Accuracy evaluation index determination method and device for demand prediction, equipment, medium and product

Abstract

The invention provides an accuracy rate evaluation index determination method and device for demand prediction, equipment, a medium and a product. Relates to the technical field of warehouse logistics, in particular to the fields of intelligent industry, intelligent supply chains, intelligent logistics and artificial intelligence. The method comprises the steps of determining a demand category of a business demand according to a demand representation feature represented by historical order data of the business demand; according to the demand category, obtaining a candidate model from a preset model pool, and obtaining a candidate accuracy rate evaluation index from a preset accuracy rate evaluation index pool; constructing a regression model by taking a candidate accuracy evaluation index determined based on the candidate model as an independent variable, taking a demand attribute feature of the business demand as a control variable and taking an inventory index of the business demand as a dependent variable; performing regression calculation on the regression model, and determining a regression coefficient determined for the candidate accuracy rate evaluation indexes in the regression model; and determining a target accuracy evaluation index according to the target regression coefficient meeting a preset condition.

Description

Technical Field

[0001] The present disclosure relates to the field of warehousing and logistics technology, and in particular to the fields of smart industry, smart supply chain, smart logistics, artificial intelligence, etc. More specifically, it relates to a method, device, equipment, medium and product for determining an accuracy evaluation index for demand forecasting. Background Art

[0002] Demand forecasting is an important basis for inventory planning. Accurate demand forecasting helps to achieve reasonable inventory planning and reduce the cost of out-of-stock due to insufficient inventory or overstocking. Therefore, it is necessary to select a suitable demand forecasting method and use a suitable forecasting accuracy evaluation index to evaluate the effect of the forecasting algorithm in order to effectively ensure the accuracy of demand forecasting.

[0003] In actual business scenarios, based on the experience of technical personnel or customer specifications, a certain algorithm and prediction indicators will be uniformly used to predict all SKUs (Stock Keeping Unit, the unit of measurement for inventory in and out) to be predicted.

[0004] In the process of implementing the concept of the present disclosure, the inventors discovered that there are at least the following problems in the related art: inappropriate selection of algorithms and indicators. Summary of the invention

[0005] In view of this, the present disclosure provides a method, apparatus, device, medium and product for determining an accuracy evaluation index for demand forecasting.

[0006] One aspect of the present disclosure provides a method for determining an accuracy evaluation index for demand forecasting, comprising: determining a demand category of a business demand based on demand characterization features represented by historical order data of the business demand; obtaining, based on the demand category, at least one candidate model that is adapted to the demand category from a preset model pool, and obtaining at least one candidate accuracy evaluation index that is adapted to the demand category from a preset accuracy evaluation index pool, wherein the candidate accuracy evaluation index represents the accuracy of a prediction result obtained by forecasting the business demand based on the candidate model; for each candidate model in the at least one candidate model and each candidate accuracy evaluation index in the at least one candidate accuracy evaluation index, determining the accuracy of the prediction result based on the candidate model; and determining the accuracy of the prediction result based on the candidate model. The candidate accuracy evaluation index determined by the candidate model is used as an independent variable, the demand attribute characteristics of the business demand are used as a control variable, and the inventory index of the business demand is used as a dependent variable to construct a regression model to obtain at least one regression model, wherein the inventory index characterizes the inventory performance of the orders in the business demand; for each regression model in the at least one regression model, regression calculation is performed on the regression model to determine the regression coefficient determined for the candidate accuracy evaluation index in the regression model to obtain at least one regression coefficient; and according to the target regression coefficient that meets the preset conditions in the at least one regression coefficient, a target accuracy evaluation index is determined as the accuracy evaluation index determined in the demand forecasting process for the business demand.

[0007] Another aspect of the present disclosure provides an accuracy evaluation index determination device for demand forecasting, comprising: a demand category determination module, used to determine the demand category of the business demand based on the demand characterization features characterized by the historical order data of the business demand; an adaptation information acquisition module, used to obtain at least one candidate model adapted to the demand category from a preset model pool according to the demand category, and to obtain at least one candidate accuracy evaluation index adapted to the demand category from a preset accuracy evaluation index pool, wherein the candidate accuracy evaluation index represents the accuracy of the prediction result obtained by performing demand forecasting on the business demand based on the candidate model; a regression model construction module, used to determine the demand category of the business demand for each candidate model in the at least one candidate model and each candidate accuracy evaluation index in the at least one candidate accuracy evaluation index; An evaluation index is provided, wherein the candidate accuracy evaluation index determined based on the candidate model is used as an independent variable, the demand attribute characteristics of the business demand are used as a control variable, and the inventory index of the business demand is used as a dependent variable to construct a regression model to obtain at least one regression model, wherein the inventory index characterizes the inventory performance of the orders in the business demand; a regression calculation module is used to perform regression calculation on the regression model for each regression model in the at least one regression model, determine the regression coefficient determined for the candidate accuracy evaluation index in the regression model, and obtain at least one regression coefficient; and an index determination module is used to determine a target accuracy evaluation index according to a target regression coefficient that meets a preset condition in the at least one regression coefficient, as an accuracy evaluation index determined for the demand forecasting process of the business demand.

[0008] Another aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the accuracy evaluation index determination method for demand forecasting disclosed in the present disclosure.

[0009] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, which, when executed, are used to implement the method for determining an accuracy evaluation index for demand forecasting of the present disclosure.

[0010] Another aspect of the present disclosure provides a computer program product, which includes computer executable instructions, and when the instructions are executed, are used to implement the method for determining the accuracy evaluation index for demand forecasting of the present disclosure.

[0011] According to the embodiments of the present disclosure, the demand characterization features characterized by the historical order data of the business demand are used to determine the demand category of the business demand; according to the demand category, the candidate model is obtained from the preset model pool, and the candidate accuracy evaluation index is obtained from the preset accuracy evaluation index pool; the candidate accuracy evaluation index determined based on the candidate model is used as the independent variable, the demand attribute characteristics of the business demand are used as the control variable, and the inventory index of the business demand is used as the dependent variable to construct a regression model; regression calculation is performed on the regression model to determine the regression coefficient determined for the candidate accuracy evaluation index in the regression model; the technical means of determining the target accuracy evaluation index according to the target regression coefficient that meets the preset conditions not only considers the demand characterization features, and realizes the selection of candidate models and candidate accuracy evaluation indicators that are compatible with the business demand according to the demand category, but also further correlates the candidate accuracy evaluation index with the inventory index that characterizes the measurement standard of downstream inventory transfer, at least partially overcomes the technical problem of inappropriate selection of algorithms and indicators, and can be conducive to finding demand forecasting methods and accuracy evaluation indicators that are suitable for business scenarios and effectively improve inventory satisfaction rates. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0013] Figure 1 An exemplary system architecture to which the accuracy evaluation index determination method for demand forecasting can be applied according to an embodiment of the present disclosure is schematically shown;

[0014] Figure 2 A flowchart of a method for determining an accuracy evaluation index for demand forecasting according to an embodiment of the present disclosure is schematically shown;

[0015] Figure 3A The formula and illustration of the absolute error evaluation index according to the embodiment of the present disclosure are schematically shown;

[0016] Figure 3B The formula and illustration of the relative error evaluation index according to the embodiment of the present disclosure are schematically shown;

[0017] Figure 4 A diagram schematically shows a result of classifying SKUs based on the SBC method according to an embodiment of the present disclosure;

[0018] Figure 5 The overall flow chart for selecting the optimal accuracy evaluation index for each type of SKU according to an embodiment of the present disclosure is schematically shown;

[0019] Fig. 6A The classification result of the SKU of the automotive aftermarket industry according to the embodiment of the present disclosure is schematically shown;

[0020] Figure 6B The OLS regression coefficient results representing the accuracy rate of all SKUs and inventory indicators according to an embodiment of the present disclosure are schematically shown;

[0021] Figure 7 A block diagram schematically shows a device for determining an accuracy evaluation index for demand forecasting according to an embodiment of the present disclosure; and

[0022] Figure 8 A block diagram of an electronic device suitable for implementing a method for determining an accuracy evaluation index for demand forecasting according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0023] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0024] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0025] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0026] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0027] In the embodiments of the present disclosure, the collection, updating, analysis, processing, use, transmission, provision, disclosure, storage, etc. of the data involved (for example, including but not limited to user personal information) are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information, network security, and national security.

[0028] In the embodiments of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0029] In the process of realizing the concept of the present disclosure, the inventors found that the selection of prediction algorithms and accuracy evaluation indicators in the related art lacked consideration of the characteristics and applicability of different types of SKUs. Selecting inappropriate prediction methods and indicators is like choosing the wrong means of transportation and the wrong ruler, measuring the wrong distance to the destination, and choosing inappropriate means of transportation, which will result in extended time to reach the destination or failure to reach the destination. This shortcoming will cause losses in two aspects: on the one hand, it will cause the prediction effect to be contrary to the actual situation, and the prediction effect will be mistakenly believed to be good during training, but the actual algorithm will perform poorly after it is launched. On the other hand, it will make it difficult to tune the algorithm.

[0030] In addition, the selection of prediction accuracy evaluation indicators in related technologies does not take into account the measurement standards of downstream business of demand forecasting. It is usually determined based on the experience of technical personnel or customer specifications, which can easily lead to good prediction results, but the inventory strategy based on the prediction results performs poorly. In other words, the evaluation system of supply chain demand forecasting runs counter to the evaluation system of downstream business, which ultimately causes inventory backlogs or severe shortages, bringing losses to the company.

[0031] The purpose of this disclosure is to solve the problem of how to select demand forecasting methods and evaluation indicators in the context of the supply chain. Due to the different demand distributions in different industries, the forecasting methods and accuracy evaluation indicators cannot be selected in general, but the sales characteristics of SKUs need to be considered. In addition, demand forecasting serves inventory replenishment and allocation. Stable and accurate demand forecast input is an important guarantee for replenishment and allocation. Therefore, the selection of accuracy evaluation indicators for demand forecasting is particularly important, and the performance of inventory indicators needs to be considered to select accuracy evaluation indicators.

[0032] The embodiments of the present disclosure provide a method, device, equipment, medium and product for determining an accuracy evaluation index for demand forecasting. The method includes:

[0033] Figure 1The exemplary system architecture 100 to which the accuracy evaluation index determination method for demand forecasting can be applied according to an embodiment of the present disclosure is schematically shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0034] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0035] The user may use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, etc. (only for example).

[0036] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0037] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0038] It should be noted that the accuracy evaluation index determination method for demand forecasting provided by the embodiment of the present disclosure can generally be performed by the server 105. Accordingly, the accuracy evaluation index determination device for demand forecasting provided by the embodiment of the present disclosure can generally be set in the server 105. The accuracy evaluation index determination method for demand forecasting provided by the embodiment of the present disclosure can also be performed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the accuracy evaluation index determination device for demand forecasting provided by the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Alternatively, the accuracy evaluation index determination method for demand forecasting provided by the embodiment of the present disclosure may also be executed by the first terminal device 101, the second terminal device 102, or the third terminal device 103, or may also be executed by other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103. Accordingly, the accuracy evaluation index determination device for demand forecasting provided by the embodiment of the present disclosure may also be set in the first terminal device 101, the second terminal device 102, or the third terminal device 103, or may be set in other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103.

[0039] For example, the historical order data of the business demand may be originally stored in any one of the first terminal device 101, the second terminal device 102, or the third terminal device 103 (for example, the first terminal device 101, but not limited thereto), or stored in an external storage device and may be imported into the terminal device 101. Then, the first terminal device 101 may locally execute the accuracy evaluation index determination method for demand forecasting provided by the embodiment of the present disclosure, or send the historical order data to other terminal devices, servers, or server clusters, and the other terminal devices, servers, or server clusters that receive the historical order data may execute the accuracy evaluation index determination method for demand forecasting provided by the embodiment of the present disclosure.

[0040] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0041] Figure 2 A flowchart of a method for determining an accuracy evaluation index for demand forecasting according to an embodiment of the present disclosure is schematically shown.

[0042] like Figure 2 As shown, the method includes operations S201 to S205.

[0043] In operation S201 , a demand category of the business demand is determined according to demand characterization features characterized by historical order data of the business demand.

[0044] According to an embodiment of the present disclosure, historical order data may include the historical sales value of all SKUs every day within a certain period of time, and may not be limited thereto. Demand characterization features may characterize the characteristics of the SKU categories, sales frequency, etc. shown in the historical order data, and may not be limited thereto. For example, the category, sales frequency, etc. of the SKU may be used as a classification basis for which type of model and accuracy evaluation index is applicable.

[0045] In operation S202, according to the demand category, at least one candidate model that is compatible with the demand category is obtained from a preset model pool, and at least one candidate accuracy evaluation index that is compatible with the demand category is obtained from a preset accuracy evaluation index pool, wherein the candidate accuracy evaluation index represents the accuracy of the prediction result obtained by performing demand prediction on the business demand based on the candidate model.

[0046] According to an embodiment of the present disclosure, a preset model may characterize a model for demand forecasting. For example, at least one of the following models may be included: a time series model, a machine learning model, a deep learning model, a combination strategy model, etc., and may not be limited thereto. The time series model may, for example, include at least one of the following: a moving average class model, Holt-winters (Holt-Winters method, a time series analysis and forecasting method), ETS (Exponential Smoothing, exponential smoothing) model, ARIMA (Autoregressive Integrated Moving Average model, differential integrated moving average autoregressive model) model, Prophet (a time series model) model, etc., and may not be limited thereto. The machine learning model may, for example, include at least one of the following: CART (Classification and Regression Trees, classification and regression tree), random forest, XGBoost (Extreme Gradient Boosting, gradient boosting model), LightGBM (Light Gradient Boosting Machine, an open source machine learning algorithm), etc., and may not be limited thereto. The deep learning model may include, for example, at least one of the following: RNN (Recurrent Neural Network), Deep AR (Probabilistic Forecasting with Autoregressive Recurrent, an autoregressive model that combines deep learning and probability models), Deep State (a model that combines state space models with deep learning), MQ-RNN (a probability prediction model), etc., and may not be limited thereto. The combination strategy model represents the combination of models according to certain rules to form an algorithm strategy for prediction. The combination strategy model may include, for example, at least one of the following: weighted ensemble strategy (Weighted Ensemble Of Statistical Models), feature-based classification selection fusion strategy (Feature-based Forecast Model Averaging), etc., and may not be limited thereto.

[0047] According to an embodiment of the present disclosure, the preset accuracy evaluation index includes at least one of the following indicators: absolute error index, deviation index, relative error index, etc., and may not be limited thereto. The absolute error index can be a predicted value ( ) and the true value ( ) to measure the prediction accuracy, that is, . The advantage of this type of indicator is that it is easy to understand; the disadvantage is that it depends on the dimension, and it is meaningless to compare time series data of different dimensions together. The goal of the deviation indicator is to measure whether the predicted value is overestimated or underestimated compared to the true value. When the overall overestimate is, the indicator symbol is +; when it is underestimated, the indicator symbol is −. Although this type of indicator can well measure whether it is overestimated or underestimated, it cannot indicate the size of the deviation. The relative error indicator uses the deviation between the predicted value and the true value divided by the true value, that is, the percentage error method ( The advantage of this type of indicator is that it is dimensionless, easy to understand in the form of percentage, and can compare the advantages and disadvantages of prediction methods on different data.

[0048] According to the embodiments of the present disclosure, the index pool to be screened can be determined in advance to prepare for the subsequent selection of the index pool by SKU type. Figure 3A and Figure 3B shown.

[0049] Figure 3A The formula and illustration of the absolute error evaluation index according to the embodiment of the present disclosure are schematically shown.

[0050] Figure 3B The formula and illustration of the relative error evaluation index according to the embodiment of the present disclosure are schematically shown.

[0051] like Figure 3A and Figure 3B As shown, the advantages and descriptions of various indicators can be determined based on multiple experimental verifications.

[0052] According to the embodiments of the present disclosure, since different models and evaluation indicators have different advantages and disadvantages when applied to different scenarios, models and evaluation indicators suitable for a certain type of SKU can be selected according to the demand category to obtain candidate models and candidate accuracy evaluation indicators. The selection process can be realized by training the selection model using the demand category, model and evaluation indicator to achieve intelligent selection, or by sending the demand category to the client and having relevant personnel perform manual selection according to the demand type, which is not limited here.

[0053] It should be noted that when only one candidate accuracy evaluation indicator and one candidate model that are adapted to the business needs are obtained, the candidate accuracy evaluation indicator can be directly determined as the target accuracy evaluation indicator determined in the demand forecasting process for the business needs, and the candidate model can be directly determined as the target model for realizing the demand forecasting of the business needs.

[0054] In operation S203, for each candidate model in at least one candidate model and each candidate accuracy evaluation indicator in at least one candidate accuracy evaluation indicator, a regression model is constructed with the candidate accuracy evaluation indicator determined based on the candidate model as an independent variable, the demand attribute characteristics of the business demand as a control variable, and the inventory indicator of the business demand as a dependent variable to obtain at least one regression model, wherein the inventory indicator characterizes the inventory performance of orders in the business demand.

[0055] According to an embodiment of the present disclosure, the demand attribute characteristics may include at least one of the following: the supply source, price, category, etc. of the SKU, and may not be limited thereto. For example, in the field of automotive spare parts, the demand attribute characteristics of business requirements may include at least one of the following: the functional group, product group, supply strategy, supply source, pilot category, supply category, price, consignor, whether to continue replenishment, etc. to which the SKU belongs, and may not be limited thereto. Different business requirements may have different demand attribute characteristics.

[0056] According to an embodiment of the present disclosure, the inventory index may be determined based on SKU inventory surplus rate, fulfillment rate, turnover rate, etc., and may not be limited thereto.

[0057] According to the embodiments of the present disclosure, for each potential candidate accuracy evaluation index, the inventory index can be regressed by OLS (ordinary least squares) respectively, and the relevant characteristics of the SKU, that is, the above-mentioned demand attribute characteristics, can be controlled to construct a regression model. When there are M candidate models and N candidate accuracy evaluation indicators, M×N regression models can be constructed.

[0058] For example, the regression model may include a model as shown in Formula (1), but may not be limited thereto.

[0059] Formula (1)

[0061] In formula (1), Can indicate the Candidate accuracy evaluation indicators, Can indicate the Inventory indicators, It can indicate the functional group to which the SKU belongs. It can indicate the product group to which the SKU belongs. can represent the supply strategy, Can indicate the source of goods, It can indicate the pilot classification. Can represent supply category, Can indicate price, Can represent suppliers, can represent the replenishment decision, It can represent the regression coefficient.

[0062] In operation S204, for each regression model in the at least one regression model, regression calculation is performed on the regression model to determine a regression coefficient determined for a candidate accuracy evaluation indicator in the regression model, thereby obtaining at least one regression coefficient.

[0063] According to an embodiment of the present disclosure, each independent variable in the regression model constructed based on the aforementioned method may have a regression coefficient, and the regression model may also include other regression parameters. The process of performing regression calculation on the regression model may be characterized as a process of calculating the coefficient values ​​of these regression coefficients.

[0064] For example, in a specific embodiment, the target candidate model and the target candidate accuracy evaluation index reflected by the candidate accuracy evaluation index in each regression model may be determined first. Then, multiple experiments may be conducted under the given scenario defined by the target candidate model and the target candidate accuracy evaluation index to calculate multiple candidate accuracy evaluation index values, multiple inventory index values, and multiple demand attribute characteristics under the given scenario. Afterwards, by inputting the multiple candidate accuracy evaluation index values, multiple inventory index values, and multiple demand attribute characteristics into the regression model for calculation, the regression coefficients of the respective variables may be obtained, thereby extracting the regression coefficients of the candidate accuracy evaluation index in the regression model.

[0065] It should be noted that the regression calculation method is not limited to the above, and may also include other types of methods, as long as the regression coefficients of the respective variables can be obtained.

[0066] In operation S205, a target accuracy evaluation index is determined according to a target regression coefficient that meets a preset condition in at least one regression coefficient, as an accuracy evaluation index determined in a demand forecasting process for business needs.

[0067] According to an embodiment of the present disclosure, the regression coefficient may have a significant feature, and the significant feature may characterize the influence of the candidate accuracy evaluation index on the next inventory index. The preset condition may include at least one of the following: a significant feature value characterizing the significant feature is greater than a preset threshold, the value of the regression coefficient is the largest among the values ​​of the at least one regression coefficient, etc., and may not be limited thereto.

[0068] According to an embodiment of the present disclosure, the significance level of the regression coefficient can be determined by calculating the P value. The P value can represent the probability of the sample observation result or more extreme result when the null hypothesis is true. In this embodiment, it can represent the influence of the candidate accuracy evaluation index on the next inventory index.

[0069] According to an embodiment of the present disclosure, through the aforementioned regression calculation, the regression coefficient of each candidate accuracy index to the corresponding inventory index and the significance of each regression coefficient can be obtained. For example, in the case where there are M×N regression models, M×N regression coefficients corresponding to the candidate accuracy evaluation index can be calculated based on the aforementioned calculation method. The candidate accuracy evaluation index corresponding to the target regression coefficient with the largest regression coefficient can be selected and determined as the target accuracy evaluation index.

[0070] Through the above-mentioned embodiments of the present disclosure, not only the demand characterization characteristics are taken into consideration to realize the selection of candidate models and candidate accuracy evaluation indicators that are adapted to business needs according to demand categories, but also the candidate accuracy evaluation indicators are further correlated with inventory indicators that characterize the measurement standards for downstream inventory transfers, which can be helpful in finding demand forecasting methods and accuracy evaluation indicators that are adapted to business scenarios and effectively improve inventory satisfaction rates.

[0071] In conjunction with specific embodiments, Figure 2 The method shown is further explained.

[0072] According to an embodiment of the present disclosure, the demand characterization feature may include at least one of the following: a demand discontinuity feature and a demand volatility feature.

[0073] According to an embodiment of the present disclosure, for historical order data, SKUs may be classified based on the SBC (demand classification method proposed by Syntetos-Boylan-Crosto) method, which can classify SKUs based on two dimensions: demand discontinuity and demand volatility.

[0074] According to an embodiment of the present disclosure, the demand discontinuity feature can be measured using ADI (average demand interval). ADI = the sum of non-zero demand intervals / the sum of the number of non-zero demand interval time periods. In short, ADI = sales days / total days.

[0075] According to an embodiment of the present disclosure, the volatility characteristics of demand can be measured using CV2 (squared coefficient of variation). CV2 = (demand standard deviation / demand average value) 2 .

[0076] According to an embodiment of the present disclosure, the above-mentioned operation S201 may include: in response to determining that a demand intermittency characteristic value characterizing demand intermittency is greater than or equal to a first threshold, and a demand volatility characteristic value characterizing demand volatility is less than a second threshold, determining that the business demand is an intermittent demand. In response to determining that a demand intermittency characteristic value is greater than or equal to a first threshold, and a demand volatility characteristic value is greater than or equal to a second threshold, determining that the business demand is a flash demand. In response to determining that a demand intermittency characteristic value is less than a first threshold, and a demand volatility characteristic value is less than a second threshold, determining that the business demand is a robust demand. In response to determining that a demand intermittency characteristic value is less than a first threshold, and a demand volatility characteristic value is greater than or equal to a second threshold, determining that the business demand is a fluctuating demand.

[0077] According to an embodiment of the present disclosure, SKUs can be divided into four types based on ADI and CV2. For example, if ADI>=1.32 is determined as large discontinuity, and CV2>=0.49 is determined as large volatility, the classification result can be as shown in FIG3.

[0078] Figure 4 The result diagram of SKU classification based on the SBC method according to an embodiment of the present disclosure is schematically shown.

[0079] like Figure 4 As shown, after classifying SKUs based on the SBC method, four types of demand can be obtained, including: the first type, intermittent demand, that is, the demand is highly intermittent but low in volatility, the interval between two demands is long, but the demand volume is basically stable each time. The second type is flash demand, that is, the demand is highly intermittent and volatile, the interval between two demands is long, and the change in each demand volume is large compared with the previous one. The third type is stable demand, that is, the demand is both low in intermittent and volatility, the demand occurs continuously and the difference between them is not much. The fourth type is fluctuating demand, that is, the demand is low in intermittent but highly volatile, the demand occurs continuously but the change each time is large compared with the previous one.

[0080] According to an embodiment of the present disclosure, by performing the aforementioned operation S202, a candidate model pool and a candidate accuracy evaluation index pool for different SKUs may be established.

[0081] For example, for the first type of SKUs with intermittent demand, the demand volatility is not large but there are intermittent ones. Since MAAPE, sMAPE, and MASE can handle intermittent demand problems with true values ​​of zero, these three indicators are suitable as alternative indicators for intermittent SKUs. In terms of model pool selection, since Croston, SBA, and ETS can solve the problem of intermittent prediction well, these three models are listed as the model pool for intermittent SKUs.

[0082] For example, for the second type of SKUs with flash demand, the time and amount of demand are irregular and difficult to predict. Therefore, we can only hope to accurately predict the total amount of demand within a certain period to help the downstream inventory to accurately prepare and replenish the total amount. Since sMAPEM is a modification of sMAPE, it can effectively consider the difference between the predicted total amount and the actual total amount. Therefore, sMAPE and sMAPEM are used as alternative indicators. In terms of model pool selection, since Croston and SBA can cope well with intermittent and volatile forecasts, but ETS is difficult to predict sudden changes in demand, unlike intermittent SKUs, flash SKUs only retain Croston and SBA to join the model pool.

[0083] For example, for the third type of robust SKU, MAPE and wMAPE are used as alternative indicators. These indicators are universal and can better measure the prediction accuracy of stable demand. In the selection of model pool, since the demand is stable and continuous, the amount of data is large. At this time, using a simpler model can effectively save the amount of calculation and ensure a certain accuracy, namely the four simple models of Naïve, SMA, ETS, and STL.

[0084] For example, the fourth type of volatile SKU has continuous but highly volatile demand, so both the indicators and the model need to capture the fluctuations caused by seasonality and events. sMAPE and MASE have good symmetry and can effectively avoid the problem of overestimation or underestimation. In terms of model pool selection, ETS, STL, and ARIMA can capture sales fluctuations caused by seasonality, promotions, external time, etc.

[0085] In summary, Figure 4 The candidate model pool and candidate accuracy evaluation pool selected for each type of SKU are summarized in Table 1.

[0086] Table 1:

[0087]

[0088] According to an embodiment of the present disclosure, for the above four types of SKUs, a model combination strategy can also be adopted, in which all models applicable to a certain type of SKU are added to the model pool for training, and the models in the model pool are combined to construct a new candidate model.

[0089] Through the above-mentioned embodiments of the present disclosure, business needs can be classified based on a relatively standard rule to select appropriate models and indicators to improve the accuracy of subsequent demand forecasting.

[0090] According to the embodiment of the present disclosure, the candidate accuracy evaluation index can be an index in the candidate accuracy evaluation index pool shown in Table 3 (denoted as )calculate.

[0091] For example, first, for each candidate model in at least one candidate model, based on the candidate model, the demand information of the business demand within a preset time period can be predicted to obtain a demand prediction result. Then, for each candidate accuracy evaluation indicator in at least one candidate accuracy evaluation indicator, according to the demand prediction result and the actual value of the business demand within the preset time period, the accuracy evaluation indicator value of the candidate model under the candidate accuracy evaluation indicator is calculated.

[0092] Based on the above method, the index value of each candidate accuracy evaluation index determined by each model can be calculated. The preset time period may include one day, one week, one month, etc., and may not be limited thereto.

[0093] According to an embodiment of the present disclosure, before building a regression model, an inventory index may be determined first. The method for determining the inventory index may include: for each order in the business demand, according to the ratio of the number of satisfied order lines and the total number of order lines in the preset time period, determining the order line fulfillment rate in the preset time period. According to the ratio of the total initial inventory and the total number of outbound items in the preset time period, determining the turnover days of the order. Determine the inventory index based on the order line fulfillment rate and the turnover days.

[0094] According to an embodiment of the present disclosure, the inventory index (denoted as ) can also consider the satisfaction rate (denoted as ) and turnover days (denoted as ).

[0095] The satisfaction rate can refer to the order line satisfaction rate. When a certain order line is satisfied, it is counted as 1, otherwise it is 0. The final statistical result is the ratio of all satisfied order lines to all order lines, which is the order line satisfaction rate. The calculation formula can be shown in formula (2).

[0096] Formula (2)

[0097] Among them, turnover days can refer to quantity turnover days. During the statistical period, it can be expressed by the ratio of the total inventory at the beginning of the period to the total number of items shipped out. The calculation formula can be shown in formula (3).

[0098] Formula (3)

[0099] According to the embodiments of the present disclosure, for SKU inventory, the higher the fulfillment rate, the better, indicating that the fulfillment of the inventory is better. The smaller the turnover days, the better, indicating that the goods stay in the warehouse for a short time and the inventory resources are fully utilized. Therefore, the inventory index is proportional to the fulfillment rate and inversely proportional to the turnover days. The expression of the Kuncun index can be shown as formula (4).

[0100] Formula (4)

[0101] It should be noted that the weights set for the order line fulfillment rate and turnover days in the inventory indicators can be adaptively adjusted according to the business scenario and are not limited here.

[0102] According to an embodiment of the present disclosure, after determining the candidate accuracy evaluation index and the candidate index based on the above method, a regression model can be constructed in combination with SKU attributes.

[0103] According to an embodiment of the present disclosure, the above operation S204 may include: for the candidate accuracy evaluation index in the regression model, obtaining multiple candidate accuracy evaluation index values ​​of the candidate accuracy evaluation index in multiple time windows. For the inventory index in the regression model, obtaining multiple inventory index values ​​of the inventory index in multiple time windows. Inputting the multiple candidate accuracy evaluation index values ​​and the multiple inventory index values ​​into the regression model to obtain a regression coefficient.

[0104] According to an embodiment of the present disclosure, assuming that the time window is one day, the candidate accuracy evaluation index value and inventory index value of each SKU per day can be calculated based on the above method. By obtaining the candidate accuracy evaluation index values ​​and inventory index values ​​of multiple days and inputting them into the regression model, the regression coefficient can be calculated.

[0105] It should be noted that the candidate accuracy evaluation index values ​​and inventory index values ​​for the specific days can be determined according to actual business needs. For example, they can be determined according to the total values ​​of the coefficients to be calculated in the regression model.

[0106] According to an embodiment of the present disclosure, a time window may include multiple preset time periods, and the preset time period included in the next time window overlaps with the preset time period included in the previous time window. The candidate accuracy evaluation index value of the candidate accuracy evaluation index in a time window may be characterized as the average of the candidate accuracy evaluation index values ​​of the candidate accuracy evaluation index in multiple preset time periods corresponding to the time window.

[0107] According to an embodiment of the present disclosure, the inventory index value of the inventory index in a time window may be characterized as an average of the inventory index values ​​in a plurality of preset time periods corresponding to the time window.

[0108] For example, suppose the preset time period is one day, with 7 days as a time window, and slide one day each time. If the SKU time span is 66 days, 60 test cycles can be selected for each SKU. The average of each candidate accuracy evaluation index value of the SKU in each window period, as well as the average order line fulfillment rate and turnover days of the inventory index in the window period, can be taken as the candidate accuracy evaluation index value and inventory index value in each time window.

[0109] Through the above-mentioned embodiments of the present disclosure, by calculating the candidate accuracy evaluation index value and inventory index value within each set time window length based on the calculation of the mean, the accuracy of the calculated inventory index value and accuracy evaluation index value can be effectively improved, thereby further improving the effectiveness of the determined accuracy evaluation index.

[0110] According to an embodiment of the present disclosure, in the process of executing the accuracy evaluation index determination method for demand forecasting, the method may further include: determining a target accuracy index evaluation value with the largest value according to at least one accuracy index evaluation value calculated for at least one candidate model under at least one candidate accuracy evaluation index. The candidate model associated with the target accuracy index evaluation value is determined as the target model for predicting demand information of the business demand.

[0111] According to the embodiments of the present disclosure, the performance of each candidate model in the candidate model pool under a given indicator can be calculated to determine the model accuracy. For example, if the given indicator is SMAP, the model accuracy of the corresponding model can be obtained by calculating the value of 1-SMAP, which represents the error between the predicted value and the true value. Then, the model with the highest accuracy can be determined as the optimal model.

[0112] According to an embodiment of the present disclosure, the process of determining the optimal model can be performed before performing the above operation S203. In this case, in the process of performing operation S203, only for each candidate accuracy evaluation indicator in at least one candidate accuracy evaluation indicator, the candidate accuracy evaluation indicator determined based on the optimal model is used as an independent variable, the demand attribute characteristics of the business demand are used as a control variable, and the inventory indicator of the business demand is used as a dependent variable to construct a regression model, which is conducive to reducing the amount of calculation.

[0113] According to an embodiment of the present disclosure, the process of determining the optimal model can also be performed after determining the target accuracy evaluation index by performing the above operation S205. In this case, in the process of determining the optimal model based on the above method, the above predetermined index can be set as the target accuracy evaluation index.

[0114] According to the embodiments of the present disclosure, the process of determining the optimal model and determining the target accuracy evaluation index may also be performed simultaneously, and there is no limitation on this.

[0115] Figure 5 The overall flow chart of selecting the optimal accuracy evaluation index for each type of SKU according to an embodiment of the present disclosure is schematically shown.

[0116] like Figure 5 As shown, the method includes operations S501 to S510.

[0117] In operation S501, historical order data of all SKUs is obtained.

[0118] In operation S502, each SKU is classified based on the SBC method.

[0119] In operation S503, a candidate model pool and a candidate accuracy evaluation index pool are determined according to the classification.

[0120] In operation S504, based on the candidate model pool, a value of each candidate accuracy evaluation indicator in the candidate accuracy evaluation indicator pool corresponding to each SKU is calculated.

[0121] In operation S505 , an inventory replenishment simulation is performed based on each candidate accuracy evaluation index value.

[0122] In operation S506, replenishment data of all SKUs is obtained.

[0123] In operation S507, the inventory indicators of each SKU are calculated, including the order line fulfillment rate and the inventory turnover days.

[0124] In operation S508, a time window is set, the window is slid, and the candidate accuracy evaluation index and inventory index within each time window are calculated.

[0125] In operation S509, OLS regression is performed with the inventory index as the dependent variable, the candidate accuracy evaluation index as the independent variable, and the SKU attribute as the control variable.

[0126] In operation S510, based on the size and significance of the regression coefficient, an optimal accuracy evaluation index is selected for each type of SKU.

[0127] In conjunction with specific embodiments, Figure 5 The method shown is further described.

[0128] The test was conducted based on the actual sales volume of a large automotive spare parts scenario. Taking SKU A as an example, the main data is shown in Table 2.

[0129] Table 2:

[0130]

[0131] The CV2 of this SKU is calculated to be 0.32, and the ADI is 1.65. Based on the SBC classification method, it is determined to be a SKU with intermittent demand. The distribution of all SKUs in this data set is as follows: Fig. 6A shown.

[0132] Fig. 6A The classification results of the automotive aftermarket SKUs according to the embodiment of the present disclosure are schematically shown.

[0133] like Fig. 6A As shown, in this data set, the proportion of SKUs with intermittent demand is 90.1%.

[0134] For each SKU, based on the category it belongs to, use the appropriate model. Figure 6B Results shown.

[0135] Figure 6B The OLS regression coefficient results characterizing the accuracy rate of all SKUs and inventory indicators according to an embodiment of the present disclosure are schematically shown.

[0136] By analyzing the results of the above solutions, the selected indicators are: absolute error indicator rmae, relative error indicators mape_2, smape_2 and mase. Among them, smape_2 performs relatively well in both methods. Smape_2 makes some rule adjustments based on the standard smape, so it is most appropriate to choose smape as the demand forecast evaluation indicator in this business scenario.

[0137] Through the above-mentioned embodiments of the present disclosure, in the context of the supply chain, standard processes and strategies can be provided for the selection of accuracy indicators for demand forecasting. Combined with the coordination and collaboration of upstream and downstream, based on the overall optimal perspective, demand forecasting optimization and inventory optimization can be synchronized to alleviate problems such as out-of-stock or inventory backlogs to the greatest extent.

[0138] Figure 7 A block diagram of a device for determining an accuracy evaluation index for demand forecasting according to an embodiment of the present disclosure is schematically shown.

[0139] like Figure 7 As shown, the accuracy evaluation index determination device 700 for demand prediction includes a demand category determination module 710, an adaptation information acquisition module 720, a regression model construction module 730, a regression calculation module 740 and an index determination module 750.

[0140] The demand category determination module 710 is used to determine the demand category of the business demand according to the demand characterization features characterized by the historical order data of the business demand.

[0141] The adaptation information acquisition module 720 is used to obtain at least one candidate model that is adapted to the demand category from the preset model pool according to the demand category, and to obtain at least one candidate accuracy evaluation index that is adapted to the demand category from the preset accuracy evaluation index pool, wherein the candidate accuracy evaluation index represents the accuracy of the prediction result obtained by performing demand prediction on the business demand based on the candidate model.

[0142] The regression model construction module 730 is used to construct a regression model for each candidate model in at least one candidate model and each candidate accuracy evaluation indicator in at least one candidate accuracy evaluation indicator, using the candidate accuracy evaluation indicator determined based on the candidate model as an independent variable, the demand attribute characteristics of the business demand as a control variable, and the inventory indicator of the business demand as a dependent variable, to obtain at least one regression model, wherein the inventory indicator represents the inventory performance of the orders in the business demand.

[0143] The regression calculation module 740 is used to perform regression calculation on each regression model in the at least one regression model, determine the regression coefficient determined for the candidate accuracy evaluation index in the regression model, and obtain at least one regression coefficient.

[0144] The indicator determination module 750 is used to determine a target accuracy evaluation indicator according to a target regression coefficient that meets a preset condition in at least one regression coefficient, as an accuracy evaluation indicator determined in a demand forecasting process for business needs.

[0145] According to an embodiment of the present disclosure, the demand characterization feature includes at least one of the following: a demand discontinuity feature and a demand volatility feature.

[0146] According to an embodiment of the present disclosure, the demand category determination module includes an intermittent demand determination unit, a flash demand determination unit, a robust demand determination unit, and a fluctuating demand determination unit.

[0147] The intermittent demand determination unit is used to determine that the business demand is intermittent demand in response to determining that the demand intermittency characteristic value characterizing the demand intermittency characteristic is greater than or equal to a first threshold, and the demand volatility characteristic value characterizing the demand volatility characteristic is less than a second threshold.

[0148] The flash demand determination unit is used to determine that the business demand is a flash demand in response to determining that the demand discontinuity characteristic value is greater than or equal to a first threshold and the demand volatility characteristic value is greater than or equal to a second threshold.

[0149] The robust demand determination unit is used to determine that the business demand is a robust demand in response to determining that the demand discontinuity characteristic value is less than a first threshold and the demand volatility characteristic value is less than a second threshold.

[0150] The fluctuating demand determining unit is used to determine that the service demand is a fluctuating demand in response to determining that the demand discontinuity characteristic value is less than a first threshold and the demand volatility characteristic value is greater than or equal to a second threshold.

[0151] According to an embodiment of the present disclosure, the accuracy evaluation index determination device for demand forecasting also includes an order line fulfillment rate determination module, a turnover days determination module and an inventory index determination module.

[0152] The order line fulfillment rate determination module is used to determine the order line fulfillment rate of each order in the business demand within a preset time period based on the ratio of the number of order lines that have been fulfilled within the preset time period to the total number of order lines.

[0153] The turnover days determination module is used to determine the turnover days of an order based on the ratio of the total initial inventory volume and the total number of shipped items for the order within a preset time period.

[0154] The inventory index determination module is used to determine the inventory index based on the order line fulfillment rate and turnover days.

[0155] According to an embodiment of the present disclosure, the regression calculation module includes an accuracy evaluation index value acquisition unit, an inventory index value acquisition unit and a regression coefficient acquisition unit.

[0156] The accuracy evaluation index value acquisition unit is used to acquire multiple candidate accuracy evaluation index values ​​of the candidate accuracy evaluation index in multiple time windows for the candidate accuracy evaluation index in the regression model.

[0157] The inventory index value acquisition unit is used to acquire multiple inventory index values ​​of the inventory index in multiple time windows for the inventory index in the regression model.

[0158] The regression coefficient obtaining unit is used to input a plurality of candidate accuracy evaluation index values ​​and a plurality of inventory index values ​​into the regression model to obtain the regression coefficient.

[0159] According to an embodiment of the present disclosure, the time window includes multiple preset time periods, and the preset time period included in the next time window overlaps with the preset time period included in the previous time window, and the candidate accuracy evaluation index value of the candidate accuracy evaluation index in a time window is characterized by the average of the candidate accuracy evaluation index values ​​of the candidate accuracy evaluation index in multiple preset time periods corresponding to the time window.

[0160] According to an embodiment of the present disclosure, the inventory index value of the inventory index in a time window is characterized by an average of the inventory index values ​​in a plurality of preset time periods corresponding to the time window.

[0161] According to an embodiment of the present disclosure, the regression coefficient has a significant feature, and the significant feature characterizes the influence of the candidate accuracy evaluation index on the next inventory index. The preset condition includes at least one of the following: the significant feature value characterizing the significant feature is greater than a preset threshold, and the value of the regression coefficient is the largest among the values ​​of at least one regression coefficient.

[0162] According to an embodiment of the present disclosure, the accuracy evaluation index determination device for demand forecasting also includes a demand forecasting result acquisition module and an accuracy evaluation index value calculation module.

[0163] The demand forecasting result obtaining module is used to predict the demand information of the business demand within a preset time period for each candidate model in at least one candidate model based on the candidate model to obtain the demand forecasting result.

[0164] The accuracy evaluation index value calculation module is used to calculate the accuracy evaluation index value of the candidate model under the candidate accuracy evaluation index for each candidate accuracy evaluation index of at least one candidate accuracy evaluation index according to the demand forecast result and the actual value of the business demand within a preset time period.

[0165] According to an embodiment of the present disclosure, the accuracy evaluation index determination device for demand forecasting also includes a target accuracy index evaluation value determination module and a target model determination module.

[0166] The target accuracy index evaluation value determination module is used to determine the target accuracy index evaluation value with the maximum value according to at least one accuracy index evaluation value calculated for at least one candidate model under at least one candidate accuracy evaluation index.

[0167] The target model determination module is used to determine the candidate model related to the target accuracy index evaluation value as the target model for predicting the demand information of the business demand.

[0168] According to the embodiments of the present invention, any multiple of the modules and units, or at least part of the functions of any multiple of them, can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules and units can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules and units can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or in any one of the three implementation methods of software, hardware and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules and units can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding function can be executed.

[0169] For example, any multiple of the demand category determination module 710, the adaptation information acquisition module 720, the regression model construction module 730, the regression calculation module 740, and the index determination module 750 can be combined in one module / unit for implementation, or any one of the modules / units can be split into multiple modules / units. Alternatively, at least part of the functions of one or more of these modules / units can be combined with at least part of the functions of other modules / units and implemented in one module / unit. According to an embodiment of the present disclosure, at least one of the demand category determination module 710, the adaptation information acquisition module 720, the regression model construction module 730, the regression calculation module 740, and the index determination module 750 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, at least one of the requirement category determination module 710, the adaptation information acquisition module 720, the regression model construction module 730, the regression calculation module 740 and the indicator determination module 750 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is executed.

[0170] It should be noted that the part of the device for determining the accuracy evaluation index for demand forecasting in the embodiments of the present disclosure corresponds to the part of the method for determining the accuracy evaluation index for demand forecasting in the embodiments of the present disclosure. The description of the part of the device for determining the accuracy evaluation index for demand forecasting specifically refers to the part of the method for determining the accuracy evaluation index for demand forecasting, which will not be repeated here.

[0171] Figure 8 A block diagram of an electronic device suitable for implementing a method for determining an accuracy evaluation index for demand forecasting according to an embodiment of the present disclosure is schematically shown. Figure 8 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0172] like Figure 8 As shown, the electronic device 800 according to an embodiment of the present disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 to a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a dedicated microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include an onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0173] In RAM 803, various programs and data required for the operation of electronic device 800 are stored. Processor 801, ROM 802 and RAM 803 are connected to each other via bus 804. Processor 801 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 802 and / or RAM 803. It should be noted that the program can also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0174] According to an embodiment of the present disclosure, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to the bus 804. The system 800 may further include one or more of the following components connected to the input / output (I / O) interface 805: an input portion 806 including a keyboard, a mouse, etc.; an output portion 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 808 including a hard disk, etc.; and a communication portion 809 including a network interface card such as a LAN card, a modem, etc. The communication portion 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that the computer program read therefrom is installed into the storage portion 808 as needed.

[0175] According to an embodiment of the present disclosure, the method flow according to an embodiment of the present disclosure can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the processor 801, the above-mentioned functions defined in the system of the embodiment of the present disclosure are executed. According to an embodiment of the present disclosure, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0176] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0177] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0178] For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 802 and / or the RAM 803 described above and / or one or more memories other than the ROM 802 and the RAM 803 .

[0179] The embodiments of the present disclosure also include a computer program product, which includes a computer program, and the computer program contains program code for executing the method provided by the embodiments of the present disclosure. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the method provided by the embodiments of the present disclosure.

[0180] When the computer program is executed by the processor 801, the above functions defined in the system / device of the embodiment of the present disclosure are executed. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0181] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 809, and / or installed from a removable medium 811. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0182] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0183] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments and / or claims of the present disclosure can be combined and / or combined in a variety of ways, even if such a combination or combination is not explicitly recorded in the present disclosure. In particular, without departing from the spirit and teaching of the present disclosure, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0184] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A method for determining an accuracy evaluation index for demand forecasting, comprising: Determine the demand category of the business demand according to the demand characterization features represented by the historical order data of the business demand; According to the demand category, at least one candidate model adapted to the demand category is obtained from a preset model pool, and at least one candidate accuracy evaluation index adapted to the demand category is obtained from a preset accuracy evaluation index pool, wherein the candidate accuracy evaluation index represents the accuracy of a prediction result obtained by performing demand prediction on the business demand based on the candidate model; For each candidate model in the at least one candidate model and each candidate accuracy evaluation indicator in the at least one candidate accuracy evaluation indicator, a regression model is constructed using the candidate accuracy evaluation indicator determined based on the candidate model as an independent variable, the demand attribute characteristics of the business demand as a control variable, and the inventory indicator of the business demand as a dependent variable to obtain at least one regression model, wherein the inventory indicator represents the inventory performance of the orders in the business demand; For each regression model in the at least one regression model, performing regression calculation on the regression model, determining a regression coefficient determined for a candidate accuracy evaluation indicator in the regression model, and obtaining at least one regression coefficient; and According to the target regression coefficient that meets the preset conditions in the at least one regression coefficient, a target accuracy evaluation index is determined as the accuracy evaluation index determined in the demand forecasting process for the business demand.

2. The method according to claim 1, wherein: The demand characterization feature includes at least one of the following: a demand discontinuity feature and a demand volatility feature.

3. The method according to claim 2, wherein: Determining the demand category of the business demand based on the demand characterization features characterized by the historical order data of the business demand includes: In response to determining that a demand intermittency characteristic value characterizing the demand intermittency characteristic is greater than or equal to a first threshold, and a demand volatility characteristic value characterizing the demand volatility characteristic is less than a second threshold, determining that the service demand is an intermittent demand; In response to determining that the demand discontinuity characteristic value is greater than or equal to the first threshold, and the demand volatility characteristic value is greater than or equal to the second threshold, determining that the service demand is a flash demand; In response to determining that the demand discontinuity characteristic value is less than the first threshold value, and the demand volatility characteristic value is less than the second threshold value, determining that the business demand is a robust demand; In response to determining that the demand discontinuity characteristic value is less than the first threshold and the demand volatility characteristic value is greater than or equal to the second threshold, it is determined that the business demand is a fluctuating demand.

4. The method according to claim 1, further comprising: Before building the regression model, For each order in the business demand, according to the ratio of the number of satisfied order lines of the order within the preset time period to the total number of order lines, determine the order line fulfillment rate of the order within the preset time period; Determine the turnover days of the order according to the ratio of the total initial inventory amount and the total number of shipments of the order within the preset time period; as well as The inventory index is determined according to the order line fulfillment rate and the turnover days.

5. The method according to claim 1, wherein: The step of performing regression calculation on the regression model to determine the regression coefficient for the candidate accuracy evaluation index in the regression model includes: For the candidate accuracy evaluation index in the regression model, obtaining multiple candidate accuracy evaluation index values ​​of the candidate accuracy evaluation index in multiple time windows; For the inventory index in the regression model, obtaining multiple inventory index values ​​of the inventory index in the multiple time windows; The multiple candidate accuracy evaluation index values ​​and the multiple inventory index values ​​are input into the regression model to obtain the regression coefficient.

6. The method according to claim 5, wherein: The time window includes multiple preset time periods, and the preset time period included in the next time window overlaps with the preset time period included in the previous time window. The candidate accuracy evaluation index value of the candidate accuracy evaluation index in a time window is characterized by the average of the candidate accuracy evaluation index values ​​of the candidate accuracy evaluation index in multiple preset time periods corresponding to the time window.

7. The method according to claim 6, wherein: The inventory index value of the inventory index in a time window is characterized by an average of the inventory index values ​​of the inventory index in a plurality of preset time periods corresponding to the time window.

8. The method according to claim 1, wherein: The regression coefficient has a significance feature, and the significance feature characterizes the influence of the candidate accuracy evaluation index on the next inventory index; the preset condition includes at least one of the following: the significance feature value characterizing the significance feature is greater than a preset threshold, and the value of the regression coefficient is the largest among the values ​​of the at least one regression coefficient.

9. The method according to any one of claims 1 to 8, further comprising: For each candidate model of the at least one candidate model, based on the candidate model, predicting demand information of the business demand within a preset time period to obtain a demand prediction result; as well as For each candidate accuracy evaluation indicator of the at least one candidate accuracy evaluation indicator, the accuracy evaluation indicator value of the candidate model under the candidate accuracy evaluation indicator is calculated according to the demand forecast result and the actual value of the business demand within the preset time period.

10. The method according to claim 9, further comprising: Determine a target accuracy index evaluation value with the largest value according to at least one accuracy index evaluation value calculated for the at least one candidate model under the at least one candidate accuracy evaluation index; as well as A candidate model associated with the target accuracy index evaluation value is determined as a target model for predicting demand information of the business demand.

11. A device for determining an accuracy evaluation index for demand forecasting, comprising: A demand category determination module, used to determine the demand category of the business demand according to the demand characterization features characterized by the historical order data of the business demand; An adaptation information acquisition module, used to obtain, according to the demand category, at least one candidate model adapted to the demand category from a preset model pool, and to obtain at least one candidate accuracy evaluation index adapted to the demand category from a preset accuracy evaluation index pool, wherein the candidate accuracy evaluation index represents the accuracy of a prediction result obtained by performing demand prediction on the business demand based on the candidate model; A regression model building module, for each candidate model in the at least one candidate model and each candidate accuracy evaluation indicator in the at least one candidate accuracy evaluation indicator, using the candidate accuracy evaluation indicator determined based on the candidate model as an independent variable, using the demand attribute characteristics of the business demand as a control variable, and using the inventory indicator of the business demand as a dependent variable, to build a regression model to obtain at least one regression model, wherein the inventory indicator represents the inventory performance of the orders in the business demand; A regression calculation module, configured to perform regression calculation on each regression model in the at least one regression model, determine a regression coefficient determined for a candidate accuracy evaluation index in the regression model, and obtain at least one regression coefficient; and The indicator determination module is used to determine a target accuracy evaluation indicator according to a target regression coefficient that meets a preset condition in the at least one regression coefficient, as an accuracy evaluation indicator determined in a demand forecasting process for the business demand.

12. An electronic device comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 10.

13. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method according to any one of claims 1 to 10.

14. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 10.

Images