the same clustering node according to the historical sales volume data of the same clustering node;
the predicted sales data is multiplied by the sales ratio of the target object for each model, respectively, to determine the second predicted sales value.
In the above-mentioned preferred technical solution of the control method of replenishment data, determining a value of a repurchase parameter of a user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value, includes:
processing the user portrait data through a lifting algorithm, determining the repurchase probability of the user in the range of the preset area of each target node, and determining the user as a possible repurchase user if the repurchase probability is higher than a preset probability threshold;
Determining the repurchase preference of the possible repurchase user according to the portrait data of the possible repurchase user;
and calculating the third predicted sales value according to the number of the possible repurchase users and the repurchase preference.
In the above preferred technical solution of the control method for replenishment data, the processing the user portrait data by a lifting algorithm to determine the repurchase probability of the user in the preset area of each target node includes:
and processing the user portrait data through an XGBoost algorithm, and determining the repurchase probability of the user in the range of the preset area of each target node.
In the above preferred technical solution of the control method for replenishment data, the determining, according to the user portrait data, a plurality of hotspot models of a target object includes:
and processing the user portrait data through an XGBoost algorithm to determine a plurality of hot spot models of the target object.
In an preferable technical solution of the control method for replenishment data, the performing on-line replenishment data control according to the final predicted sales value of the target node includes:
from the resource server, obtaining: a stock quantity of the hotspot model within the target node, and a quantity of the hotspot model being shipped to the target node;
Calculating the replenishment quantity of the target node according to a replenishment quantity calculation formula, wherein the replenishment quantity calculation formula is as follows:
B＝L 1 -K-T
wherein B is the replenishment quantity of the target node, L 1 K is the stock quantity of the hot spot model numbers in the target node, and T is the quantity of the hot spot model numbers which are being transported to the target node;
and carrying out online replenishment data control according to the replenishment quantity of the target node.
In a second aspect, the present application provides a control device for restocking data, the device comprising:
the data acquisition module is used for:
acquiring data from a resource server;
a data processing module for:
controlling the data acquisition module to acquire user portrait data of a user in a preset area range of a target node from the resource server, determining a plurality of hot spot models of a target object according to the user portrait data, then controlling the data acquisition module to acquire historical sales data of the hot spot models from the resource server, and determining a first predicted sales value of the target node according to the historical sales data of the hot spot models;
the data acquisition module is controlled to acquire node portrait data of nodes with the same type as the target node from the resource server, cluster the node portrait data to obtain a plurality of co-clustered nodes, then acquire historical sales data of the co-clustered nodes from the resource server, and determine a second predicted sales value of the target node according to the historical sales data of the co-clustered nodes;
Determining a repurchase parameter value of a user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value;
weighting the first predicted pin value, the second predicted pin value and the third predicted pin value to determine a final predicted pin value of the target node;
the replenishment data control module is used for:
and performing online replenishment data control according to the final predicted sales value of the target node.
In a third aspect, the present application provides a control apparatus for restocking data, the apparatus comprising: a memory, a processor; the memory is used for storing a computer program;
the processor is configured to execute the computer program stored in the memory, and implement the method for controlling replenishment data as described above.
In a fourth aspect, the present application provides a readable storage medium having a computer program stored thereon; the computer program is used for realizing the control method of the replenishment data.
According to the control method, the device, the equipment and the storage medium for the replenishment data, provided by the application, a plurality of hot spot models of target objects in users around target nodes are determined, and target node sales values are predicted according to the hot spot models, so that replenishment can be carried out on the free sales, and stock quantity can be reduced, and stock age and stock structure can be optimized; further, when the final predicted sales value of the target node is calculated, not only the historical sales value of the target node is considered, but also the repurchase parameter values of the same cluster nodes in the nodes with the same target node type and the users around the target node are considered, so that compared with the prior art, only the historical factors or other partial factors are considered, and the replenishment data can be more accurately scheduled.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application.
FIG. 1 is a schematic diagram of a system architecture provided by the present application;
FIG. 2 is a flowchart of a method for controlling replenishment data according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a weight calculation according to an embodiment of the present application;
FIG. 4 is a method for clustering node portrait data to obtain a plurality of co-clustered nodes according to an embodiment of the present application;
FIG. 5 is a schematic diagram of a mean shift algorithm calculation process according to an embodiment of the present application;
FIG. 6 is a flow chart of a method for determining a second predicted sales value for a target node based on historical sales data for co-clustered nodes, provided by an embodiment of the application;
FIG. 7 is a flowchart of a method for determining a user's repurchase parameter value based on user profile data and determining a third predicted sales value for a target node based on the repurchase parameter value, according to an embodiment of the present application;
FIG. 8 is a schematic diagram of a control device for replenishment data according to an embodiment of the present application;
fig. 9 is a schematic diagram of a control device for replenishment data according to an embodiment of the present application.
Specific embodiments of the present application have been shown by way of the above drawings and will be described in more detail below. The drawings and the written description are not intended to limit the scope of the inventive concepts in any way, but rather to illustrate the inventive concepts to those skilled in the art by reference to the specific embodiments.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the application. Rather, they are merely examples of apparatus and methods consistent with aspects of the application as detailed in the accompanying claims.
Traditional off-line stores are affected by a plurality of factors, and traditional channel shop goods modes are difficult to meet the requirements of fine and efficient operation of the off-line stores.
The existing dispatching method for the replenishment data of the off-line store only considers historical factors or other partial factors, and cannot accurately dispatch the replenishment data, so that on one hand, the supply of the products in the store is insufficient, on the other hand, the excessive supply of the common products is caused, the sales are lost, the whole stock quantity is large, the stock age is high, the stock structure is unreasonable, and the like, and the management risk and the management cost are greatly increased.
In order to solve the technical problems, the technical conception of the application is as follows: firstly, determining a plurality of hot spot models of target objects in users around a target node, calculating a final predicted sales value according to historical sales data of the hot spot models, historical sales data of the same cluster nodes in the nodes with the same target node type and the repurchase parameter values of the users around the target node, and controlling replenishment data according to the final predicted sales value.
The following describes the technical scheme of the present application and how the technical scheme of the present application solves the above technical problems in detail with specific embodiments. The following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Fig. 1 is a schematic diagram of a system architecture provided by the present application, and first, a system to which the method of the present application is applied will be described with reference to fig. 1:
as shown in fig. 1, the system comprises a central processing platform, a resource server and a replenishment data scheduling server, wherein the central processing platform is realized based on the server, the resource server can comprise a plurality of servers, and the method of the application is applied to the central processing platform.
In the application, a central processing platform firstly interacts with a resource server, and acquires various data of a target node from the resource server. Note that, in the present application, the node is usually an off-line sales store, and if the target node that needs to perform replenishment data scheduling is a terminal retail store, the data that may be obtained from the resource server includes B-domain data, O-domain data, purchase-sale-storage data, and the like. The central processing platform calculates the final predicted sales value of the target node according to various data acquired from the resource server, and then interacts with the replenishment data scheduling server according to the final predicted sales value of the target node to schedule and control the replenishment data of the target node.
The central processing platform performing scheduling control on the replenishment data of the target node may include:
calculating the replenishment quantity of a target object of the target node according to the final predicted sales value of the target node, wherein the target object is usually a commodity needing replenishment data scheduling in the target node;
and sending the replenishment quantity of the target object of the target node to a replenishment data scheduling server, controlling the replenishment data scheduling server to perform replenishment data scheduling control on the target node based on the inventory data according to the replenishment quantity of the target object.
The method embodiments of the present application will be described in detail with reference to the accompanying drawings, based on the system shown in fig. 1.
In embodiment 1 of the present application, a method for controlling replenishment data is provided. Fig. 2 is a flowchart of a control method for replenishment data according to an embodiment of the present application, as shown in fig. 2, where the method includes:
s201, user portrait data of a user in a preset area range of a target node are obtained from a resource server, a plurality of hot spot models of a target object are determined according to the user portrait data, and a first predicted sales value of the target node is determined according to historical sales data of the hot spot models;
firstly, according to service requirements, a preset area range of a target node is determined, for example, a preset area range within 3 km or 5 km around the target node is taken as the preset area range, and users in the preset area range are called radiation users, namely, the target node can radiate affected users. Taking a terminal retail store as an example, the user portrait data of the radiating user may include user basic information, contract information, terminal usage behavior, historical machine switching, internet surfing preferences, traffic, and the like.
Optionally, the user image data is processed through an XGBoost algorithm to determine a plurality of hotspot models of the target object. The hotspot model is usually the model of a plurality of target objects which are most popular among radiation users, the type and the number of the hotspot models can be determined according to business rules, terminal retail stores are taken as an example, and five types of the most popular terminal on the market in the last three years can be taken as the hotspot model.
S202, acquiring node portrait data of nodes with the same type as the target node from a resource server, clustering the node portrait data to obtain a plurality of same-cluster nodes, and determining a second predicted sales value of the target node according to the historical sales data of the same-cluster nodes;
in this step, the nodes with the same target node type may be determined according to the service rule, for example, if the target node is a town node, the nodes with the same target node type may be other town nodes, and if the target node is a rural node, the nodes with the same target node type may be other rural nodes. The determination of the nodes of the same type as the target node requires the determination of the type of the target node and the corresponding business rules according to the specific application of the method of the application, and in general, the nodes of the same type as the target node have commonality with the target node in at least one dimension. For ease of description, nodes of the same type of target node will be referred to as like nodes hereinafter.
The node portrayal data of similar nodes can include data of recent hot spot model sales, number of staff, age of radiating users and store users, traffic usage, network age, terminal usage duration, etc. Before node portrait data of similar nodes are processed by a clustering algorithm, the node portrait data of the similar nodes are required to be preprocessed, and preprocessed node portrait data are obtained.
And after the preprocessing node portrait data are processed through a clustering algorithm, a plurality of clustering clusters are obtained, wherein similar nodes in the clustering cluster with the maximum data density are selected as the same clustering nodes. The same cluster node represents the "majority" of the similar nodes of the target node, and the basic idea of step S202 is to predict the sales of a single target node by the sales of the "majority" of the similar nodes of the target node. Because the co-clustered nodes represent the common states in the similar nodes, the historical sales of the co-clustered nodes are used for predicting the sales of the target nodes with high accuracy.
S203, determining a repurchase parameter value of the user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value;
optionally, the repurchase parameter values determined from the user profile data of the radiating user include a potential repurchase user and a repurchase preference of the potential repurchase user. The basic idea of step S203 is to determine possible repurchase users among the radiating users, and predict sales of the target node according to the repurchase preference of the possible repurchase users.
S204, carrying out weighted calculation on the first predicted pin value, the second predicted pin value and the third predicted pin value, and determining a final predicted pin value of the target node;
Specifically, the weighted calculation formula is:
L 1 ＝aX+bY+cZ
wherein L is 1 For the final predicted pin value, X is the first predicted pin value, Y is the second predicted pin value, Z is the third predicted pin value, a is the first weight coefficient, b is the second weight coefficient, c is the third weight coefficientA number. The weight coefficient in this embodiment may be adjusted according to the actual sales feedback of the target node.
Fig. 3 is a schematic diagram of weight calculation provided by the embodiment of the present application, as shown in fig. 3, assuming that a hotspot model of a target object includes brand 1 and brand 2, then:
the first predicted sales value calculated according to the target node historical sales data comprises a brand 1 predicted sales value X1 and a brand 2 predicted sales value X2;
the second predicted sales value calculated according to the historical sales data of the same cluster node comprises a brand 1 predicted sales value Y1 and a brand 2 predicted sales value Y2;
the third predicted sales value calculated according to the radiation user repurchase parameter comprises a brand 1 predicted sales value Z1 and a brand 2 predicted sales value Z2;
bringing the above X1, X2, Y1, Y2, Z1, Z2 into the above weighted calculation formula, the final predicted sales value of the target object for each hotspot model, such as the final predicted sales value L of brand 1, is finally obtained 11 Final predicted sales value L for Brand 2 12 。
S205, performing online replenishment data control according to the final predicted sales value of the target node.
The technical effects of this embodiment are: determining a plurality of hot spot models of target objects in users around a target node, and predicting a target node sales value according to the hot spot models, so that replenishment can be carried out on smooth sales, and stock quantity can be reduced, and stock age and stock structure can be optimized; further, when the final predicted sales value of the target node is calculated, not only the historical sales value of the target node is considered, but also the repurchase parameter values of the same cluster nodes in the nodes with the same target node type and the users around the target node are considered.
In embodiment 2 of the present application, there is provided a method of determining a first predicted sales value of a target node from historical sales data of a hotspot model, the method including:
in the historical sales data of the hot spot model, a historical sales time sequence is formed in a fixed period, the historical sales time sequence is processed through an exponential smoothing formula, a first predicted sales value is calculated, and the exponential smoothing formula is as follows:
F t+1 ＝AX t +(1-A)F t
Wherein F is t+1 Predicting the pin value, X, for a t+1 cycle t For t period of actual pin value, F t And predicting a pin value for a t period, wherein A is a weight coefficient.
For example, in this embodiment, the historical sales data of the hot spot model of the last 28×2 days may be selected from the historical sales data of the target node, and the historical sales data takes 7 days as a period, so that 8 periods are total, where the first predicted sales value is a predicted sales value of the hot spot model of the target node for 7 days in the future. The present application is not limited to a specific method of dividing the period.
It should be noted that the exponential smoothing method is a time-series algorithm. The method is characterized in that firstly, the effect of a recent observed value in an observation period on a predicted value is enhanced by an exponential smoothing method, weights given to observed values in different times are unequal, so that the weight of the recent observed value is increased, the predicted value can rapidly reflect the actual change of a market, and the weights are reduced according to an geometric progression. Second, the exponential smoothing method is flexible to the weights given by the observations, and can take different a values to change the rate of change of the weights.
Therefore, in the embodiment, the historical sales data of the target node hot spot model is processed by adopting the exponential smoothing method, the weight of the recent observed value is increased, the obtained first predicted sales value is more in line with the change of the market, and the stock backlog of the target node is effectively avoided.
In embodiment 3 of the present application, a method for clustering node portrait data to obtain a plurality of co-clustered nodes is provided, and fig. 4 is a method for clustering node portrait data to obtain a plurality of co-clustered nodes, as shown in fig. 4, where the method includes:
s401, preprocessing node image data to obtain preprocessed node image data;
optionally, preprocessing the node image data to obtain preprocessed node image data, including:
performing feature value conversion processing on the node portrait data to obtain feature value conversion data, wherein the feature value conversion comprises: normalization processing and/or regularization processing;
and carrying out principal component analysis on the eigenvalue conversion data to obtain preprocessing node image data.
Since the obtained node portrait data contains various data and cannot be clustered directly, after the feature value conversion processing is performed on the node portrait data, the data with correlation among the feature value conversion data are processed by a principal component analysis method due to the correlation among the feature value conversion data, so that the types of data needing to be analyzed are reduced, and meanwhile, the loss of the information contained in the original data can be reduced as much as possible. Taking a terminal sales store as an example, the preprocessing node portrait data obtained after principal component analysis is typically the number of target node shops, the target node scale, and the like.
S402, preprocessing node image data through a mean shift algorithm improved by a Gaussian kernel function, and determining co-clustering nodes.
The Mean Shift algorithm, i.e. the Mean-Shift algorithm, is used to find a dense region of data points, and the basic principle is similar to that of k-means, and is an iterative process, i.e. the Shift Mean of the current point is calculated first, the point is moved to the Shift Mean, and the Shift Mean is taken as a new starting point, and the Shift is continued until the final condition is met. Fig. 5 is a schematic diagram of a calculation process of a mean shift algorithm according to an embodiment of the present application, as shown in fig. 5, the calculation process is as follows:
(1) Consider that a center point is initially determined in a feature space having N sample points;
(2) Calculate all points (x i ) Vector with center point;
(3) Calculating the average value of all vectors in the whole circular space to obtain an offset average value;
(4) Moving the center point to an offset mean position;
(5) And repeating the movement until a certain condition is met.
The objective of introducing a kernel function in the Mean-Shift algorithm is to make the contribution of the offset to the Mean-Shift vector different according to the difference of the distance between the sample and the offset point, so that the point which is away from the center in the calculation has a larger weight, and the shorter the distance is, the larger the weight is.
After the pretreatment node image data are processed by using the Mean-Shift algorithm, a cluster with the maximum data density can be found in the pretreatment image data set, and the shortest distance between the pretreatment node image data of different similar nodes in the cluster with the maximum data density means that the smaller the difference between the pretreatment node image data in the cluster is, the common state in the similar nodes can be represented. And finally, similar nodes in the cluster with the maximum data density are found, namely the same cluster nodes. The Mean-Shift algorithm is used, the calculated co-clustered nodes represent 'majority' in similar nodes of the target node, and the historical sales of the co-clustered nodes are used for predicting sales of single target nodes with high accuracy due to the fact that the co-clustered nodes represent general states in the similar nodes.
In embodiment 4 of the present application, a method for determining a second predicted pin value of a target node according to historical sales data of co-clustered nodes is provided, and fig. 6 is a flowchart of a method for determining a second predicted pin value of a target node according to historical sales data of co-clustered nodes, as shown in fig. 6, where the method includes:
S601, multiplying historical sales data of the same clustering nodes by experience coefficients to obtain predicted sales data;
and multiplying the historical sales data of the hot spot models of the same cluster nodes by an experience coefficient to obtain predicted sales data, wherein the experience coefficient is determined according to the business rule. The predicted sales data is the predicted sales data of the hot spot model in the future preset period of the target node, and the predicted period is equal to the period of the exponential smoothing method in embodiment 2, that is, if the sales data of the target node for 7 days is calculated according to the time sequence algorithm in embodiment 2, the predicted sales data in this embodiment is also the sales data of the target node for 7 days.
S602, calculating the sales volume ratio of the target object of each model in the hot spot model in the historical sales volume data of the same clustering node according to the historical sales volume data of the same clustering node;
in order to perform replenishment data scheduling control on different hot spot models, sales volume duty ratios of the different hot spot models need to be determined, the specific method is to obtain historical sales volume data of target objects of each hot spot model, and divide the historical sales volume data of the same clustering node, so that the sales volume duty ratios of the different hot spot models are obtained.
S603, multiplying the predicted sales data by the sales ratio of the target object of each model respectively to determine a second predicted sales value.
In the step, according to the sales volume ratio of the target object of each hot spot model, the predicted sales volume value of the target object of each hot spot model in the second predicted sales volume value is determined, so that replenishment data scheduling control can be performed for different hot spot models. For example, taking a terminal retail store as an example, the target object of the hotspot model includes a brand 1 terminal, a brand 2 terminal, and a brand 3 terminal, and the second predicted sales value includes a predicted sales value of the brand 1 terminal, a predicted sales value of the brand 2 terminal, and a predicted sales value of the brand 3 terminal.
In embodiment 5 of the present application, a method is provided for determining a user's repurchase parameter value from user representation data and determining a third predicted sales value for a target node from the repurchase parameter value. FIG. 7 is a flowchart of a method for determining a value of a repurchase parameter of a user according to user portrait data and determining a third predicted sales value of a target node according to the value of the repurchase parameter, according to an embodiment of the present application, as shown in FIG. 7, the method includes:
s701, processing user image data through a lifting algorithm, determining the re-purchase probability of a user in a preset area range of each target node, and determining the user as a possible re-purchase user if the re-purchase probability is higher than a preset probability threshold;
Users within the preset area of the target node are called radiating users. Specifically, the probability of repurchase of the radiation user within a preset time period in the future, such as the probability of repurchase of 1 month in the future to 3 months in the future, can be determined.
Preferably, the user image data is processed through a lifting algorithm, and the determination of the repurchase probability of the user in the preset area range of each target node comprises the following steps:
and processing the user image data through an XGBoost algorithm, and determining the repurchase probability of the user in the range of the preset area of each target node.
S702, determining the repurchase preference of the possible repurchase user according to the portrait data of the possible repurchase user;
specifically, in this embodiment, considering the purchase preference of the radiation user, the repurchase preference of the possible repurchase user in the hot spot model is determined according to the portrait data of the possible repurchase user. Taking the terminal retail store as an example, it may be that repurchase user a prefers brand 1 terminals in the hotspot model and repurchase user B prefers brand 2 terminals in the hotspot model.
S703, calculating a third predicted sales value according to the number of possible repurchase users and the repurchase preference.
In embodiment 6 of the present application, on-line replenishment data control is performed according to a final predicted sales value of a target node, including:
From the resource server, acquire: the inventory of hot spot models in the target node, and the number of hot spot models being shipped to the target node;
calculating the goods replenishment quantity of the target node according to a goods replenishment quantity calculation formula, wherein the goods replenishment quantity calculation formula is as follows:
B＝L 1 -K-T
wherein B is the replenishment quantity of the target node, L 1 K is the stock quantity of hot spot models in the target node, and T is the quantity of hot spot models which are being transported to the target node;
in this embodiment, the real-time performance of the target node replenishment data scheduling control may be ensured by considering the inventory of the hot spot models in the target node and the number of hot spot models being transported to the target node.
And carrying out online replenishment data control according to the replenishment quantity of the target node.
In embodiment 7 of the present application, a control device for replenishment data is provided, fig. 8 is a schematic diagram of a control device for replenishment data provided in an embodiment of the present application, and as shown in fig. 8, the device 80 includes: a data acquisition module 801, a data processing module 802, and a restocking data control module 803;
a data acquisition module 801, configured to:
acquiring data from a resource server;
a data processing module 802 for:
the control data acquisition module 801 acquires user portrait data of a user in a preset area range of a target node from a resource server, determines a plurality of hot spot models of a target object according to the user portrait data, then the control data acquisition module 801 acquires historical sales data of the hot spot models from the resource server, and determines a first predicted sales value of the target node according to the historical sales data of the hot spot models;
The control data acquisition module 801 acquires node portrait data of nodes with the same type as the target node from the resource server, clusters the node portrait data to obtain a plurality of co-clustered nodes, and then the control data acquisition module 801 acquires historical sales data of the co-clustered nodes from the resource server, and determines a second predicted sales value of the target node according to the historical sales data of the co-clustered nodes;
determining a repurchase parameter value of the user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value;
weighting the first predicted pin value, the second predicted pin value and the third predicted pin value to determine a final predicted pin value of the target node;
the replenishment data control module 803 is configured to:
and performing online replenishment data control according to the final predicted sales value of the target node.
In embodiment 8 of the present application, a control device for replenishment data is provided, fig. 9 is a schematic diagram of a control device for replenishment data provided in an embodiment of the present application, and as shown in fig. 9, the device 90 includes: the memory 901, the processor 902 and the interaction interface 903 are connected through the bus 904;
The memory 901 is for storing a computer program;
the processor 902 is configured to execute a computer program stored in the memory 901 to implement the control method of the replenishment data described above.
The specific implementation process of the processor 902 may refer to the above-mentioned method embodiment, and its implementation principle and technical effects are similar, and this embodiment will not be described herein again.
In embodiment 9 of the present application, there is provided a readable storage medium having a computer program stored thereon; the computer program is used for realizing the control method of the replenishment data.
It should be noted that, for simplicity of description, the foregoing method embodiments are all described as a series of acts, but it should be understood by those skilled in the art that the present application is not limited by the order of acts described, as some steps may be performed in other orders or concurrently in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are alternative embodiments, and that the acts and modules referred to are not necessarily required for the present application.
It should be further noted that, although the steps in the flowchart are sequentially shown as indicated by arrows, the steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least a portion of the steps in the flowcharts may include a plurality of sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, the order in which the sub-steps or stages are performed is not necessarily sequential, and may be performed in turn or alternately with at least a portion of the sub-steps or stages of other steps or other steps.
It will be appreciated that the device embodiments described above are merely illustrative and that the device of the application may be implemented in other ways. For example, the division of the units/modules in the above embodiments is merely a logic function division, and there may be another division manner in actual implementation. For example, multiple units, modules, or components may be combined, or may be integrated into another system, or some features may be omitted or not performed.
In addition, each functional unit/module in each embodiment of the present application may be integrated into one unit/module, or each unit/module may exist alone physically, or two or more units/modules may be integrated together, unless otherwise specified. The integrated units/modules described above may be implemented either in hardware or in software program modules.
The integrated units/modules, if implemented in hardware, may be digital circuits, analog circuits, etc. Physical implementations of hardware structures include, but are not limited to, transistors, memristors, and the like. The processor may be any suitable hardware processor, such as CPU, GPU, FPGA, DSP and ASIC, etc., unless otherwise specified. Unless otherwise indicated, the storage elements may be any suitable magnetic or magneto-optical storage medium, such as resistive Random Access Memory RRAM (Resistive Random Access Memory), dynamic Random Access Memory DRAM (Dynamic Random Access Memory), static Random Access Memory SRAM (Static Random-Access Memory), enhanced dynamic Random Access Memory EDRAM (Enhanced Dynamic Random Access Memory), high-Bandwidth Memory HBM (High-Bandwidth Memory), hybrid Memory cube HMC (Hybrid Memory Cube), etc.
The integrated units/modules may be stored in a computer readable memory if implemented in the form of software program modules and sold or used as a stand-alone product. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in whole or in part in the form of a software product stored in a memory, comprising several instructions for causing a computer device (which may be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method of the various embodiments of the present application. And the aforementioned memory includes: a U-disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments. The technical features of the foregoing embodiments may be arbitrarily combined, and for brevity, all of the possible combinations of the technical features of the foregoing embodiments are not described, however, all of the combinations of the technical features should be considered as being within the scope of the disclosure.
Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. This application is intended to cover any variations, uses, or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the application pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
It is to be understood that the application is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the application is limited only by the appended claims.
Claims (12)
1. A method of controlling restocking data, the method comprising:
user portrait data of a user in a preset area range of a target node are obtained from a resource server, a plurality of hot spot models of a target object are determined according to the user portrait data, and a first predicted sales value of the target node is determined according to historical sales data of the hot spot models;
Acquiring node portrait data of nodes with the same type as the target node from the resource server, clustering the node portrait data to obtain a plurality of same clustering nodes, and determining a second predicted sales value of the target node according to the history sales data of the same clustering nodes;
determining a repurchase parameter value of a user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value;
weighting the first predicted pin value, the second predicted pin value and the third predicted pin value to determine a final predicted pin value of the target node;
and performing online replenishment data control according to the final predicted sales value of the target node.
2. The method of claim 1, wherein the determining a first predicted sales value for the target node based on historical sales data for the hotspot model comprises:
in the historical sales data of the hot spot model, a historical sales time sequence is formed in a fixed period, the historical sales time sequence is processed through an exponential smoothing formula, and the first predicted sales value is calculated, wherein the exponential smoothing formula is as follows:
F t+1 ＝AX t +(1-A)F t
Wherein F is t+1 Predicting the pin value, X, for a t+1 cycle t For t period of actual pin value, F t And predicting a pin value for a t period, wherein A is a weight coefficient.
3. The method of claim 1, wherein clustering the node representation data to obtain a plurality of co-clustered nodes comprises:
preprocessing the node image data to obtain preprocessed node image data;
and processing the preprocessing node portrait data through a mean shift algorithm improved by a Gaussian kernel function, and determining the co-clustered nodes.
4. A method according to claim 3, wherein preprocessing the node representation data to obtain preprocessed node representation data comprises:
performing feature value conversion processing on the node portrait data to obtain feature value conversion data, wherein the feature value conversion includes: normalization processing and/or regularization processing;
and carrying out principal component analysis on the eigenvalue conversion data to obtain the preprocessing node portrait data.
5. The method of claim 1, wherein said determining a second predicted pin value for the target node based on historical pin data for the co-clustered nodes comprises:
Multiplying the historical sales volume data of the same cluster node by an experience coefficient to obtain predicted sales volume data;
calculating the sales volume ratio of the target object of each model in the hot spot model in the historical sales volume data of the same clustering node according to the historical sales volume data of the same clustering node;
the predicted sales data is multiplied by the sales ratio of the target object for each model, respectively, to determine the second predicted sales value.
6. The method of claim 1, wherein determining a user's repurchase parameter value from the user representation data and determining a third predicted sales value for the target node from the repurchase parameter value comprises:
processing the user portrait data through a lifting algorithm, determining the repurchase probability of the user in the range of the preset area of each target node, and determining the user as a possible repurchase user if the repurchase probability is higher than a preset probability threshold;
determining the repurchase preference of the possible repurchase user according to the portrait data of the possible repurchase user;
and calculating the third predicted sales value according to the number of the possible repurchase users and the repurchase preference.
7. The method of claim 6, wherein the processing the user representation data by the lifting algorithm to determine the probability of repurchase of the user within the predetermined area of each target node comprises:
and processing the user portrait data through an XGBoost algorithm, and determining the repurchase probability of the user in the range of the preset area of each target node.
8. The method of claim 1, wherein said determining a number of hotspot models for a target object from the user representation data comprises:
and processing the user portrait data through an XGBoost algorithm to determine a plurality of hot spot models of the target object.
9. The method according to any one of claims 1-8, wherein said performing on-line restocking data control based on the final predicted sales value of the target node comprises:
from the resource server, obtaining: a stock quantity of the hotspot model within the target node, and a quantity of the hotspot model being shipped to the target node;
calculating the replenishment quantity of the target node according to a replenishment quantity calculation formula, wherein the replenishment quantity calculation formula is as follows:
B＝L 1 -K-T
wherein B is the replenishment quantity of the target node, L 1 K is the stock quantity of the hot spot model numbers in the target node, and T is the quantity of the hot spot model numbers which are being transported to the target node;
and carrying out online replenishment data control according to the replenishment quantity of the target node.
10. A control device for replenishment data, the device comprising:
the data acquisition module is used for:
acquiring data from a resource server;
a data processing module for:
controlling the data acquisition module to acquire user portrait data of a user in a preset area range of a target node from the resource server, determining a plurality of hot spot models of a target object according to the user portrait data, then controlling the data acquisition module to acquire historical sales data of the hot spot models from the resource server, and determining a first predicted sales value of the target node according to the historical sales data of the hot spot models;
the data acquisition module is controlled to acquire node portrait data of nodes with the same type as the target node from the resource server, cluster the node portrait data to obtain a plurality of co-clustered nodes, then acquire historical sales data of the co-clustered nodes from the resource server, and determine a second predicted sales value of the target node according to the historical sales data of the co-clustered nodes;
Determining a repurchase parameter value of a user according to the user portrait data, and determining a third predicted sales value of the target node according to the repurchase parameter value;
weighting the first predicted pin value, the second predicted pin value and the third predicted pin value to determine a final predicted pin value of the target node;
the replenishment data control module is used for:
and performing online replenishment data control according to the final predicted sales value of the target node.
11. A control apparatus for restocking data, the apparatus comprising: a memory, a processor;
the memory is used for storing a computer program;
the processor is configured to execute a computer program stored in the memory to implement the control method of restocking data according to any one of claims 1 to 9.
12. A readable storage medium having a computer program stored thereon; the computer program is for implementing a control method of restocking data according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311245803.9A CN117236850A (en) | 2023-09-25 | 2023-09-25 | Control method, device, equipment and storage medium for replenishment data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311245803.9A CN117236850A (en) | 2023-09-25 | 2023-09-25 | Control method, device, equipment and storage medium for replenishment data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117236850A true CN117236850A (en) | 2023-12-15 |
Family
ID=89094549
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311245803.9A Pending CN117236850A (en) | 2023-09-25 | 2023-09-25 | Control method, device, equipment and storage medium for replenishment data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117236850A (en) |
-
2023
- 2023-09-25 CN CN202311245803.9A patent/CN117236850A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10936947B1 (en) | Recurrent neural network-based artificial intelligence system for time series predictions | |
| CN110503531B (en) | Dynamic social scene recommendation method based on time sequence perception | |
| CN108351999B (en) | System and method for providing multi-channel inventory distribution avenues for retailers | |
| US10748072B1 (en) | Intermittent demand forecasting for large inventories | |
| JP6443858B2 (en) | Calculation device, calculation method, learning device, learning method, and program | |
| Sun et al. | Dynamic matrix factorization: A state space approach | |
| WO2013071414A1 (en) | System, method and computer program for forecasting energy price | |
| CN111402013A (en) | Commodity collocation recommendation method, system, device and storage medium | |
| WO2023036184A1 (en) | Methods and systems for quantifying client contribution in federated learning | |
| CN113743971B (en) | Data processing method and device | |
| CN108665156B (en) | Supply chain selection evaluation method based on Markov chain under block chain | |
| CN112579876A (en) | Information pushing method, device and system based on user interest and storage medium | |
| CN111080206A (en) | Method, device and equipment for generating replenishment list and storage medium | |
| WO2021146802A1 (en) | Method and system for optimizing an objective having discrete constraints | |
| US20160171365A1 (en) | Consumer preferences forecasting and trends finding | |
| WO2018088277A1 (en) | Prediction model generation system, method, and program | |
| US20180268352A1 (en) | Method and system for retail stock allocation | |
| CN103782290A (en) | Generation of recommendation values | |
| CN114240052A (en) | Combined sales strategy optimization method and system based on genetic algorithm | |
| CN117710049A (en) | Integral commodity recommendation method and system based on user portrait | |
| Huang et al. | Taylor approximation of inventory policies for one-warehouse, multi-retailer systems with demand feature information | |
| CN117236850A (en) | Control method, device, equipment and storage medium for replenishment data | |
| Alamdar et al. | A deep Q-learning approach to optimize ordering and dynamic pricing decisions in the presence of strategic customers | |
| CN115049458A (en) | Commodity pushing method and device based on user crowd modeling, medium and equipment | |
| CN115829624A (en) | Price prediction method and device based on store periodic copybook and related medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |