WO2023020255A1 - 数据处理方法、装置、设备及存储介质 - Google Patents

数据处理方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2023020255A1
WO2023020255A1 PCT/CN2022/108897 CN2022108897W WO2023020255A1 WO 2023020255 A1 WO2023020255 A1 WO 2023020255A1 CN 2022108897 W CN2022108897 W CN 2022108897W WO 2023020255 A1 WO2023020255 A1 WO 2023020255A1
Authority
WO
WIPO (PCT)
Prior art keywords
time period
time
stocking
shipments
goods
Prior art date
Application number
PCT/CN2022/108897
Other languages
English (en)
French (fr)
Inventor
詹昌飞
Original Assignee
北京沃东天骏信息技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 北京沃东天骏信息技术有限公司 filed Critical 北京沃东天骏信息技术有限公司
Publication of WO2023020255A1 publication Critical patent/WO2023020255A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0633—Lists, e.g. purchase orders, compilation or processing
    • G06Q30/0635—Processing of requisition or of purchase orders

Definitions

  • the present application relates to the field of data processing, involving but not limited to a data processing method, device, equipment and storage medium.
  • the e-commerce platform estimates the stocking quantity in scenarios such as large-scale promotional activities on June 18 and November 11, based on manual experience. In this way, the estimated stocking quantity is accurate. low-level problem.
  • the present application provides a data processing method, device, device, and storage medium to solve at least one problem existing in the related art.
  • the embodiment of the present application provides a data processing method, the method comprising:
  • the target proportion sequence determines the target proportion sequence from at least one proportion sequence;
  • the first time period includes at least two time sub-segments;
  • the second time period A time sub-segment is any historical time sub-segment in the at least two time sub-segments;
  • the target proportion sequence is the ratio of the stocking quantity of different time sub-segments in the at least one time sub-segment;
  • the target ratio is the target ratio in the target ratio sequence
  • an embodiment of the present application provides a data processing device, the device comprising:
  • the determination unit is configured to determine the target proportion sequence from at least one proportion sequence according to the total amount of goods stocked in the first time period and the goods in and out in the first time period;
  • the first time period includes at least two time periods segment;
  • the first time sub-section is any historical time sub-section in the at least two time sub-sections;
  • the target proportion sequence is the stocking amount of different time sub-sections in the at least one time sub-section Proportion;
  • the prediction unit is configured to predict the stocking amount of the second time sub-segment in the first time period according to the total stocking amount and the target proportion in the target proportion sequence, and the target proportion is the target proportion The proportion of the stocking quantity corresponding to the second time sub-segment in the proportion sequence.
  • an embodiment of the present application provides an electronic device, a memory, a processor, and a computer program stored on the memory and operable on the processor, and the processor implements the above-mentioned data when executing the computer program.
  • Approach
  • the embodiment of the present application provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the foregoing data processing method is implemented.
  • FIG. 1 is an optional structural schematic diagram of a data processing system provided by an embodiment of the present application
  • FIG. 2 is an optional schematic flowchart of a data processing method provided in an embodiment of the present application
  • FIG. 3 is an optional schematic flowchart of a data processing method provided in an embodiment of the present application.
  • FIG. 4 is a schematic diagram of an optional interface provided by the embodiment of the present application.
  • FIG. 5 is an optional schematic flowchart of a data processing method provided in an embodiment of the present application.
  • FIG. 6 is an optional structural schematic diagram of a data processing device provided in an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of an optional electronic device provided by an embodiment of the present application.
  • SKU Stock Keeping Unit
  • a time period is a continuous time period, for example, a time period with a set duration such as one year, one month, etc., wherein a time period can be divided into multiple time sub-segments, for example, if the time period is one year, then the time sub-segment is a month, and for another example, the time period is one month, and the time sub-section is one day.
  • the stocking amount may be the quantity of goods that need to be stocked in the first week of May.
  • Inventory quantity which is used to represent the quantity of goods remaining in the warehouse.
  • the data processing method of the embodiment of the present application can be applied to the data processing system 100 shown in FIG. 1 .
  • the data processing system 100 includes: a server 10 and a client 20 .
  • the server 10 and the client 20 communicate through the network 30 .
  • the data processing method provided in the embodiment of the present application may be applied to a data processing device, and the data processing device may be a server 10 or a client 20 .
  • the data processing device determines a target proportion sequence from at least one proportion sequence according to the total amount of goods stocked in the first time period and the amount of goods in and out of the first time sub-segment, and then determines the target proportion sequence according to the total quantity of goods stocked and the target proportion sequence
  • the target ratio in the first time period is predicted to be the stocking quantity of the second time subsection in the first time period, wherein the first time period includes at least two time subsections; the first time subsection is the at least two time subsections Any historical time sub-segment in the two time sub-segments; the target proportion sequence is the proportion of the stocking quantity of different time sub-segments in the at least one time sub-segment; the target proportion is the target proportion The proportion of the stocking quantity corresponding to the second time sub-segment in the comparison sequence.
  • the server 10 sends the predicted stocking quantity of the second time sub-segment in the first time period to the client 20 through the network 30, and the client 20 receives the second time sub-segment After the stocking quantity of the sub-segment is displayed, the stocking quantity of the second time sub-segment is displayed to the user.
  • the client 20 directly displays the stocking quantity of the second time sub-segment to the user.
  • FIG. 2 is a schematic diagram of the implementation flow of a data processing method provided in the embodiment of the present application. The method is applied to a data processing device. As shown in FIG. 2, the method may include the following steps:
  • the data processing device determines a target proportion sequence from at least one proportion sequence according to the total amount of goods stocked in the first time period and the inbound and outbound quantity of goods in the first time subsegment.
  • the first time period is a continuous time period, and the continuous time period may be one day, one month, or one year, which is not limited in this embodiment of the present application.
  • the first time period may include at least two time subsections, and the first time subsection is any historical time subsection in the at least two time subsections, wherein the historical time subsection is a time subsection that has already occurred.
  • the first period of time is one month
  • at least two time subsections included in this month may include: the first week, the second week, the third week and the fourth week; The time subsection of , then the first week is the first time subsection.
  • the first time period is one year, and the at least two time subsections included in this year may include: the first month, the second month, the third month, ..., the eleventh month and the twelfth month; if the first month is a time sub-segment that has occurred, then the first month is the first time sub-segment.
  • the first time period is one year
  • the at least two time sub-sections included in the year may include: the first quarter, the second quarter, the third quarter and the fourth quarter; if the first quarter is already occurred the past time sub-segment, the first quarter is the first time sub-segment.
  • the data processing device may determine the first time subsection included in the first time period according to the first time period, and after determining the first time subsection , to determine the amount of goods in and out of the first time sub-segment.
  • the quantity of goods in and out includes: shipments and inventory, where the shipment is used to represent the quantity of goods sold, and the quantity of inventory is used to represent the quantity of goods remaining in the warehouse.
  • the method for determining the shipment volume may include: the data processing device acquires at least one order data; the data processing device determines the shipment volume by parsing the at least one order data; wherein, each of the at least one order data Order data is used to represent the quantity of goods purchased by users.
  • obtaining at least one order data includes: obtaining at least one order data from a distributed file system (Hadoop Distributed File System, HDFS).
  • a distributed file system Hadoop Distributed File System, HDFS.
  • Determining the shipment volume by analyzing the at least one order data includes: determining the shipment volume by summarizing each order data in the at least one order data.
  • the data processing device obtains 3 order data from HDFS.
  • the first order data it is recorded that the quantity of goods A purchased by user A is 10.
  • the second order data it is recorded that user B
  • the third order data it is recorded that the quantity of goods A purchased by user C is 10
  • the data processing device records the quantity of goods A purchased by the user recorded in the three order data Summarize and determine that the quantity of goods A sold is 40.
  • the goods may be SKUs, or all commodities under a commodity category, which is not limited in this embodiment of the present application.
  • the goods when the goods are all commodities under a commodity category, and the commodity category is sweaters, the goods may include sweaters of all sizes and colors of sweaters under the sweaters category.
  • the at least one proportion sequence is at least one proportion sequence corresponding to the first time period, and for each of the proportion sequences in the at least one proportion sequence, the proportion The ratio sequence includes at least one ratio, and the ratio is the ratio of the stocking amount of the corresponding time sub-segment to the total stocking amount of the first time segment.
  • the first time period is May
  • the at least two time subsections included in May may include: the first week, the second week, the third week and the fourth week;
  • the comparison sequence may include: [20%, 30%, 40%, 10%], for the proportion sequence [20%, 30%, 40%, 10%], the proportion sequence includes the proportion of 20%, the proportion of The proportion is 30%, 40% and 10%, among which, 20% is the ratio of the stock in the first week to the total stock in May, and 30% is the stock in the second week.
  • the ratio of the total stocking in the month, 40% is the ratio of the stocking in the third week to the total stocking in May
  • 40% is the ratio of the stocking in the fourth week to the total stocking in May.
  • the data processing device may receive at least one proportion sequence; and determine the at least one proportion sequence as the at least one proportion sequence.
  • receiving at least one proportion sequence input includes: the data processing device receives at least one proportion sequence input by a user.
  • At least one proportion sequence received includes: [20%, 30%, 40%, 10%] and [30%, 20%, 10%, 40%], then the determined at least one proportion The sequences were [20%, 30%, 40%, 10%] and [30%, 20%, 10%, 40%].
  • the data processing device receives the historical proportion sequence; on the basis of the historical proportion sequence, for the proportion corresponding to each time sub-segment, the adjustment ratio can be increased or decreased, and the at least one proportion can be determined.
  • Comparison sequence wherein, the historical proportion sequence is the proportion sequence of the first historical time period corresponding to the first time period; the adjustment step is used to represent the adjustable ratio within the adjustable range; the adjustable range It is used to characterize a maximum adjustment range when adjusting each proportion;
  • the first time period is May
  • the first historical time period corresponding to the first time period May is May of last year
  • the proportion sequence of May last year that is, the historical proportion sequence is [30%, 20% %, 10%, 40%]
  • the adjustable range is ⁇ 5%
  • the adjustment ratio is 1%.
  • the target proportion sequence is the proportion of stocking quantities in different time sub-segments in at least one time sub-segment.
  • At least one proportion sequence includes: proportion sequence 1 [20%, 30%, 40%, 10%] and proportion sequence 2 [30%, 20%, 10%, 40%], then it can From proportion sequence 1 and proportion sequence 2, determine the target proportion sequence as proportion sequence 1.
  • the data processing device predicts the stocking amount in the second time sub-segment in the first time period according to the total stocking amount and the target proportion in the target proportion sequence.
  • the target proportion is the proportion of the stocking quantity corresponding to the second time sub-segment in the target proportion sequence.
  • the target proportion sequence is [20%, 30%, 40%, 10%], wherein, 30% is the proportion of the stocking amount corresponding to the second time sub-segment, then Based on the total stocking amount of 400 and the target proportion of 30%, it can be predicted that the stocking amount in the second time sub-period is 120.
  • the embodiment of the present application provides a data processing method, which determines the target proportion sequence from at least one proportion sequence according to the total amount of stock in the first time period and the goods in and out in the first time sub-segment; the first time period Including at least two time sub-sections; the first time sub-section is any historical time sub-section in the at least two time sub-sections; the target proportion sequence is different time sub-sections in the at least one time sub-section
  • the proportion of the stocking amount of the segment; according to the target proportion in the total stocking amount and the target proportion sequence, predict the stocking amount of the second time sub-segment in the first time period, and the target proportion is The proportion of the stocking quantity corresponding to the second time sub-segment in the target proportion sequence.
  • the target proportion sequence can be determined according to the stocking quantity of the first time period and the goods in and out of the first time sub-segment that have occurred , and then predict the stocking amount of the second time sub-segment that has not occurred according to the total stocking amount and the target proportion in the target proportion sequence. In this way, the accuracy of the estimated stocking quantity in the second time sub-period can be improved.
  • the above S201 may include:
  • S301 Determine at least one reference stocking amount according to the total stocking amount in the first time period and the reference proportion in the at least one proportion sequence.
  • the total amount of stock in the first time period is 400, and at least one proportion sequence includes: proportion sequence 1: [20%, 30%, 40%, 10%] and proportion sequence 2: [25%, 35% , 25%, 25%], for the first time sub-segment in the first time period, the reference proportion of the first time sub-segment can be 20%, or 25%.
  • the total stocking amount of 400 and the reference ratio of the first time sub-segment account for 20%.
  • the reference stocking amount of the first time sub-segment is determined to be 80.
  • the reference proportion is 25%, and the reference stocking quantity for a first time sub-segment is determined to be 100, and the reference stocking quantities 80 and 100 for the two first time sub-segments constitute at least one reference stocking quantity.
  • S302. Determine at least one spot rate for the first time sub-segment according to the at least one reference stocking amount and the goods in and out of the first time sub-segment.
  • At least one spot rate for the first time sub-segment can be determined according to at least one reference stocking quantity of the first time sub-segment and the quantity of goods in and out of the first time sub-segment, wherein, The spot rate is the ratio of the quantity shipped to the quantity to be sold.
  • the target spot rate that satisfies the condition may include: taking the highest spot rate among at least one spot rate as the target spot rate.
  • At least one stock rate includes: 30%, 50% and 60%, wherein, 30% corresponds to a reference stock quantity of 50, 50% corresponds to a reference stock quantity of 80, and 60% corresponds to a reference stock quantity of 100. If the target stock rate is 60%, then use the reference stocking quantity 100 corresponding to the target spot rate of 60% as the target reference stocking quantity.
  • At least one proportion sequence includes: proportion sequence 1: [20%, 30%, 40%, 10%] and proportion sequence 2: [25%, 35%, 25%, 25%], If the proportion series of the target reference stocking quantity is obtained as proportion series 1, then proportion series 1 is the target proportion series.
  • the target proportion can be determined from at least one proportion sequence according to the total amount of goods stocked in the first time period, the inbound and outbound quantity of goods in the first time period, and the supplier delivery time (vendor lead time, VLT). Compare sequence.
  • the amount of goods in and out includes shipments and inventory
  • the above S302 may include: determining at least one amount to be sold according to the at least one reference stocking amount and the inventory; based on the shipment and the inventory The at least one quantity for sale determines the at least one spot rate.
  • the inventory of the first time sub-segment is 100, and at least one reference stocking quantity includes: 80 and 100, then it can be based on a With reference to stocking quantity 80 and inventory quantity 100, determine that a quantity for sale is 180, and then based on shipment quantity 50 and quantity for sale 180, determine a spot rate of 27%; also according to a reference stock quantity 100 and stock quantity 100, It is determined that the quantity for sale is 200, and then based on the shipment quantity of 50 and the quantity for sale of 200, a spot rate of 25% is determined.
  • determining at least one spot rate of the first time sub-period may be determined first.
  • the spot rate of each time sub-segment in the at least two time sub-segments included in the first time period is calculated, and then the spot rate of each time sub-segment is summed to obtain the spot rate of the first time period.
  • the data processing device determining at least one spot rate in the first time period includes: inputting each proportion sequence in at least one proportion sequence into a mixed integer programming (Mixed-Integer Programming, MIP) model, and obtaining each A spot rate corresponding to a proportion sequence.
  • MIP Mixed integer programming
  • the model objective of the MIP model is to minimize turnover Among them, R is the proportion sequence, i is at least one proportion sequence, itoi represents the turnover rate corresponding to the i-th proportion sequence, and min represents the minimum turnover rate.
  • the constraint condition of the MIP model is the maximum spot rate, wherein the maximum spot rate can be realized by the following formula (1):
  • st is the abbreviation of subject to, which is used to represent the constraint conditions
  • cri represents the spot rate corresponding to the i-th proportion sequence
  • crmax represents the maximum spot rate
  • stcr i ⁇ cr max is used to represent the i-th proportion sequence corresponding to The spot rate of should be greater than or equal to the maximum spot rate.
  • a maximum spot rate may be determined from the at least one spot rate by using a mathematical optimization technique (CPLEX).
  • CPLEX mathematical optimization technique
  • the method may further include: according to the inventory quantity at the first time in the first time period, the inventory quantity at the second time in the first time period, and the output in the first time period
  • the quantity of goods is to determine the total amount of stock; the first time is the start time of the first time period, and the second time is the end time of the first time period.
  • determining the total stocking amount may include : The inventory of the second time minus the inventory of the first time, plus the shipment of the first time period, determines the total amount of stock.
  • the method may further include: determining the inventory amount at the second time according to the shipment amount in the second time period and the stock-to-sales ratio coefficient.
  • the stock-to-sales ratio coefficient is used to characterize the ratio between inventory and shipments;
  • the second time period is a time period adjacent to the first time period and after the first time period.
  • the determining the inventory at the second time according to the shipments in the second time period and the stock-to-sales ratio coefficient may include: multiplying the shipments in the second time period by the stock-to-sales ratio coefficient to determine the inventory at the second time quantity.
  • the inventory at the second time can be determined by multiplying the shipment in the second time period of 100 by the stock-to-sales ratio coefficient of 0.4 The amount is 40.
  • the method may further include: according to the shipments in the first historical time period corresponding to the first time period and the shipments in the second historical time period corresponding to the third time period, determining the The first growth rate of the shipments in the first historical time period relative to the shipments in the second historical time period; according to the first growth rate and the shipments in the third time period, predict the Shipments for the first time period described above.
  • the first historical time period corresponding to the first time period is the time period of the same period in previous years.
  • the first time period is May of this year
  • the 5 The first historical time period corresponding to the month may be May of last year or May of the previous year, which is not limited in this embodiment of the present application.
  • the second historical time period corresponding to the third time period is the time period of the same period in previous years.
  • the third time period is April of this year
  • the second historical time period corresponding to April may be last April, It may also be April of the previous year, which is not limited in this embodiment of the present application.
  • the first time period is May this year
  • the first historical time period is May last year
  • the shipment volume in May last year is 200
  • the third time period is April this year
  • the 4 The monthly shipment volume is 100
  • the second historical time period is April last year
  • the shipment volume in April last year is 100
  • it can be based on the shipment volume of 200 in May last year and the shipment volume of 100 in April last year.
  • the quantity is 200.
  • the method may further include: the data processing device inputs the promotional data of the first time period into a fitting regression model to predict the reference information of the goods in the first time period;
  • the reference information of the first time period is input into the sales forecast model to predict the shipment volume of the first time period.
  • the promotional data is used to represent the promotional information and promotion plan of the goods, wherein the promotional information is used to represent a specific preferential form, and the promotional plan is used to represent the total amount of goods expected to be sold during the entire promotion process; the reference The information is used to represent the exposure and price of the goods; the sales forecast model is used to predict the shipments.
  • Promotional data can be entered through a given template.
  • the given template may include: goods, promotional information and the duration of the promotional information.
  • the promotional information may include: full discount discount information, full gift discount information, and flash kill information.
  • the offer information for full discounts is: 50 off when you spend 100.
  • the discount information for free gifts is: 100 free for orders over 100.
  • the seckill information is: time-limited seckill.
  • the sales promotion plan may be: the total amount of goods expected to be sold during the two-month promotion process is 1000 pieces.
  • the fitting regression model may include: a logistic regression (Logistic Regression, LR) model, a maximum gradient boosting (Extreme Gradient Boosting, XGBoost) model, a lightweight gradient boosting machine (Light Gradient Boosting Machine, LightGBM) model And Convolutional Neural Networks (CNN), etc.
  • logistic regression Logistic Regression, LR
  • maximum gradient boosting Extreme Gradient Boosting, XGBoost
  • XGBoost lightweight gradient boosting machine
  • Light Gradient Boosting Machine LightGBM
  • CNN Convolutional Neural Networks
  • Determining the sales forecast model may include: the data processing device acquires the time-series features and attribute features of the goods; inputting the time-series features, the attribute features, the exposure of the goods, and the price of the goods into the reference sales forecast model, for The reference sales forecast model is trained to obtain a sales forecast model.
  • the sales forecast model may include: linear regression model, time series model and deep learning model, etc.
  • the time series model may include: prophecy (Prophet) model and Holt-Winter (Holt Winter) model, etc.
  • deep learning Models can include: Recurrent Neural Network (RNN) model, Multi-Quantile Recurrent Neural Network (MQRNN) model, Long Short-Term Memory (LSTM) network model, etc.
  • RNN Recurrent Neural Network
  • MQRNN Multi-Quantile Recurrent Neural Network
  • LSTM Long Short-Term Memory
  • the time-series features of the goods may include: historical sales features of the goods, where the historical sales features are used to represent the average sales volume of the goods within a historical time period.
  • the historical sales feature may include: the average sales volume of the historical 3 days, the historical 5 days, the historical 7 days, the historical 14 days or the historical 30 days.
  • the attribute characteristics of the goods may include information such as the brand, body shape, and volume of the goods.
  • Obtaining the attribute characteristics of the goods may include: the data processing device obtains the attribute information of the goods from the commodity information system, stores the acquired attribute information of the goods in a relational database (MYSQL database) or HDFS, and performs Processing, so as to obtain the attribute characteristics of the goods, wherein the attribute characteristics of the goods can be expressed by attribute values.
  • a relational database MYSQL database
  • HDFS high definition framework
  • the method may further include: fitting the promotion data of the first historical time period corresponding to the first time period with the reference information of the first historical time period to obtain the fitting regression model.
  • the first time period is May
  • the first historical time period corresponding to May is May of last year
  • the sales promotion data in May of last year includes full discount discounts
  • the promotion data of full discount discounts in May last year The data was fitted with the exposure volume of the goods in May last year and the actual received price to obtain a fitted regression model.
  • the method may further include: according to the shipments in the second historical time period corresponding to the third time period and the shipments in the third historical time period corresponding to the second time period, determining the The second growth rate of the shipment volume in the third historical time period relative to the shipment volume in the second historical time period; according to the second growth rate and the shipment volume in the third time period, predict the Shipments for the second time period.
  • the third historical time period corresponding to the second time period is the time period of the same period in previous years.
  • the third historical time period corresponding to June may be June of last year.
  • the month may also be June of the previous year, which is not limited in this embodiment of the present application.
  • the third time period is April this year, the second historical time period is April last year, and the shipment volume in April last year is 100, the second time period is June this year, and the third historical time period is In June last year, the shipment volume in June last year was 200, and the shipment volume in June last year was 200 and the shipment volume in April last year was 100, to determine the shipment volume in June last year relative to the shipment volume in April last year.
  • the second growth rate of shipment volume is 100%, and based on the second growth rate of 100% and the shipment volume of 100 in April this year, it is predicted that the shipment volume in June this year will be 200.
  • the second growth rate R1 of the SKU after determining the second growth rate R1 of the SKU, determine the second growth rate R2 of all goods under the brand represented by the SKU, and determine the second growth rate R1 of the SKU and the The size relationship of the second growth rate R2 of all goods under the brand, if the second growth rate R1 of the SKU is greater than three times the second growth rate R2 of the goods under the brand, then the SKU’s The second growth rate R1 is replaced by the second growth rate R2 of all goods under the brand.
  • the method may further include: inputting the promotional data of the second time period into a fitting regression model to predict the reference information of the goods in the second time period;
  • the reference information of the time period is input into the sales forecast model to predict the shipment volume in the second time period.
  • the promotion data is used to represent the promotion information and promotion plan of the goods
  • the reference information is used to represent the exposure and price of the goods
  • the sales forecast model is used to predict the shipment volume.
  • the three kinds of data can be stored in the MYSQL database and HDFS, and the data stored in the MYSQL database and HDFS can be stored on a platform , such as in the big data mart platform.
  • the original data is used to represent the data before the data is processed, for example, the attribute information of the goods, the commodity number and the commodity name, etc.
  • the intermediate data is used to represent all the data used before the result data is obtained, for example, the first The data of the goods in and out of the time sub-segment
  • the result data is used to characterize the data obtained after processing the intermediate data, for example, the predicted stocking quantity of the second time sub-segment in the first time segment.
  • the original data, intermediate data and result data can be stored in the form of data tables in the MYSQL database and HDFS.
  • the data table is a way to store data in a structured manner.
  • the data After the data is stored in the big data mart platform, the data can be pushed to the MYSQL database of the replenishment system by using the Plumber through train, and the original data, intermediate data and result data will be displayed on the display interface. At the same time, users can Select the data to be displayed.
  • a schematic diagram of the original data, intermediate data and result data displayed on the display interface may be as shown in FIG. 4 .
  • Computer visualization shows the calculated data, calculation process and calculation results to the user.
  • Purchasing in transit means that the supplier is purchasing goods, but has not yet sent them to the warehouse.
  • the replenishment plan is mainly divided into two parts.
  • One is the sales estimate.
  • the sales of future big promotions are manually estimated based on historical experience. Due to e-commerce There are tens of millions of SKUs. Manuals can only stratify SKUs based on the original sales, and estimate SKUs at different levels, which is difficult to refine.
  • the total amount of replenishment is estimated based on the estimated sales; the second is the storage rhythm arrangement , due to the sharp increase in the overall warehousing volume in the early stage of the big promotion, it is a huge challenge for the warehousing arrangement.
  • Premature warehousing will increase inventory turnover and increase e-commerce storage and cash flow costs. Too late warehousing may lead to Insufficient warehousing capacity, unable to warehousing, resulting in sales loss.
  • the labor will refer to the warehousing rhythm based on the same historical period, and communicate with the logistics to determine the approximate warehousing rhythm.
  • the warehousing logic and rhythm adopted by different buyers for replenishment may be different, without coordination and planning.
  • This application proposes a data processing method, which is mainly implemented based on machine learning and operational research optimization methods.
  • the main implementation steps are data acquisition, sales forecast, storage rhythm optimization, big promotion and replenishment plan output, and data storage and output.
  • data acquisition includes data analysis and product information acquisition, etc.
  • sales forecasting uses statistical methods and machine learning models to output future sales estimates of products
  • storage rhythm optimization uses simulation and operational research optimization to solve the optimal storage rhythm ratio (that is, the target ratio sequence described in the above-mentioned embodiment), the output of the big promotion replenishment plan, using the sales forecast and the optimal warehousing rhythm to calculate the recommended amount of replenishment for multiple order days during the big promotion period.
  • This application combines the sales forecasting module and the warehousing rhythm optimization module to carry out automatic decision-making for replenishment of products during major promotions.
  • First obtain all kinds of information known about the product, including basic attribute information (commodity brand, category, etc.), time series information (sales data that has occurred in history), and business input (sales plan, promotion information, etc.). Then, the input information is input into the sales estimation device and the warehouse-in rhythm optimization device, and the future sales forecast and the optimal warehouse-in rhythm are respectively output.
  • the sales forecast and the optimal warehousing rhythm the recommended amount of SKU stocking for major promotions in the future is given.
  • the specific process is as follows:
  • the necessary hardware conditions for the model building process are computers, servers, and at least one database.
  • the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to perform "data acquisition", Any one of the five steps of "regression of sales forecast time series model", “optimization of storage rhythm”, “replenishment suggestion output for big promotion” and “data storage and output”.
  • Step 1 data acquisition.
  • the product attribute data can be obtained from the product information system, mainly crawled through the product master station information, stored in MYSQL and HDFS, and processed and stored by selecting the required product information through database operations.
  • the user order data is read through the distributed file system HDFS, and processed into the daily sales data of the product by analyzing the user order data. For example, a total of 10 users placed orders for product A on January 1, 2020.
  • the total sales volume of a user, the total sales volume of commodity A on that day is x pieces, and the analysis data is stored in the database, which is convenient for subsequent model calls and processing.
  • the promotional data entered by the business is input through a given template and processed by the system.
  • the data required for this application needs to be updated regularly, so establish a data flow task from the data source to the database storage, obtain the data regularly, and store it in the required database.
  • An example of the relevant data structure obtained is as follows.
  • Step two sales forecast.
  • Sales forecasting uses business input to estimate future sales. Since the big promotion period is different from daily sales, the fluctuation of overall demand will be much larger, and the main reason for demand fluctuations comes from changes in promotional prices and traffic. Therefore, overall estimates are divided into the following two types:
  • the first type is that the business does not enter promotional plans and sales plans that have a greater impact on sales.
  • the smoothing process is replaced by the overall chain ratio coefficient of the brand dimension; Sales volume and chain factor, get sales estimates for May and June.
  • the second type provides promotional information and sales plans for the business.
  • the specific steps include: according to the historical promotion data and sales data of the SKU dimension, and Fit the actual traffic and the final price to obtain a factor fitting regression model.
  • the model is not limited to LR, Xgboost, LightGBM, CNN, etc., and then obtain the estimated future traffic and final price according to the promotional information and plans input by the business; Time series features and attribute features of SKU, as well as historical traffic and arrival price features for sales forecasting model training, model linear regression, time series forecasting algorithm (Prophet or Holt winter) and deep learning model (RNN, MQRNN, LSTM), etc.; combined The future estimated traffic, the price and the trained sales forecast model can be used to obtain the output of future time series forecast results.
  • Step 3 Determine the storage ratio.
  • the warehousing ratio is the proportion sequence described in the above embodiment.
  • the warehousing ratio can be input by business, or can be calculated by using the historical purchase data in the same period to calculate the warehousing ratio from the first week W1 to the nth week Wn.
  • the adjustable space Range of the storage ratio is given; according to the storage ratio and the adjustable space Range, the adjustment step is set to combine all possible storage ratios.
  • optimization solver CPLEX solver is used to solve the optimal warehousing rhythm, and the target warehousing ratio with optimal turnover is obtained under the condition of optimal spot rate.
  • optimization solver CPLEX solver is used to solve the optimal warehousing rhythm, and the ratio of the warehousing rhythm with the optimal turnover is obtained under the condition of the optimal spot rate.
  • the target stock-in ratio is a stock-in ratio corresponding to a reference spot rate
  • the reference spot rate is the highest spot rate among the at least one spot rate.
  • Step 4 Forecast the recommended replenishment quantity for the big promotion.
  • the overall SKU stocking plan for the big promotion can be calculated.
  • the recommended amount of replenishment the estimated total stocking amount * the ratio of the warehousing rhythm of this node.
  • Step five data storage and output.
  • the storage of results is mainly divided into two steps.
  • the first step is to store the original data (attribute characteristics and promotion characteristics), intermediate data (intermediate variable data) and result data through the distributed file management system HDFS, and store the data table in Big data marketplace platform.
  • the second step is to use Plumber to push the data to the MYSQL database of the replenishment system, and use the front end of the replenishment system to output the original data, intermediate data and result data.
  • users can choose the system to display data according to their own needs.
  • the data processing method provided in the embodiment of the present application may include:
  • the data processing device receives target data.
  • the target data may include: sales plan, promotional information or sales data of goods.
  • the data processing device judges whether the target data includes sales plan and promotion information.
  • target data includes the sales plan and promotion information
  • perform S502a use a machine learning model to estimate sales.
  • perform S502b use historical sales data to estimate sales.
  • the data processing device outputs a monthly sales estimate.
  • the data processing device calculates the ratio of historical storage in the same period according to the sales data.
  • the data processing device forms at least one storage ratio according to the adjustment ratio.
  • the data processing device uses simulation to simulate inventory sales to obtain stock quantities corresponding to different stock-in ratios.
  • the data processing device uses the operational research model to find an optimal storage ratio.
  • the data processing device determines a big promotion replenishment plan according to the monthly sales estimate and the optimal storage ratio.
  • FIG. 6 is a data processing device provided by an embodiment of the present application. As shown in FIG. 6, the data processing device 600 includes:
  • the determination unit 601 is configured to determine a target proportion sequence from at least one proportion sequence according to the total amount of stock in the first time period and the inbound and outbound quantity of goods in the first time sub-segment; the first time period includes at least two time periods Subsection; the first time subsection is any historical time subsection in the at least two time subsections; the target proportion sequence is the stocking amount of different time subsections in the at least one time subsection proportion of
  • the predicting unit 602 is configured to predict the stocking amount in the second time sub-segment in the first time period according to the total stocking amount and the target proportion in the target proportion sequence; the target proportion is the The proportion of the stocking quantity corresponding to the second time sub-segment in the target proportion sequence.
  • the determining unit is further configured to:
  • the proportion sequence that obtains the target reference stocking quantity is determined as the target proportion sequence.
  • the quantity of goods in and out includes shipments and inventory
  • the determining unit is further configured to:
  • the at least one spot rate is determined based on the shipped quantity and the at least one pending sale quantity.
  • the determining unit is further configured to:
  • the first time is a start time of the first time period
  • the second time is an end time of the first time period
  • the determining unit is further configured to:
  • the stock-to-sales ratio coefficient is used to characterize the ratio between the stock and the shipment;
  • the second time A segment is a time segment adjacent to the first time segment and after the first time segment.
  • the determining unit is further configured to:
  • the shipments in the first historical time period corresponding to the first time period and the shipments in the second historical time period corresponding to the third time period determine that the shipments in the first historical time period are relative to the a first growth rate of shipments for the second historical time period;
  • the prediction unit is further configured to:
  • the prediction unit 602 is further configured to:
  • the promotional data of the first time period into the fitting regression model to predict the reference information of the goods in the first time period;
  • the promotional data is used to characterize the promotional information and promotion plan of the goods;
  • the above reference information is used to represent the exposure and price of the goods;
  • the sales forecast model Inputting the reference information of the first time period into a sales forecast model to predict shipments in the first time period; the sales forecast model is used to predict shipments.
  • the device further includes a processing unit configured to simulate the promotional data of the first historical time period corresponding to the first time period with the reference information of the first historical time period combined to obtain the fitted regression model.
  • the determining unit is further configured to:
  • the shipments in the second historical time period corresponding to the third time period and the shipments in the third historical time period corresponding to the second time period determine that the shipments in the third historical time period are relative to the a second growth rate in shipments for the second historical time period;
  • the prediction unit is further configured to:
  • the promotional data of the second time period into a fitting regression model to predict the reference information of the goods in the second time period;
  • the promotional data is used to characterize the promotional information and promotion plan of the goods;
  • the above reference information is used to represent the exposure and price of the goods;
  • the sales forecast model Inputting the reference information of the second time period into a sales forecast model to predict shipments in the second time period; the sales forecast model is used to predict shipments.
  • An embodiment of the present application also provides an electronic device, including a memory and a processor, the memory stores a computer program that can run on the processor, and the processor implements the data processing provided in the above-mentioned embodiments when executing the program method.
  • the embodiment of the present application also provides a storage medium, that is, a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the data processing method provided in the above-mentioned embodiment is implemented.
  • FIG. 7 is a schematic diagram of a hardware entity of an electronic device according to an embodiment of the present application.
  • the electronic device 700 includes: a processor 701, at least one communication bus 702, and at least one external communication interface 704 and memory 705.
  • the communication bus 702 is configured to realize connection and communication between these components.
  • the electronic device 700 further includes: a user interface 703, wherein the user interface 703 may include a display screen, and the external communication interface 704 may include a standard wired interface and a wireless interface.
  • the memory 705 is configured to store instructions and applications executable by the processor 701, and can also cache data to be processed or processed by the processor 701 and various modules in the electronic device (for example, image data, audio data, voice communication data and video data) Communication data), which can be realized by flash memory (FLASH) or random access memory (Random Access Memory, RAM).
  • the disclosed devices and methods can be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division.
  • the coupling, or direct coupling, or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be electrical, mechanical or other forms of.
  • the units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed to multiple network units; Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be used as a single unit, or two or more units can be integrated into one unit; the above-mentioned integration
  • the unit can be realized in the form of hardware or in the form of hardware plus software functional unit.
  • the above-mentioned integrated units of the present application are realized in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium.
  • the computer software products are stored in a storage medium and include several instructions to make A computer device (which may be a personal computer, a server, or a network device, etc.) executes all or part of the methods described in the various embodiments of the present application.
  • the aforementioned storage medium includes various media capable of storing program codes such as removable storage devices, ROMs, magnetic disks or optical disks.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Accounting & Taxation (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Finance (AREA)
  • Theoretical Computer Science (AREA)
  • Development Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本申请公开了一种数据处理方法，所述方法包括：根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；第一时间段包括至少两个时间子段；第一时间子段为至少两个时间子段中的任一历史时间子段；目标占比序列为至少一个时间子段中不同时间子段的备货量的占比；根据备货总量和目标占比序列中的目标占比，预测第一时间段中第二时间子段的备货量；目标占比为目标占比序列中第二时间子段对应的备货量的占比。另外，本申请还公开了一种数据处理装置、设备及存储介质。

Description

数据处理方法、装置、设备及存储介质
相关申请的交叉引用
本申请基于申请号为202110942369.4，申请日为2021年08月17日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此以引入方式并入本申请。
技术领域
本申请涉及数据处理领域，涉及但不限于一种数据处理处理方法、装置、设备及存储介质。
背景技术
相关技术中，电商平台在例如6月18日、11月11日的大型促销活动场景下，对于备货量的预估，是根据人工经验来进行的，这样，造成预估的备货量的准确度不高的问题。
发明内容
本申请为解决相关技术中存在的至少一个问题而提供一种数据处理方法、装置、设备及存储介质。
本申请的技术方案是这样实现的：
在一种实施方式中，本申请实施例提供一种数据处理方法，所述方法包括：
根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中的任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；
根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量；所述目标占比为所述目标占比序列中所述第二时间子段对应的备货量的占比。
在一种实施方式中，本申请实施例提供一种数据处理装置，所述装置包括：
确定单元，配置为根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中的任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；
预测单元，配置为根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量，所述目标占比为所述目标占比序列中所述第二时间子段对应的备货量的占比。
在一种实施方式中，本申请实施例提供一种电子设备，存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现上述数据处理方法。
在一种实施方式中，本申请实施例提供一种存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现上述数据处理方法。
附图说明
图1为本申请实施例提供的一种数据处理系统的可选的结构示意图；
图2为本申请实施例提供的数据处理方法的可选的流程示意图；
图3为本申请实施例提供的数据处理方法的可选的流程示意图；
图4为本申请实施例提供的可选的界面示意图；
图5为本申请实施例提供的数据处理方法的可选的流程示意图；
图6为本申请实施例提供的数据处理装置的可选的结构示意图；
图7为本申请实施例提供的电子设备的可选的结构示意图。
具体实施方式
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附图，对申请的具体技术方案做进一步详细描述。以下实施例用于说明本申请，但不用来限制本申请的范围。
除非另有定义，本文所使用的所有的技术和科学术语与属于本申请的技术领域的技术人员通常理解的含义相同。本文中在本申请的说明书中所使用的术语只是为了描述具体的实施例的目的，不是旨在于限制本申请。本文所使用的术语“和/或”包括一个或多个相关的所列项目的任意的和所有的组合。
对本申请进行进一步详细说明之前，对本申请实施例中涉及的名词和术语进行说明，本申请实施例中涉及的名词和术语适用于如下的解释。
1)、库存量单位(Stock Keeping Unit，SKU)，用于表征库存管理中的最小可用单元，比如，如果一个商品有至少一种颜色，那么所述至少一种颜色中的每一种颜色都可以为一个SKU，例如，对于卫衣而言，若所述卫衣有黑色和白色两种颜色，则黑色卫衣可以为一个SKU，白色卫衣可以为一个SKU。
2)、备货总量，用于表征在一时间段内所需要备的货的总量。时间段为一连续的时间段，例如、一年、一个月等设定时长的时间段，其中，一个时间段可划分为多个时间子段，比如：时间段为一年，则时间子段为一个月，再比如，时间段为一个月，时间子段为一天。
3)、备货量，用于表征在一时间子段内所需要备的货的数量。
在一示例中，备货量可以是在五月份的第一周内，所需要备的货物的数量。
4)、库存量，用于表征在仓库中剩余的货物的数量。
本申请实施例的数据处理方法可应用于图1所示的数据处理系统100，如图1所示，该数据处理系统100包括：服务器10和客户端20。其中，服务器10和客户端20之间通过网络30进行通信。
本申请实施例提供的数据处理方法可应用于数据处理设备，数据处理设备可为服务器10，也可为客户端20。
数据处理设备根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列，再根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量，其中，所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中的任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；所述目标占比为所述目标占比序列中所述第二时间子段对应的备货量的占比。
在数据处理设备为服务器10的情况下，服务器10将预测的第一时间段中第二时间子段的备货量通过网络30发送至客户端20，客户端20在接收到所述第二时间子段的备货量后，向用户展示所述第二时间子段的备货量。
在数据处理设备为客户端20的情况下，客户端20直接向用户展示第二时间子段的备货量。
下面，结合图1所示的数据处理系统100的示意图，对本申请实施例提供的数据处理方法、装置、设备和存储介质的各实施例进行说明。
图2为本申请实施例提供的一种数据处理方法的实现流程示意图，该方法应用于数据处理设备，如图2所示，该方法可以包括如下步骤：
S201、数据处理设备根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列。
这里，第一时间段为一个连续的时间段，所述连续的时间段可以为一天，也可以为一个月，还可以为一年，本申请实施例对此不进行限定。
第一时间段可以包括至少两个时间子段，第一时间子段为至少两个时间子段中的任一历史时间子段，其中，历史时间子段为已经发生过的时间子段。
在一示例中，第一时间段为一个月，该月包括的至少两个时间子段可包括：第一周、第二周、第三周和第四周；若第一周为已经发生过的时间子段，则第一周为第一时间子段。
在另一示例中，第一时间段为一年，该年包括的至少两个时间子段可包括：第一个月、第二个月、第三个月、……、第十一个月和第十二个月；若第一个月为已经发生过的时间子段，则第一个月为第一时间子段。
在又一示例中，第一时间段为一年，该年包括的至少两个时间子段可包括：第一季度、第二季度、第三季度和第四季度；若第一季度为已经发生过的时间子段，则第一季度为第一时间子段。
本申请实施例中，数据处理设备在确定第一时间段后，可以根据第一时间段，确定所述第一时间段中包括的第一时间子段，在确定所述第一时间子段后，确定所述第一时间子段的货物进出量。
这里，货物进出量包括：出货量和库存量，其中，出货量用于表征卖出的货物的数量，库存量 用于表征在仓库中剩余的货物的数量。
本申请实施例中，确定出货量的方法可以包括：数据处理设备获取至少一个订单数据；数据处理设备通过解析所述至少一个订单数据，确定出货量；其中，至少一个订单数据中每一订单数据用于表征用户购买货物的数量。
这里，获取至少一个订单数据，包括：从分布式文件系统(Hadoop Distributed File System，HDFS)中获取至少一个订单数据。
通过解析所述至少一个订单数据，确定出货量包括：通过对至少一个订单数据中每一订单数据进行汇总，确定出货量。
在一示例中，数据处理设备从HDFS中获取3个订单数据，在第一个订单数据中，记录有用户A购买货物A的数量为10个，在第二个订单数据中，记录有用户B购买货物A的数量为20个，在第三个订单数据中，记录有用户C购买货物A的数量为10个，数据处理设备通过对所述三个订单数据中记录的用户购买货物A的数量进行汇总，确定卖出的货物A的数量为40个。
本申请实施例中，货物可以为SKU，也可以为一个商品类别下的所有商品，本申请实施例对此不进行任何限定。
在一示例中，在货物为一个商品类别下的所有商品，且商品类别为卫衣的情况下，货物可以包括该卫衣类别下所有尺码的卫衣和所有颜色的卫衣。
本申请实施例中，所述至少一个占比序列为所述第一时间段对应的至少一个占比序列，对于所述至少一个占比序列中每一所述占比序列而言，所述占比序列包括至少一个占比，所述占比为对应时间子段的备货量占所述第一时间段的备货总量的比值。
在一示例中，第一时间段为5月，5月包括的至少两个时间子段可以包括：第一周、第二周、第三周和第四周；所述5月的至少一个占比序列可以包括：[20％、30％、40％、10％]，对于占比序列[20％、30％、40％、10％]而言，该占比序列包括占比20％、占比30％、占比40％和占比10％，其中，占比20％为第一周的备货量占5月的备货总量的比值，占比30％为第二周的备货量占5月的备货总量的比值，占比40％为第三周的备货量占5月的备货总量的比值，占比40％为第四周的备货量占5月的备货总量的比值。
本申请实施例中，数据处理设备可接收至少一个占比序列；将所述至少一个占比序列确定为所述至少一个占比序列。
这里，接收输入的至少一个占比序列包括：数据处理设备接收用户输入的至少一个占比序列。
在一示例中，若接收的至少一个占比序列包括：[20％、30％、40％、10％]和[30％、20％、10％、40％]，则确定的至少一个占比序列为[20％、30％、40％、10％]和[30％、20％、10％、40％]。
本申请实施例中，数据处理设备接收历史占比序列；在所述历史占比序列的基础上，对于每一时间子段对应的占比，可以增加或减少调节比例，确定所述至少一个占比序列；其中，所述历史占比序列为第一时间段对应的第一历史时间段的占比序列；调节步长用于表征在可调节范围内，每次可调节的比例；可调节范围用于表征在对每个占比进行调节时的一个最大调节范围；。
在一示例中，第一时间段为5月，第一时间段5月对应的第一历史时间段为去年5月，去年5月的占比序列，即历史占比序列为[30％、20％、10％、40％]，可调节范围为±5％，调节比例为1％，5月包括的至少两个时间子段可以包括：第一周、第二周、第三周和第四周。确定5月的至少一个占比序列可以包括：对于确定第一周对应的占比而言，可以增加调节比例1％，得到占比31％，对于确定第二周对应的占比而言，可以减小调节比例1％，得到占比19％，对于确定第三周对应的占比而言，可以增加调节比例1％，得到占比11％，对于确定第四周对应的占比而言，可以减小调节比例1％，得到占比39％，从而可以确定一占比序列为[31％、29％、11％、39％]。
目标占比序列为至少一个时间子段中不同时间子段的备货量的占比。
在一示例中，至少一个占比序列包括：占比序列1[20％、30％、40％、10％]和占比序列2[30％、20％、10％、40％]，则可以从占比序列1和占比序列2中，确定目标占比序列为占比序列1。
S202、数据处理设备根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量。
这里，目标占比为目标占比序列中第二时间子段对应的备货量的占比。
在一示例中，若备货总量为400，目标占比序列为[20％、30％、40％、10％]，其中，30％为第二时间子段对应的备货量的占比，则可以根据备货总量400和目标占比30％，预测第二时间子段的备货量为120。
本申请实施例提供一种数据处理方法，根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量，所述目标占比为所述目标占比序列中所述第二时间子段对应的备货量的占比。这样，在预测第一时间段中第二时间子段的备货量的过程中，可以先根据第一时间段的备货量和已发生的第一时间子段的货物进出量，确定目标占比序列，再根据所述备货总量和目标占比序列中的目标占比，预测未发生的第二时间子段的备货量。这样，可以提高预估的第二时间子段的备货量的准确度。
在一些实施例中，如图3所示，上述S201可以包括：
S301、根据第一时间段的备货总量和所述至少一个占比序列中的参考占比，确定至少一个参考备货量。
这里，第一时间段的备货总量为400，至少一个占比序列包括：占比序列1：[20％、30％、40％、10％]和占比序列2：[25％、35％、25％、25％]，对于第一时间段中的第一时间子段而言，第一时间子段的参考占比可以为20％，也可以为25％，可以根据第一时间段的备货总量400和第一时间子段的参考占比20％，确定一第一时间子段的参考备货量为80，还可以根据第一时间段的备货总量400和第一时间子段的参考占比25％，确定一第一时间子段的参考备货量为100，所述两个第一时间子段的参考备货量80和100组成至少一个参考备货量。
S302、根据所述至少一个参考备货量和所述第一时间子段的货物进出量，确定针对所述第一时间子段的至少一个现货率。
这里，对于第一时间子段而言，可以根据第一时间子段的至少一个参考备货量和第一时间子段的货物进出量，确定针对第一时间子段的至少一个现货率，其中，现货率为出货量和待售量的比值。
S303、将所述至少一个现货率中满足条件的目标现货率对应的参考备货量作为目标参考备货量。
这里，满足条件的目标现货率可以包括：将至少一个现货率中现货率最高的作为目标现货率。
在一示例中，至少一个现货率包括：30％、50％和60％，其中，30％对应的参考备货量为50，50％对应的参考备货量为80，60％对应的参考备货量为100，若目标现货率为60％，则将目标现货率60％对应的参考备货量100作为目标参考备货量。
S304、将至少一个占比序列中，得到所述目标参考备货量的占比序列确定为所述目标占比序列。
在一示例中，至少一个占比序列包括：占比序列1：[20％、30％、40％、10％]和占比序列2：[25％、35％、25％、25％]，若得到目标参考备货量的占比序列为占比序列1，则占比序列1为目标占比序列。
本申请实施例中，可以根据第一时间段的备货总量、第一时间子段的货物进出量和供应商送货时长(vendor lead time，VLT)，从至少一个占比序列中确定目标占比序列。
在一些实施例中，货物进出量包括出货量和库存量，上述S302可以包括：根据所述至少一个参考备货量和所述库存量，确定至少一个待售量；基于所述出货量和所述至少一个待售量，确定所述至少一个现货率。
这里，对于第一时间子段而言，若第一时间子段的出货量为50，第一时间子段的库存量为100，至少一个参考备货量包括：80和100，则可根据一参考备货量80和库存量100，确定一待售量为180，再基于出货量50和待售量180，确定一现货率为27％；还可根据一参考备货量100和库存量100，确定一待售量为200，再基于出货量50和待售量200，确定一现货率为25％。
上述在对确定至少一个现货率进行解释时，是以确定第一时间子段的至少一个现货率为例进行解释的，在实际应用中，确定第一时间段的现货率，可以是先确定所述第一时间段包括的至少两个时间子段中每一时间子段的现货率，再对每一时间子段的现货率进行求和计算，得到第一时间段的现货率。
本申请实施例中，数据处理设备确定第一时间段的至少一个现货率包括：将至少一个占比序列中每一占比序列输入至混合整数规划(Mixed-Integer Programming，MIP)模型，得到每一占比序列对应的一个现货率。
所述MIP模型的模型目标为最小化周转
Figure PCTCN2022108897-appb-000001
其中，R为占比序列，i为至少一个占比序列，itoi表示第i个占比序列对应的周转率，min表示最小的周转率。
所述MIP模型的约束条件为最大现货率，其中，最大现货率可以通过下述公式(1)实现：
s.t.cr i≥cr max,i∈[1,T]       公式(1)；
其中，s.t.为subjuct to的缩写，用于表征约束条件，cri表示第i个占比序列对应的现货率，crmax 表示最大的现货率，s.t.cr i≥cr max用于表征第i个占比序列对应的现货率应大于或等于最大现货率。
在确定至少一个现货率后，可以利用数学优化技术(CPLEX)，从所述至少一个现货率确定出最大的现货率。
在一些实施例中，所述方法还可以包括：根据所述第一时间段的第一时间的库存量、所述第一时间段的第二时间的库存量和所述第一时间段的出货量，确定所述备货总量；所述第一时间为所述第一时间段的起始时间，所述第二时间为所述第一时间段的截止时间。
这里，根据所述第一时间段的第一时间的库存量、所述第一时间段的第二时间的库存量和所述第一时间段的出货量，确定所述备货总量可以包括：第二时间的库存量减去第一时间的库存量，再加上第一时间段的出货量，确定备货总量。
在一示例中，第一时间段为5月，第一时间5月的起始时间5月1日，第二时间为5月的截止时间5月31日，若5月1日的库存量为100，5月31日的库存量为200，5月的出货量为200，则确定的备货总量为200-100+200＝300。
在一些实施例中，所述方法还可以包括：根据第二时间段的出货量和存销比系数，确定所述第二时间的库存量。
这里，所述存销比系数用于表征库存量与出货量之间的比值；所述第二时间段为与所述第一时间段相邻且位于所述第时间段之后的时间段。
所述根据第二时间段的出货量和存销比系数，确定所述第二时间的库存量可以包括：第二时间段的出货量乘以存销比系数，确定第二时间的库存量。
在一示例中，若第二时间段的出货量为100，存销比系数为0.4，则可以根据第二时间段的出货量100乘以存销比系数0.4，确定第二时间的库存量为40。
在一些实施例中，所述方法还可以包括：根据所述第一时间段对应的第一历史时间段的出货量和第三时间段对应的第二历史时间段的出货量，确定所述第一历史时间段的出货量相对于所述第二历史时间段的出货量的第一增长率；根据所述第一增长率和所述第三时间段的出货量，预测所述第一时间段的出货量。
这里，在第一时间段为一个月的情况下，第一时间段对应的第一历史时间段为往年同期的时间段，比如，在第一时间段为今年5月的情况下，所述5月对应的第一历史时间段可以为去年5月，也可以为前年5月，本申请实施例对此不进行限定。
第三时间段对应的第二历史时间段为往年同期的时间段，比如，在第三时间段为今年4月的情况下，所述4月对应的第二历史时间段可以为去年4月，也可以为前年4月，本申请实施例对此不进行限定。
在一示例中，若第一时间段为今年5月，第一历史时间段为去年5月，所述去年5月的出货量为200，第三时间段为今年4月，所述今年4月的出货量为100，第二历史时间段为去年4月，所述去年4月的出货量为100，则可以根据去年5月的出货量200和去年4月的出货量100，确定去年5月的出货量相对于去年4月的出货量的第一增长率为100％，再根据第一增长率100％和今年4月的 出货量100，预测5月的出货量为200。
在一些实施例中，所述方法还可以包括：数据处理设备将所述第一时间段的促销数据输入至拟合回归模型，预测所述货物在所述第一时间段的参考信息；将所述第一时间段的参考信息输入至销量预测模型，预测所述第一时间段的出货量。
这里，所述促销数据用于表征货物的促销信息和促销计划，其中，促销信息用于表征一具体的优惠形式，促销计划用于表征在整个促销过程中期望销售的货物总量；所述参考信息用于表征所述货物的曝光量和价格；所述销量预测模型用于对出货量进行预测。
促销数据可以通过给定模板输入。其中，给定模板可包括：货物、促销信息和所述促销信息所持续的时间。
促销信息可以包括：满减优惠信息、满赠优惠信息和秒杀信息。在一示例中，满减优惠信息为：满100减50。在一示例中，满赠优惠信息为：满100赠100。在一示例中，秒杀信息为：限时秒杀。
在一示例中，促销计划可以为：在为期两个月的促销过程中期望销售的货物总量为1000件。
本申请实施例中，拟合回归模型可以包括：逻辑回归(Logistic Regression，LR)模型、最大梯度提升(Extreme Gradient Boosting，XGBoost)模型、轻量级梯度提升机(Light Gradient Boosting Machine，LightGBM)模型和卷积神经网络(Convolutional Neural Networks，CNN)等。
确定销量预测模型可以包括：数据处理设备获取货物的时序特征和属性特征；将所述时序特征、所述属性特征、所述货物的曝光量和所述货物的价格输入至参考销量预测模型，对所述参考销量预测模型进行训练，得到销量预测模型。
这里，销量预测模型可以包括：线性回归模型、时序序列模型和深度学习模型等，其中，时序序列模型可以包括：预言(Prophet)模型和霍尔特-温特(Holt Winter)模型等；深度学习模型可以包括：循环神经网络(Recurrent Neural Network，RNN)模型、多分位循环神经网络(Multi-Quantile Recurrent Neural Network，MQRNN)模型和长短期记忆(Long Short-Term Memory，LSTM)网络模型等。
这里，货物的时序特征可以包括：货物的历史销量特征，所述历史销量特征用于表征货物在历史时间段内的平均销售量。
在一示例中，历史销量特征可以包括：历史3天、历史5天、历史7天、历史14天或者历史30天的平均销售量。
货物的属性特征可以包括：货物的品牌、体型和体积等信息。
获取货物的属性特征可以包括：数据处理设备从商品信息系统中获取货物的属性信息，将获取的商品的属性信息存储到关系型数据库(MYSQL数据库)或HDFS中，对需要的货物的属性信息进行加工，从而得到货物的属性特征，其中，货物的属性特征可以通过属性值表示。
在一些实施例中，所述方法还可以包括：将所述第一时间段对应的第一历史时间段的促销数据与所述第一历史时间段的参考信息进行拟合，得到所述拟合回归模型。
在一示例中，第一时间段为5月，所述5月对应的第一历史时间段为去年5月，去年5月的促 销数据包括满减优惠，将去年5月的满减优惠的促销数据与去年5月的货物的曝光量和实际的到手价格进行拟合，得到拟合回归模型。
在一些实施例中，所述方法还可以包括：根据第三时间段对应的第二历史时间段的出货量和所述第二时间段对应的第三历史时间段的出货量，确定所述第三历史时间段的出货量相对于所述第二历史时间段的出货量的第二增长率；根据所述第二增长率和所述第三时间段的出货量，预测所述第二时间段的出货量。
这里，第二时间段对应的第三历史时间段为往年同期的时间段，比如，在第二时间段为今年6月的情况下，所述6月对应的第三历史时间段可以为去年6月，也可以为前年6月，本申请实施例对此不进行限定。
在一示例中，若第三时间段为今年4月，第二历史时间段为去年4月，去年4月的出货量为100，第二时间段为今年6月，第三历史时间段为去年6月，去年6月的出货量为200，则可以根据去年6月的出货量200和去年4月的出货量100，确定去年6月的出货量相对于去年4月的出货量的第二增长率为100％，再根据第二增长率100％和今年4月的出货量100，预测今年6月的出货量为200。
本申请实施例中，在确定SKU的第二增长率R1后，确定所述SKU所表示的品牌下的所有货物的第二增长率R2，并判断所述SKU的第二增长率R1和所述品牌下的所有货物的第二增长率R2的大小关系，若所述SKU的第二增长率R1大于三倍的所述品牌下的所述货物的第二增长率R2，则将所述SKU的第二增长率R1替换为所述品牌下的所有货物的第二增长率R2。
在一些实施例中，所述方法还可以包括：将所述第二时间段的促销数据输入至拟合回归模型，预测所述货物在所述第二时间段的参考信息；将所述第二时间段的参考信息输入至销量预测模型，预测所述第二时间段的出货量。
这里，所述促销数据用于表征货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；所述销量预测模型用于对出货量进行预测。其中，对于促销数据、参考信息和销量预测模型的解释，请参见上述实施例，此处不再赘述。
本申请实施例中，在得到货物的原始数据、中间数据和结果数据后，可以将所述三种数据存储至MYSQL数据库和HDFS中，并将存储到MYSQL数据库和HDFS中的数据存储至一平台，例如大数据集市平台中。其中，原始数据用于表征对数据进行加工前的数据，例如，货物的属性信息、商品编号和商品名称等；中间数据用于表征在得到结果数据之前，所利用的所有数据，例如，第一时间子段的货物进出量数据；结果数据用于表征对所述中间数据进行处理后所得到数据，例如，预测的第一时间段中第二时间子段的备货量。
这里，在将原始数据、中间数据和结果数据存储至MYSQL数据库和HDFS后，在所述MYSQL数据库和HDFS中，可以以数据表的形式，对所述原始数据、中间数据和结果数据进行存储。其中，数据表是一种以结构化存储数据的方式。
将数据存储到大数据集市平台后，可以利用Plumber直通车将数据推送至补货系统的MYSQL数据库中，并在显示界面上显示原始数据、中间数据和结果数据，同时，用户可根据自身需求选取 需要显示的数据。其中，在显示界面上显示的原始数据、中间数据和结果数据的示意图可以如图4所示。
在图4中，历史BAND中那一列中，字母A至F分别表示历史的销量排序，其中，字母A所表示的历史销量排序最高，字母F表示的历史销量排序最低。
计算机可视化表示将计算的数据、计算的过程以及计算的结果展示给用户。
采购在途表示供应商在采购货物，但是还没有送到仓库。
在电商零售场景下，618和双11是常见的全国购物狂欢节，由于电商平台商品丰富，大促场景下商品的需求波动大，这给补货管理带来了极大的挑战。补货不仅需要考虑千万种商品巨大的需求波动，同时需要解决大批量补货带来的入库产能安排，保证大促期间商品不断货，电商平台仓储成本和现金流占用得到合理控制。
为了完成商家和零售平台达成的销售目标，大促前双方会沟通相关的销售计划，根据计划进行促销活动的规划和安排，以及对应的补货方案。补货方案主要分为两部分，一为销售的预估，对于主推的爆品和畅销品SKU的价格优惠、活动粒度等的安排，人工根据历史经验预估未来大促的销售，由于电商SKU高达千万种，人工只能对SKU依据原有销售进行分层，对不同层级SKU进行预估，难以细化进行，最后根据预估销售进行补货总量估计；二为入库节奏安排，由于大促前期整体入库量陡增，对于入库安排是个巨大的挑战，过早的入库会带来库存周转的增加，增加电商的仓储和现金流成本，过晚入库可能导致入库产能不足，无法入库，产生销售损失。人工会参考根据历史同期的入库节奏，与物流进行沟通，确定大致的入库节奏。不同采购者补货采用的入库逻辑和节奏可能都是不同的，没有统筹和规划。
本申请提出了一种数据处理方法，主要基于机器学习和运筹优化方法进行实现，主要实现步骤为数据获取、销售预估、入库节奏优化、大促补货方案输出和数据存储及输出共五个步骤，其中数据获取包括数据解析和商品信息获取等，销售预估利用统计方法和机器学习模型输出商品未来的销售预估，入库节奏优化利用仿真和运筹优化求解最优的入库节奏比例(即为上述实施例中所述的目标占比序列)，大促补货方案输出，利用销售预估和最优入库节奏计算商品大促期间多次下单日的补货建议量。
本申请结合了销量预估模块和入库节奏优化模块，进行自动化的大促补货商品决策。首先，获取商品已知的各类信息，包括基本属性信息(商品品牌、品类等)、时序信息(历史已发生的销量数据)、业务输入(销售计划、促销信息等)。然后，将输入信息输入销售预估装置和入库节奏优化装置，分别输出未来的销售预估和最优的入库节奏。最后，根据销售预估和最优入库节奏，给出SKU未来多次的大促备货建议量。具体流程如下所示：
模型的建立过程必要的硬件条件为电脑、服务器，以及至少一个数据库，所述存储器存储有可被所述处理器执行的程序指令，所述处理器调用所述程序指令能够执行“数据获取”、“销售预估时序模型回归”、“入库节奏优化求解”、“大促补货建议输出”以及“数据存储及输出”五步骤中任意一步。
步骤一、数据获取。
本申请需要商品属性数据、用户订单数据、业务输入的促销数据等。其中，商品属性数据可以从商品信息系统获取，主要通过商品主站信息进行爬取，存储到MYSQL和HDFS中，通过数据库操作，选取需要的商品信息进行加工和存储。用户订单数据通过分布式文件系统HDFS进行数据读取，通过解析用户订单数据，加工成商品每日的销量数据，例如，商品A在2020年1月1日共有10个用户下单，通过汇总10个用户的总销量，得到当日商品A的总销量为x件，对于解析数据存储到数据库中，方便后续模型进行调用和处理。业务输入的促销数据通过给定模板输入，系统进行加工处理。
本申请所需数据，需要定期更新，因此建立从数据源到数据库存储的数据流任务，定期获取数据，存储到所需的数据库内。获取的相关数据结构示例如下。
步骤二、销售预估。
销售预估通过业务输入进行未来销售的预估，由于大促时段不同于日常销售，整体需求的波动性会变大很多，而需求波动的主要原因来源于促销价格和流量的改变。因此，将整体预估分为以下两种类型：
第一种为业务没有输入对销售影响较大的促销计划和销售计划，对于此类SKU利用统计方法，预估未来的销售，具体步骤包括：计算SKU维度历史同期销量环比系数R1和R2，以及品牌维度整体环比系数；利用规则判断环比系数合理性，对于环比系数超过品牌整体N(N＝3)倍的进行平滑处理，所述平滑处理为用品牌维度整体环比系数代替；利用近期已发生的销量和环比系数，得到5月和6月的销售预估。
第二种为业务提供了促销信息和销售计划，对于有丰富信息的SKU，利用机器学习模型对未来的销售进行预估，具体步骤包括：根据SKU维度历史已发生的促销数据、销量数据，与实际的流量和到手价进行拟合，得到因子拟合回归模型，模型不限于LR、Xgboost、LightGBM、CNN等，然后根据业务输入的促销信息和计划，得到未来预估的流量和到手价；根据SKU的时序特征和属性特征，以及历史发生的流量和到手价特征进行销量预测模型训练，模型线性回归、时序预测算法(Prophet或Holt winter)和深度学习模型(RNN、MQRNN、LSTM)等；结合未来预估流量、到手价和训练好的销量预测模型，得到未来时序预测结果输出。
步骤三、确定入库比例。
这里，入库比例即为上述实施例中所述的占比序列，入库比例可以业务输入，也可以利用历史同期采购数据进行统计，计算第一周W1到第n周Wn的入库比例，同时给定入库比例可调节空间Range；根据入库比例和可调节空间Range，设置调节步长，组合所有可能的入库比例。
根据大促总备货量、所有的入库比例、期初库存I0+VLT，对SKU未来库存销量进行仿真，得到不同入库节奏对应的仿真结果；利用MIP模型求解SKU维度最优的入库节奏：目标为最小化周转，约束为最优现货率。
最后，利用优化求解器CPLEX求解器进行最优入库节奏求解，得到现货率最优情况下，周转最 优的目标入库比例。
最后，利用优化求解器CPLEX求解器进行最优入库节奏求解，得到现货率最优情况下，周转最优的入库节奏比例。
这里，目标入库比例为参考现货率所对应的入库比例，所述参考现货率为所述至少一个现货率中最高的现货率。
步骤四、预测大促补货建议量。
根据大促总备货量和最优入库节奏，可以算出SKU整体的大促备货计划。在每个大促备货节点，补货建议量＝预估的总备货量*此次节点的入库节奏比例。
步骤五、数据存储及输出。
结果的存储主要分为2个步骤，第一步为通过分布式文件管理系统HDFS对原始数据(属性特征和促销特征)、中间数据(中间变量数据)和结果数据进行存储，将数据表存储到大数据集市平台。第二步为，利用Plumber直通车将数据推送到补货系统的MYSQL数据库，利用补货系统前端进行原始数据、中间数据和结果数据的输出，同时，用户可根据自身需求选取系统展示数据。
如图5所示，本申请实施例提供的数据处理方法可包括：
S501、数据处理设备接收目标数据。
这里，所述目标数据可以包括：货物的销售计划、促销信息或销量数据。
S502、数据处理设备判断目标数据中是否包括销售计划和促销信息。
若目标数据中包括所述销售计划和促销信息，则执行S502a：利用机器学习模型进行销售预估。
若目标数据中不包括所述销售计划和促销信息，则执行S502b：利用历史销量数据进行销售预估。
S503、数据处理设备输出月度销售预估。
S504、数据处理设备根据销量数据，计算历史同期入库比例。
S505、数据处理设备根据调节比例，组成至少一个入库比例。
S506、对于不同入库比例，数据处理设备利用仿真对库存销量进行模拟，得到不同入库比例对应的库存量。
S507、数据处理设备利用运筹模型求解最优入库比例。
S508、数据处理设备根据月度销售预估和最优入库比例，确定大促补货计划。
图6为本申请实施例的提供的一种数据处理装置，如图6所示，该数据处理装置600包括：
确定单元601，配置为根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中的任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；
预测单元602，配置为根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量；所述目标占比为所述目标占比序列中所述第二时间子段对应的备货 量的占比。
在一些实施例中，所述确定单元，还配置为：
根据第一时间段的备货总量和所述至少一个占比序列中的参考占比，确定至少一个参考备货量；
根据所述至少一个参考备货量和所述第一时间子段的货物进出量，确定针对所述第一时间子段的至少一个现货率；
将所述至少一个现货率中满足条件的目标现货率对应的参考备货量作为目标参考备货量；
将至少一个占比序列中，得到所述目标参考备货量的占比序列确定为所述目标占比序列。
在一些实施例中，所述货物进出量包括出货量和库存量，所述确定单元，还配置为：
根据所述至少一个参考备货量和所述库存量，确定至少一个待售量；
基于所述出货量和所述至少一个待售量，确定所述至少一个现货率。
在一些实施例中，所述确定单元，还配置为：
根据所述第一时间段的第一时间的库存量、所述第一时间段的第二时间的库存量和所述第一时间段的出货量，确定所述备货总量；
所述第一时间为所述第一时间段的起始时间，所述第二时间为所述第一时间段的截止时间。
在一些实施例中，所述确定单元，还配置为：
根据第二时间段的出货量和存销比系数，确定所述第二时间的库存量；所述存销比系数用于表征库存量与出货量之间的比值；所述第二时间段为与所述第一时间段相邻且位于所述第时间段之后的时间段。
在一些实施例中，所述确定单元，还配置为：
根据所述第一时间段对应的第一历史时间段的出货量和第三时间段对应的第二历史时间段的出货量，确定所述第一历史时间段的出货量相对于所述第二历史时间段的出货量的第一增长率；
所述预测单元，还配置为：
根据所述第一增长率和所述第三时间段的出货量，预测所述第一时间段的出货量。
在一些实施例中，所述预测单元602，还配置为：
将所述第一时间段的促销数据输入至拟合回归模型，预测所述货物在所述第一时间段的参考信息；所述促销数据用于表征所述货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；
将所述第一时间段的参考信息输入至销量预测模型，预测所述第一时间段的出货量；所述销量预测模型用于对出货量进行预测。
在一些实施例中，所述装置还包括处理单元，所述处理单元配置为将所述第一时间段对应的第一历史时间段的促销数据与所述第一历史时间段的参考信息进行拟合，得到所述拟合回归模型。
在一些实施例中，所述确定单元，还配置为：
根据第三时间段对应的第二历史时间段的出货量和所述第二时间段对应的第三历史时间段的出货量，确定所述第三历史时间段的出货量相对于所述第二历史时间段的出货量的第二增长率；
根据所述第二增长率和所述第三时间段的出货量，预测所述第二时间段的出货量。
在一些实施例中，所述预测单元，还配置为：
将所述第二时间段的促销数据输入至拟合回归模型，预测所述货物在所述第二时间段的参考信息；所述促销数据用于表征所述货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；
将所述第二时间段的参考信息输入至销量预测模型，预测所述第二时间段的出货量；所述销量预测模型用于对出货量进行预测。
本申请实施例还提供一种电子设备，包括存储器和处理器，所述存储器存储有可在处理器上运行的计算机程序，所述处理器执行所述程序时实现上述实施例中提供的数据处理方法。
本申请实施例还提供一种存储介质，也就是计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现上述实施例中提供的数据处理方法。
这里需要指出的是：以上存储介质和设备实施例的描述，与上述方法实施例的描述是类似的，具有同方法实施例相似的有益效果。对于本申请存储介质和设备实施例中未披露的技术细节，请参照本申请方法实施例的描述而理解。
需要说明的是，图7为本申请实施例电子设备的一种硬件实体示意图，如图7所示，所述电子设备700包括：一个处理器701、至少一个通信总线702、至少一个外部通信接口704和存储器705。其中，通信总线702配置为实现这些组件之间的连接通信。在一示例中，电子设备700还包括：用户接口703、其中，用户接口703可以包括显示屏，外部通信接口704可以包括标准的有线接口和无线接口。
存储器705配置为存储由处理器701可执行的指令和应用，还可以缓存待处理器701以及电子设备中各模块待处理或已经处理的数据(例如，图像数据、音频数据、语音通信数据和视频通信数据)，可以通过闪存(FLASH)或随机访问存储器(Random Access Memory，RAM)实现。
应理解，说明书通篇中提到的“一个实施例”或“一实施例”意味着与实施例有关的特定特征、结构或特性包括在本申请的至少一个实施例中。因此，在整个说明书各处出现的“在一个实施例中”或“在一些实施例中”未必一定指相同的实施例。此外，这些特定的特征、结构或特性可以任意适合的方式结合在一个或多个实施例中。应理解，在本申请的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。
在本申请所提供的几个实施例中，应该理解到，所揭露的设备和方法，可以通过其它的方式实 现。以上所描述的设备实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，如：多个单元或组件可以结合，或可以集成到另一个系统，或一些特征可以忽略，或不执行。另外，所显示或讨论的各组成部分相互之间的耦合、或直接耦合、或通信连接可以是通过一些接口，设备或单元的间接耦合或通信连接，可以是电性的、机械的或其它形式的。
上述作为分离部件说明的单元可以是、或也可以不是物理上分开的，作为单元显示的部件可以是、或也可以不是物理单元；既可以位于一个地方，也可以分布到多个网络单元上；可以根据实际的需要选择其中的部分或全部单元来实现本实施例方案的目的。
另外，在本申请各实施例中的各功能单元可以全部集成在一个处理单元中，也可以是各单元分别单独作为一个单元，也可以两个或两个以上单元集成在一个单元中；上述集成的单元既可以采用硬件的形式实现，也可以采用硬件加软件功能单元的形式实现。
本领域普通技术人员可以理解：实现上述方法实施例的全部或部分步骤可以通过程序指令相关的硬件来完成，前述的程序可以存储于计算机可读取存储介质中，该程序在执行时，执行包括上述方法实施例的步骤；而前述的存储介质包括：移动存储设备、只读存储器(Read Only Memory，ROM)、磁碟或者光盘等各种可以存储程序代码的介质。
或者，本申请上述集成的单元如果以软件功能模块的形式实现并作为独立的产品销售或使用时，也可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请实施例的技术方案本质上或者说对相关技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机、服务器、或者网络设备等)执行本申请各个实施例所述方法的全部或部分。而前述的存储介质包括：移动存储设备、ROM、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (22)

  1. 一种数据处理方法，所述方法包括：
    根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中的任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；
    根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量；所述目标占比为所述目标占比序列中所述第二时间子段对应的备货量的占比。
  2. 根据权利要求1所述的方法，其中，所述根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列，包括：
    根据第一时间段的备货总量和所述至少一个占比序列中的参考占比，确定至少一个参考备货量；
    根据所述至少一个参考备货量和所述第一时间子段的货物进出量，确定针对所述第一时间子段的至少一个现货率；
    将所述至少一个现货率中满足条件的目标现货率对应的参考备货量作为目标参考备货量；
    将至少一个占比序列中，得到所述目标参考备货量的占比序列确定为所述目标占比序列。
  3. 根据权利要求2所述的方法，其中，所述货物进出量包括出货量和库存量；所述根据所述至少一个参考备货量和第一时间子段的货物进出量，确定针对所述第一时间子段的至少一个现货率，包括：
    根据所述至少一个参考备货量和所述库存量，确定至少一个待售量；
    基于所述出货量和所述至少一个待售量，确定所述至少一个现货率。
  4. 根据权利要求1所述的方法，其中，所述方法还包括：
    根据所述第一时间段的第一时间的库存量、所述第一时间段的第二时间的库存量和所述第一时间段的出货量，确定所述备货总量；
    所述第一时间为所述第一时间段的起始时间，所述第二时间为所述第一时间段的截止时间。
  5. 根据权利要求4所述的方法，其中，所述方法还包括：
    根据第二时间段的出货量和存销比系数，确定所述第二时间的库存量；所述存销比系数用于表征库存量与出货量之间的比值；所述第二时间段为与所述第一时间段相邻且位于所述第时间段之后的时间段。
  6. 根据权利要求4所述的方法，其中，所述方法还包括：
    根据所述第一时间段对应的第一历史时间段的出货量和第三时间段对应的第二历史时间段的出货量，确定所述第一历史时间段的出货量相对于所述第二历史时间段的出货量的第一增长率；
    根据所述第一增长率和所述第三时间段的出货量，预测所述第一时间段的出货量。
  7. 根据权利要求4所述的方法，其中，所述方法还包括：
    将所述第一时间段的促销数据输入至拟合回归模型，预测所述货物在所述第一时间段的参考信息；所述促销数据用于表征所述货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；
    将所述第一时间段的参考信息输入至销量预测模型，预测所述第一时间段的出货量；所述销量预测模型用于对出货量进行预测。
  8. 根据权利要求7所述的方法，其中，所述方法还包括：
    将所述第一时间段对应的第一历史时间段的促销数据与所述第一历史时间段的参考信息进行拟合，得到所述拟合回归模型。
  9. 根据权利要求5所述的方法，其中，所述方法还包括：
    根据第三时间段对应的第二历史时间段的出货量和所述第二时间段对应的第三历史时间段的出货量，确定所述第三历史时间段的出货量相对于所述第二历史时间段的出货量的第二增长率；
    根据所述第二增长率和所述第三时间段的出货量，预测所述第二时间段的出货量。
  10. 根据权利要求5所述的方法，其中，所述方法还包括：
    将所述第二时间段的促销数据输入至拟合回归模型，预测所述货物在所述第二时间段的参考信息；所述促销数据用于表征所述货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；
    将所述第二时间段的参考信息输入至销量预测模型，预测所述第二时间段的出货量；所述销量预测模型用于对出货量进行预测。
  11. 一种数据处理装置，所述装置包括：
    确定单元，配置为根据第一时间段的备货总量和第一时间子段的货物进出量，从至少一个占比序列中确定目标占比序列；所述第一时间段包括至少两个时间子段；所述第一时间子段为所述至少两个时间子段中的任一历史时间子段；所述目标占比序列为所述至少一个时间子段中不同时间子段的备货量的占比；
    预测单元，配置为根据所述备货总量和所述目标占比序列中的目标占比，预测所述第一时间段中第二时间子段的备货量，所述目标占比为所述目标占比序列中所述第二时间子段对应的备货量的占比。
  12. 根据权利要求11所述的数据处理装置，其中，所述确定单元，还配置为：
    根据第一时间段的备货总量和所述至少一个占比序列中的参考占比，确定至少一个参考备货量；
    根据所述至少一个参考备货量和所述第一时间子段的货物进出量，确定针对所述第一时间子段的至少一个现货率；
    将所述至少一个现货率中满足条件的目标现货率对应的参考备货量作为目标参考备货量；
    将至少一个占比序列中，得到所述目标参考备货量的占比序列确定为所述目标占比序列。
  13. 根据权利要求12所述的数据处理装置，其中，所述货物进出量包括出货量和库存量；所述确定单元，还配置为：
    根据所述至少一个参考备货量和所述库存量，确定至少一个待售量；
    基于所述出货量和所述至少一个待售量，确定所述至少一个现货率。
  14. 根据权利要求11所述的数据处理装置，其中，所述确定单元，还配置为：
    根据所述第一时间段的第一时间的库存量、所述第一时间段的第二时间的库存量和所述第一时间段的出货量，确定所述备货总量；
    所述第一时间为所述第一时间段的起始时间，所述第二时间为所述第一时间段的截止时间。
  15. 根据权利要求14所述的数据处理装置，其中，所述确定单元，还配置为：
    根据第二时间段的出货量和存销比系数，确定所述第二时间的库存量；所述存销比系数用于表征库存量与出货量之间的比值；所述第二时间段为与所述第一时间段相邻且位于所述第时间段之后的时间段。
  16. 根据权利要求14所述的数据处理装置，其中，所述确定单元，还配置为：
    根据所述第一时间段对应的第一历史时间段的出货量和第三时间段对应的第二历史时间段的出货量，确定所述第一历史时间段的出货量相对于所述第二历史时间段的出货量的第一增长率；
    所述预测单元，还配置为：
    根据所述第一增长率和所述第三时间段的出货量，预测所述第一时间段的出货量。
  17. 根据权利要求14所述的数据处理装置，其中，所述预测单元，还配置为：
    将所述第一时间段的促销数据输入至拟合回归模型，预测所述货物在所述第一时间段的参考信息；所述促销数据用于表征所述货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；
    将所述第一时间段的参考信息输入至销量预测模型，预测所述第一时间段的出货量；所述销量预测模型用于对出货量进行预测。
  18. 根据权利要求17所述的数据处理装置，其中，所述数据处理装置还包括：处理单元，配置为：
    将所述第一时间段对应的第一历史时间段的促销数据与所述第一历史时间段的参考信息进行拟合，得到所述拟合回归模型。
  19. 根据权利要求15所述的数据处理装置，其中，所述确定单元，还配置为：
    根据第三时间段对应的第二历史时间段的出货量和所述第二时间段对应的第三历史时间段的出货量，确定所述第三历史时间段的出货量相对于所述第二历史时间段的出货量的第二增长率；
    根据所述第二增长率和所述第三时间段的出货量，预测所述第二时间段的出货量。
  20. 根据权利要求15所述的数据处理装置，其中，所述预测单元，还配置为：
    将所述第二时间段的促销数据输入至拟合回归模型，预测所述货物在所述第二时间段的参考信息；所述促销数据用于表征所述货物的促销信息和促销计划；所述参考信息用于表征所述货物的曝光量和价格；
    将所述第二时间段的参考信息输入至销量预测模型，预测所述第二时间段的出货量；所述销量预测模型用于对出货量进行预测。
  21. 一种电子设备，包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述计算机程序时，实现权利要求1至10任一项所述的数据处理方法。
  22. 一种存储介质，存储有计算机程序，，所述计算机程序被处理器执行时，实现权利要求1至10任一项所述的数据处理方法。
PCT/CN2022/108897 2021-08-17 2022-07-29 数据处理方法、装置、设备及存储介质 WO2023020255A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110942369.4A CN113657667A (zh) 2021-08-17 2021-08-17 一种数据处理方法、装置、设备及存储介质
CN202110942369.4 2021-08-17

Publications (1)

Publication Number Publication Date
WO2023020255A1 true WO2023020255A1 (zh) 2023-02-23

Family

ID=78479940

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/108897 WO2023020255A1 (zh) 2021-08-17 2022-07-29 数据处理方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN113657667A (zh)
WO (1) WO2023020255A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116342042A (zh) * 2023-05-25 2023-06-27 北京京东乾石科技有限公司 一种补货方法及装置、存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113657667A (zh) * 2021-08-17 2021-11-16 北京沃东天骏信息技术有限公司 一种数据处理方法、装置、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160092828A1 (en) * 2014-09-26 2016-03-31 Paul Sofsky System and Method for Taking Inventory and Ordering
CN107123017A (zh) * 2017-03-22 2017-09-01 重庆允升科技有限公司 一种工业品原料商品推荐方法
CN109978428A (zh) * 2017-12-28 2019-07-05 北京京东尚科信息技术有限公司 数据估算方法和装置以及计算机可读存储介质
CN113657667A (zh) * 2021-08-17 2021-11-16 北京沃东天骏信息技术有限公司 一种数据处理方法、装置、设备及存储介质

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2919294B2 (ja) * 1995-03-07 1999-07-12 株式会社神戸製鋼所 在庫計画策定支援装置
US7840454B2 (en) * 2005-09-07 2010-11-23 Sap Ag Methods and systems for dynamic calculation of safety stock quantity
CN106971348B (zh) * 2016-01-14 2021-04-30 阿里巴巴集团控股有限公司 一种基于时间序列的数据预测方法和装置
CN110363454B (zh) * 2018-04-09 2022-01-11 杉数科技(北京)有限公司 用于确定商品补货量的方法及装置
CN109272276B (zh) * 2018-09-29 2020-10-27 石化盈科信息技术有限责任公司 一种库存补货管理方法以及装置
CN113159355A (zh) * 2020-01-07 2021-07-23 北京京邦达贸易有限公司 数据预测方法、装置、物流货量预测方法、介质、设备
CN113095745A (zh) * 2020-01-09 2021-07-09 北京沃东天骏信息技术有限公司 补货决策模型训练和补货决策方法、系统、设备及介质
CN111582788A (zh) * 2020-04-29 2020-08-25 北京小米松果电子有限公司 物品调度方法、装置及计算机可读存储介质
CN112085442B (zh) * 2020-08-27 2023-12-22 太原市万尺云数智科技有限公司 设备操作方法、装置、电子设备和计算机可读介质
CN113159665A (zh) * 2021-03-04 2021-07-23 深圳市腾辉信息技术有限公司 基于电商平台的仓库备货方法、电子设备及存储介质
CN114742505A (zh) * 2022-04-20 2022-07-12 农夫山泉股份有限公司 库存控制方法和装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160092828A1 (en) * 2014-09-26 2016-03-31 Paul Sofsky System and Method for Taking Inventory and Ordering
CN107123017A (zh) * 2017-03-22 2017-09-01 重庆允升科技有限公司 一种工业品原料商品推荐方法
CN109978428A (zh) * 2017-12-28 2019-07-05 北京京东尚科信息技术有限公司 数据估算方法和装置以及计算机可读存储介质
CN113657667A (zh) * 2021-08-17 2021-11-16 北京沃东天骏信息技术有限公司 一种数据处理方法、装置、设备及存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116342042A (zh) * 2023-05-25 2023-06-27 北京京东乾石科技有限公司 一种补货方法及装置、存储介质
CN116342042B (zh) * 2023-05-25 2024-04-19 北京京东乾石科技有限公司 一种补货方法及装置、存储介质

Also Published As

Publication number Publication date
CN113657667A (zh) 2021-11-16

Similar Documents

Publication Publication Date Title
JP7340456B2 (ja) モデル生成のための動的な特徴選択
WO2023020255A1 (zh) 数据处理方法、装置、设备及存储介质
Papachristos et al. An inventory model with deteriorating items, quantity discount, pricing and time-dependent partial backlogging
US20210334844A1 (en) Method and system for generation of at least one output analytic for a promotion
US11010798B2 (en) System and method for integrating retail price optimization for revenue and profit with business rules
US10181138B2 (en) System and method for determining retail-business-rule coefficients from current prices
US20090327037A1 (en) System and Methods for Pricing Markdown with Model Refresh and Reoptimization
Sana Optimal production lot size and reorder point of a two-stage supply chain while random demand is sensitive with sales teams' initiatives
CN102282551A (zh) 娱乐入场券定价的自动决策支持
Priyan et al. Continuous review inventory model with controllable lead time, lost sales rate and order processing cost when the received quantity is uncertain
Kreng et al. An innovation diffusion of successive generations by system dynamics—An empirical study of Nike Golf Company
CN109784984A (zh) 一种生鲜门店智能订货系统
US20210312488A1 (en) Price-Demand Elasticity as Feature in Machine Learning Model for Demand Forecasting
van Donselaar et al. Heuristics for setting reorder levels in periodic review inventory systems with an aggregate service constraint
Hanukov et al. A multi-server system with inventory of preliminary services and stock-dependent demand
Smith Clearance pricing in retail chains
Ferreira et al. Demand learning and pricing for varying assortments
US11354686B2 (en) Short life cycle sales curve estimation
US20230419184A1 (en) Causal Inference Machine Learning with Statistical Background Subtraction
US20130073341A1 (en) Pricing markdown optimization system
Hosseinalifam et al. Network capacity control under a nonparametric demand choice model
US20220222689A1 (en) Methods and apparatus for surge-adjusted forecasting
CA2909957A1 (en) Large-scale customer-product relationship mapping and contact scheduling
JP2021039542A (ja) 情報処理方法および情報処理装置
JP6886186B2 (ja) 食品仕入れ数決定支援システム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22857567

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE