Prediction model construction method, article flow prediction method, equipment and medium

Abstract

The embodiment of the invention discloses a prediction model construction method, an article flow prediction method, equipment and a medium. A specific embodiment of the prediction model construction method comprises the following steps: training each article flow volume prediction base model in a base model pool according to a training set included in training data corresponding to a target article to obtain each trained article flow volume prediction base model; performing verification processing on each trained article flow volume prediction base model through a verification set included in the training data to obtain each verification result; according to each verification result, generating each model weight corresponding to each trained article flow prediction base model; and according to each determined model weight, integrating each trained article flow prediction base model to obtain an integrated model as an article flow prediction model. The implementation mode is related to the intelligent supply chain, the time complexity of predicting the article flow volume and the number of failure times of the article flow volume prediction task are reduced, and the accuracy of predicting the article flow volume is improved.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technologies, and particularly to a method for constructing a prediction model, a method for predicting the flow volume of items, a device, and a medium. Background Art

[0002] With the advancement of the digitalization process of the supply chain, predicting the flow volume of items is particularly important as the first step in supply chain planning. Currently, when predicting the flow volume of items, the commonly adopted methods are: integrating different types of time series or machine learning models to predict the flow volume of items, or directly using simple models to predict the flow volume of items.

[0003] However, the inventors have found that when predicting the flow volume of items using the above methods, the following technical problems often exist: for the method of integrating different types of time series or machine learning models, the time complexity is relatively high, and when the computing resources are insufficient, the number of failures in the item flow volume prediction task is relatively large; for the method of directly using simple models, the information contained in the time series data cannot be effectively learned, resulting in relatively low accuracy of the output results.

[0004] The above information disclosed in this background art section is only used to enhance the understanding of the background of the inventive concept, and thus, it may include information that does not constitute the prior art known to those of ordinary skill in the art in this country. Summary of the Invention

[0005] This summary of the disclosure is used to introduce concepts in a brief form, and these concepts will be described in detail in the following detailed implementation section. This summary of the disclosure is not intended to identify the key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.

[0006] Some embodiments of the present disclosure propose a method for constructing an item flow volume prediction model, a method for predicting the item flow volume, a device, an electronic device, a computer-readable medium, and a computer program product to solve one or more of the technical problems mentioned in the above background art section.

[0007] In a first aspect, some embodiments of the present disclosure provide a method for constructing an article throughput prediction model. The method includes: training each article throughput prediction base model in a base model pool according to a training set included in training data corresponding to a target article, to obtain each trained article throughput prediction base model. Among them, the training data includes a sample historical article throughput sequence of the target article. The article throughput prediction base models in the base model pool correspond to model types and model parameter information. Each of the model types corresponding to each article throughput prediction base model includes an extreme learning machine and a moving average; performing a verification process on each of the trained article throughput prediction base models through a verification set included in the training data, to obtain each verification result. Among them, the verification result in each of the verification results corresponds to the article throughput prediction base model in each of the trained article throughput prediction base models; generating respective model weights corresponding to each of the trained article throughput prediction base models according to each of the verification results; integrating each of the trained article throughput prediction base models according to the determined respective model weights, to obtain an integrated model as the article throughput prediction model.

[0008] Optionally, the performing a verification process on each of the trained article throughput prediction base models through a verification set included in the training data, to obtain each verification result includes: for each trained article throughput prediction base model, generating a prediction error of the trained article throughput prediction base model through the verification set; performing a normalization process on each of the obtained prediction errors, to obtain each of the normalized prediction errors as each verification result.

[0009] Optionally, the generating respective model weights corresponding to each of the trained article throughput prediction base models according to each of the verification results includes: for each trained article throughput prediction base model, determining the reciprocal of the verification result corresponding to the trained article throughput prediction base model as the model weight.

[0010] Optionally, the generating respective model weights corresponding to each of the trained article throughput prediction base models according to each of the verification results includes: extracting features from the training set to obtain a feature vector set; selecting, from each of the trained article throughput prediction base models, an article throughput prediction base model whose corresponding verification result meets a preset optimization condition as a sample label; training a multi-classification model according to the feature vector set and the sample label; inputting the feature vector set into the multi-classification model to obtain respective probability values as respective model weights. Among them, the probability value in each of the probability values corresponds to the article throughput prediction base model in each of the trained article throughput prediction base models.

[0011] Optionally, the number of each item throughput prediction base model corresponding to the extreme learning machine in the above base model pool is the first number, and the number of each item throughput prediction base model corresponding to the moving average in the above base model pool is the second number, and the first number is greater than the second number.

[0012] In a second aspect, some embodiments of the present disclosure provide an apparatus for constructing an item throughput prediction model. The apparatus includes: a training unit configured to train each item throughput prediction base model in the base model pool according to a training set included in training data corresponding to a target item, to obtain each trained item throughput prediction base model. Wherein, the training data includes sample historical item throughput sequences of each item in the item set, and the item throughput prediction base models in the base model pool correspond to model types and model parameter information, and each model type corresponding to each item throughput prediction base model includes an extreme learning machine and a moving average; a verification unit configured to perform verification processing on each trained item throughput prediction base model through a verification set included in the training data, to obtain each verification result. Wherein, the verification result in each of the verification results corresponds to the item throughput prediction base model in each of the trained item throughput prediction base models; a generation unit configured to generate respective model weights corresponding to each of the trained item throughput prediction base models according to each of the verification results; an integration unit configured to integrate each of the trained item throughput prediction base models according to the determined respective model weights, to obtain an integrated model as the item throughput prediction model.

[0013] Optionally, the verification unit is further configured to: for each trained item throughput prediction base model, generate a prediction error of the trained item throughput prediction base model through the verification set; perform normalization processing on the obtained respective prediction errors, to obtain the respective normalized prediction errors as each verification result.

[0014] Optionally, the generation unit is further configured to: for each trained item throughput prediction base model, determine the reciprocal of the verification result corresponding to the trained item throughput prediction base model as the model weight.

[0015] Optionally, the generating unit is further configured to: perform feature extraction on the above training set to obtain a feature vector set; select, from the above respective trained item throughput prediction base models, an item throughput prediction base model whose corresponding verification result meets a preset optimization condition as a sample label; train a multi-classification model according to the above feature vector set and the above sample label; input the above feature vector set into the above multi-classification model to obtain respective probability values as respective model weights, where the probability value in the above respective probability values corresponds to the item throughput prediction base model in the above respective trained item throughput prediction base models.

[0016] Optionally, the number of respective item throughput prediction base models in the above base model pool whose corresponding model type is extreme learning machine is a first number, the number of respective item throughput prediction base models in the above base model pool whose corresponding model type is moving average is a second number, and the first number is greater than the second number. In a third aspect, some embodiments of the present disclosure provide an item throughput prediction method, the method including: obtaining a historical item throughput sequence of a target item within a preset historical time period as a target historical item throughput sequence; inputting the above target historical item throughput sequence into an item throughput prediction model pre-constructed for the above target item to obtain a predicted item throughput, where the above item throughput prediction model is constructed by the method described in any implementation manner of the first aspect, and the above predicted item throughput is the weighted sum of the output results of the respective trained item throughput prediction base models in the above item throughput prediction model, and the weighting coefficient of the output result of each trained item throughput prediction base model is the model weight corresponding to the trained item throughput prediction base model.

[0017] Optionally, the method further includes: obtaining the remaining inventory of the above target item; in response to determining that the above predicted item throughput is greater than the above remaining inventory, generating a replenishment quantity for the above target item according to the above remaining inventory and the above predicted item throughput; controlling an associated item scheduling device to schedule the above target item with the above replenishment quantity to the inventory location of the above target item.

[0018] Fourth aspect, some embodiments of the present disclosure provide an article turnover prediction device, the device includes: an acquisition unit configured to acquire a historical article turnover volume sequence of a target article within a preset historical time period as a target historical article turnover volume sequence; an input unit configured to input the above-mentioned target historical article turnover volume sequence into a pre-constructed article turnover volume prediction model corresponding to the above-mentioned target article to obtain a predicted article turnover volume, wherein the above-mentioned article turnover volume prediction model is constructed by the method described in any implementation manner of the first aspect, and the above-mentioned predicted article turnover volume is the weighted sum of the output results of each trained article turnover volume prediction base model in the above-mentioned article turnover volume prediction model, and the weighting coefficient of the output result of each trained article turnover volume prediction base model is the model weight corresponding to the trained article turnover volume prediction base model.

[0019] Optionally, the article turnover volume prediction device further includes: a remaining inventory quantity acquisition unit, a replenishment quantity generation unit, and a control unit. Among them, the remaining inventory quantity acquisition unit is configured to acquire the remaining inventory quantity of the above-mentioned target article; the replenishment quantity generation unit is configured to, in response to determining that the above-mentioned predicted article turnover volume is greater than the above-mentioned remaining inventory quantity, generate a replenishment quantity corresponding to the above-mentioned target article according to the above-mentioned remaining inventory quantity and the above-mentioned predicted article turnover volume; the control unit is configured to control an associated article scheduling device to schedule the above-mentioned target article with the above-mentioned replenishment quantity to the inventory location of the above-mentioned target article.

[0020] Fifth aspect, some embodiments of the present disclosure provide an electronic device, including: one or more processors; a storage device having stored thereon one or more programs, when the one or more programs are executed by the one or more processors, enabling the one or more processors to implement the method described in any implementation manner of the first aspect or the third aspect.

[0021] Sixth aspect, some embodiments of the present disclosure provide a computer-readable medium having stored thereon a computer program, wherein when the program is executed by a processor, it implements the method described in any implementation manner of the first aspect or the third aspect.

[0022] Seventh aspect, some embodiments of the present disclosure provide a computer program product, including a computer program, which when executed by a processor, implements the method described in any implementation manner of the first aspect or the third aspect.

[0023] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: The article turnover volume prediction model obtained by the article turnover volume prediction model construction method of some embodiments of the present disclosure reduces the time complexity and the number of failures of the article turnover volume prediction task, and improves the accuracy of the predicted article turnover volume. Specifically, the reasons for the relatively large number of failures in the article turnover volume prediction task and the relatively low accuracy of the predicted article turnover volume are as follows: The method of integrating different types of time series or machine learning models has a high time complexity. When computing resources are insufficient, the number of failures in the article turnover volume prediction task is relatively large. The method of directly using a simple model cannot effectively learn the information contained in the time series data, resulting in a relatively low accuracy of the output result. Based on this, in the article turnover volume prediction model construction method of some embodiments of the present disclosure, first, according to the training set included in the training data corresponding to the target article, each article turnover volume prediction base model in the base model pool is trained to obtain each trained article turnover volume prediction base model. Among them, the above-mentioned training data includes the sample historical article turnover volume sequence of the target article. The article turnover volume prediction base models in the above-mentioned base model pool correspond to model types and model parameter information. Each of the model types corresponding to each of the above-mentioned article turnover volume prediction base models includes an extreme learning machine and a moving average. Thus, the training set can be used to pre-train each base model in the base model pool. Then, through the validation set included in the above-mentioned training data, the above-mentioned each trained article turnover volume prediction base model is verified to obtain each verification result. Among them, the verification result in each of the above-mentioned verification results corresponds to the article turnover volume prediction base model in each of the above-mentioned trained article turnover volume prediction base models. Thus, the validation set can be used to verify the accuracy of each trained base model. Secondly, according to each of the above-mentioned verification results, each model weight corresponding to each of the above-mentioned trained article turnover volume prediction base models is generated. Thus, the accuracy verification results of each base model can be used to determine each model weight. Finally, according to the determined each model weight, the above-mentioned each trained article turnover volume prediction base model is integrated to obtain an integrated model as the article turnover volume prediction model. Thus, each base model can be integrated using each model weight. Also because the base model pool includes an extreme learning machine model and a moving average model, directly using a simple model to predict the article turnover volume can be avoided, and the extreme learning machine model can effectively learn the information contained in the time series data, improving the accuracy of the predicted article turnover volume. Also because the time complexity of the extreme learning machine is relatively low compared to time series or machine learning models, the time complexity can be reduced, and further the number of failures of the article turnover volume prediction task can be reduced. Thus, the time complexity and the number of failures of the article turnover volume prediction task are reduced, and the accuracy of the predicted article turnover volume is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In combination with the accompanying drawings and with reference to the following specific embodiments, the above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic and the elements and elements are not necessarily drawn to scale.

[0025] Figure 1 FIG. 4 is a schematic diagram of an application scenario of a method for constructing an item turnover volume prediction model according to some embodiments of the present disclosure;

[0026] Figure 2 FIG. 8 is a flowchart of some embodiments of a method for constructing an item turnover volume prediction model according to the present disclosure;

[0027] Figure 3 FIG. 12 is a flowchart of some other embodiments of a method for constructing an item turnover volume prediction model according to the present disclosure;

[0028] Figure 4 FIG. 16 is a flowchart of some embodiments of a method for predicting item turnover volume according to the present disclosure;

[0029] Figure 5 FIG. 20 is a schematic structural diagram of some embodiments of an item turnover volume prediction model construction device according to the present disclosure;

[0030] Figure 6 FIG. 24 is a schematic structural diagram of some embodiments of an item turnover volume prediction device according to the present disclosure;

[0031] Figure 7 FIG. 28 is a schematic structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. Specific Embodiments

[0032] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0033] In addition, it should be noted that for the sake of convenience of description, only the parts related to the relevant invention are shown in the drawings. Without conflict, the embodiments in the present disclosure and the features in the embodiments can be combined with each other.

[0034] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependent relationships.

[0035] It should be noted that the modification of "one" and "multiple" mentioned in this disclosure is illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise clearly specified in the context, it should be understood as "one or more".

[0036] The names of the messages or information exchanged between multiple devices in the embodiments of this disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0037] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0038] Figure 1 It is a schematic diagram of an application scenario of a method for constructing an item throughput prediction model according to some embodiments of the present disclosure.

[0039] In Figure 1 the application scenario, first, the computing device 101 can train each item throughput prediction base model in the base model pool 103 according to the training set 1021 included in the training data 102 corresponding to the target item, and obtain each trained item throughput prediction base model 104. Among them, the above training data 102 includes the sample historical item throughput sequence of the target item. The item throughput prediction base models in the above base model pool 103 correspond to model types and model parameter information. The model types corresponding to each of the above item throughput prediction base models include extreme learning machines and moving averages. Then, the computing device 101 can perform a verification process on each of the above trained item throughput prediction base models 104 through the verification set 1022 included in the above training data 102 to obtain each verification result 105. Among them, the verification result in each of the above verification results 105 corresponds to the item throughput prediction base model in each of the above trained item throughput prediction base models 104. Secondly, the computing device 101 can generate each model weight 106 corresponding to each of the above trained item throughput prediction base models 104 according to each of the above verification results 105. Finally, the computing device 101 can integrate each of the above trained item throughput prediction base models 104 according to the determined each model weight 106 to obtain an integrated model as the item throughput prediction model 107.

[0040] It should be noted that the above computing device 101 can be hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster composed of multiple servers or terminal devices, or can be implemented as a single server or a single terminal device. When the computing device is embodied as software, it can be installed in the above-listed hardware devices. It can be implemented as, for example, multiple software or software modules for providing distributed services, or can be implemented as a single software or software module. No specific limitation is made here.

[0041] It should be understood that Figure 1 the number of computing devices in

[0042] Continuing to refer to Figure 2 , a flow 200 of some embodiments of a method for constructing an article throughput prediction model according to the present disclosure is shown. The method for constructing an article throughput prediction model includes the following steps:

[0043] Step 201, training each article throughput prediction base model in the base model pool according to the training set included in the training data corresponding to the target article, to obtain each trained article throughput prediction base model.

[0044] In some embodiments, an execution subject of the method for constructing an article throughput prediction model (such as Figure 1 the computing device 101 shown) may train each article throughput prediction base model in the base model pool according to the training set included in the training data corresponding to the target article, to obtain each trained article throughput prediction base model. Among them, the above training data may be data related to the article throughput of the target article. The above training data may include the sample historical article throughput sequence of the above target article. The target article may be any article. For example, the above target article may be any accessory. The above sample historical article throughput sequence may be a set of each article throughput (such as sales volume) of the above target article within a historical time period. Each sample historical article throughput may be used as a sample for training the model. The above training data may include a training set and a test set. For example, the training data may be the sample historical article throughput sequence of the target article during the period from January to August 2022. Then the training set may be the historical article throughput sequence composed of each historical article throughput of the target article during the period from January to June 2022. The test set may be the historical article throughput sequence composed of each historical article throughput of the target article during the period from July to August 2021.

[0045] The above-mentioned basic model pool can be used to store each item turnover prediction basic model with a preset model type and model parameter information. The item turnover prediction basic model can be an untrained model for predicting the item turnover. Each model type corresponding to the above-mentioned item turnover prediction basic models includes extreme learning machine and moving average. For example, the item turnover prediction basic model with the model type of extreme learning machine can be an online recurrent extreme learning machine model. The item turnover prediction basic model with the model type of moving average can be a weighted moving average model. The model parameter information corresponding to the item turnover prediction basic model of the extreme learning machine can include the activation function type and the memory factor. The activation function type can be sigmoid or tanh. The memory factor can be a value between 0 and 1. The model parameter information corresponding to the item turnover prediction basic model of the moving average can include each historical time range and the weighted ratio corresponding to each historical time range. As an example, each item turnover prediction basic model in the basic model pool can be as shown in the following table:

[0046]

[0047] Optionally, the number of each item turnover prediction basic model corresponding to the model type of extreme learning machine in the above-mentioned basic model pool is the first number. The number of each item turnover prediction basic model corresponding to the model type of moving average in the above-mentioned basic model pool is the second number. The first number is greater than the second number. For example, the first number can be greater than twice the second number. Thus, an item turnover prediction model can be integrated with the extreme learning machine model as the main part and the moving average model as the auxiliary part. Also, because the number of extreme learning machine models is relatively large, the accuracy of the predicted item turnover can be further improved.

[0048] Step 202: Perform a verification process on each trained item turnover prediction basic model through the validation set included in the training data to obtain each verification result.

[0049] In some embodiments, the execution subject can perform a verification process on each trained item turnover prediction basic model through the validation set included in the training data to obtain each verification result. Among them, the verification result in each of the above-mentioned verification results corresponds to the item turnover prediction basic model in each of the above-mentioned trained item turnover prediction basic models. In practice, for each trained item turnover prediction basic model, the execution subject can generate the accuracy rate of the trained item turnover prediction basic model through the above-mentioned validation set. Then, the accuracy rate can be determined as the verification result corresponding to the trained item turnover prediction basic model. Thus, each verification result corresponding to each trained item turnover prediction basic model can be obtained.

[0050] Step 203: Generate respective model weights corresponding to each trained item throughput prediction base model according to each verification result.

[0051] In some embodiments, the above-mentioned execution entity may generate respective model weights corresponding to each of the above-mentioned trained item throughput prediction base models according to each of the above-mentioned verification results. In practice, for each trained item throughput prediction base model, the above-mentioned execution entity may determine the verification result corresponding to the above-mentioned trained item throughput prediction base model as the model weight corresponding to the above-mentioned trained item throughput prediction base model.

[0052] In some optional implementation manners of some embodiments, the above-mentioned execution entity may generate respective model weights corresponding to each of the above-mentioned trained item throughput prediction base models according to each of the above-mentioned verification results through the following steps, including:

[0053] First step: Extract features from the above-mentioned training set to obtain a feature vector set. In practice, the above-mentioned execution entity may use tsfresh to extract features from the above-mentioned training set to obtain a feature vector set.

[0054] Second step: Select, from each of the above-mentioned trained item throughput prediction base models, an item throughput prediction base model whose corresponding verification result meets a preset optimization condition as a sample label. Among them, the above-mentioned preset optimization condition may be that the verification result represents the best model effect. When the verification result is the accuracy rate, the largest verification result may represent the best model effect. When the verification result is the prediction error, the smallest verification result may represent the best model effect. In practice, the model identifier of the selected item throughput prediction base model may be used as the sample label. The model identifier may uniquely represent the model. For example, the model identifier may be the model name.

[0055] Third step: Train a multi-classification model according to the above-mentioned feature vector set and the above-mentioned sample label. In practice, the above-mentioned execution entity may use the above-mentioned feature vector set as feature data, the above-mentioned sample label as the true label, and the model identifiers of each item throughput prediction base model in the base model pool as optional labels to train the multi-classification model.

[0056] Fourth step: Input the above-mentioned feature vector set into the above-mentioned multi-classification model to obtain respective probability values as respective model weights. Among them, the probability value in each of the above-mentioned probability values corresponds to the item throughput prediction base model in each of the above-mentioned trained item throughput prediction base models. Each of the above-mentioned probability values may represent the probability that the feature data is applicable to the corresponding item throughput prediction base model. Thus, the feature pool of historical item throughput can be extracted to determine the weights for weighting the model, thereby further utilizing the features of historical item throughput, improving the accuracy of the model weights, and further improving the accuracy of the predicted item throughput.

[0057] Step 204: Integrate each trained item transfer volume prediction base model according to the determined respective model weights to obtain an integrated model as the item transfer volume prediction model.

[0058] In some embodiments, the above-mentioned execution entity may integrate each of the trained item transfer volume prediction base models according to the determined respective model weights to obtain an integrated model as the item transfer volume prediction model. In practice, the above-mentioned execution entity may use the above-mentioned respective model weights as weighting coefficients to weight and integrate each of the trained item transfer volume prediction base models into an integrated model. Thus, the integrated model can be determined as the constructed item transfer volume prediction model.

[0059] The above - mentioned various embodiments of the present disclosure have the following beneficial effects: The article turnover volume prediction model obtained by the article turnover volume prediction model construction method of some embodiments of the present disclosure reduces the time complexity and the number of failures of the article turnover volume prediction task, and improves the accuracy of the predicted article turnover volume. Specifically, the reasons for the relatively large number of failures in the article turnover volume prediction task and the relatively low accuracy of the predicted article turnover volume are as follows: The method of integrating different types of time - series or machine - learning models has a high time complexity. When computing resources are insufficient, the number of failures of the article turnover volume prediction task is relatively large. The method of directly using simple models cannot effectively learn the information contained in time - series data, resulting in relatively low accuracy of the output results. Based on this, in the article turnover volume prediction model construction method of some embodiments of the present disclosure, first, according to the training set included in the training data corresponding to the target article, each article turnover volume prediction base model in the base model pool is trained to obtain each trained article turnover volume prediction base model. Among them, the above - mentioned training data includes the sample historical article turnover volume sequence of the target article. The article turnover volume prediction base models in the above - mentioned base model pool correspond to model types and model parameter information. Each of the model types corresponding to each of the article turnover volume prediction base models includes extreme learning machine and moving average. Thus, the training set can be used to pre - train each base model in the base model pool. Then, through the validation set included in the above - mentioned training data, the above - mentioned trained article turnover volume prediction base models are verified to obtain each verification result. Among them, the verification result in each of the above - mentioned verification results corresponds to the article turnover volume prediction base model in each of the above - mentioned trained article turnover volume prediction base models. Thus, the validation set can be used to verify the accuracy of each trained base model. Secondly, according to each of the above - mentioned verification results, each model weight corresponding to each of the above - mentioned trained article turnover volume prediction base models is generated. Thus, the accuracy verification results of each base model can be used to determine each model weight. Finally, according to the determined model weights, the above - mentioned trained article turnover volume prediction base models are integrated to obtain an integrated model as the article turnover volume prediction model. Thus, each base model can be integrated using each model weight. Also, because the base model pool includes an extreme learning machine model and a moving average model, directly using a simple model to predict the article turnover volume can be avoided, and the extreme learning machine model can effectively learn the information contained in time - series data, improving the accuracy of the predicted article turnover volume. Also, because the time complexity of the extreme learning machine is relatively low compared to time - series or machine - learning models, the time complexity can be reduced, and further, the number of failures of the article turnover volume prediction task can be reduced. Thus, the time complexity and the number of failures of the article turnover volume prediction task are reduced, and the accuracy of the predicted article turnover volume is improved.

[0060] Further refer to Figure 3, which shows the process 300 of some other embodiments of the method for constructing an item turnover volume prediction model. The process 300 of the method for constructing an item turnover volume prediction model includes the following steps:

[0061] Step 301: Train each item turnover volume prediction base model in the base model pool according to the training set included in the training data corresponding to the target item, and obtain each trained item turnover volume prediction base model.

[0062] In some embodiments, the specific implementation of step 301 and the technical effects brought thereby can refer to Figure 2 Step 201 in the corresponding embodiments, which will not be elaborated here.

[0063] Step 302: For each trained item turnover volume prediction base model, generate the prediction error of the trained item turnover volume prediction base model through the validation set.

[0064] In some embodiments, for each trained item turnover volume prediction base model, the execution subject (such as Figure 1 the computing device 101 shown) of the item turnover volume prediction model construction method can generate the prediction error of the above-mentioned trained item turnover volume prediction base model through the above-mentioned validation set. Among them, the above-mentioned prediction error can be the mean absolute percentage error.

[0065] Step 303: Perform normalization processing on the obtained prediction errors to obtain the normalized prediction errors as the respective verification results.

[0066] In some embodiments, the above-mentioned execution subject can perform normalization processing on the obtained prediction errors to obtain the normalized prediction errors as the respective verification results. In practice, the above-mentioned execution subject can perform z-score normalization processing on the obtained prediction errors to obtain the normalized prediction errors as the respective verification results. Thus, the prediction errors can be mapped to the range of 0-1.

[0067] Step 304: For each trained item turnover volume prediction base model, determine the reciprocal of the verification result of the corresponding trained item turnover volume prediction base model as the model weight.

[0068] In some embodiments, for each trained item turnover volume prediction base model, the above-mentioned execution subject can determine the reciprocal of the verification result of the corresponding above-mentioned trained item turnover volume prediction base model as the model weight.

[0069] Step 305: Integrate each trained item turnover volume prediction base model according to the determined model weights to obtain an integrated model as the item turnover volume prediction model.

[0070] In some embodiments, for the specific implementation of step 305 and the technical effects brought thereby, reference may be made to Figure 2 step 204 in the corresponding embodiments, which will not be elaborated herein.

[0071] From Figure 3 it can be seen that compared with the description of some corresponding embodiments of Figure 2 , the process 300 of the method for constructing the article turnover volume prediction model in some corresponding embodiments reflects the steps of expanding the model weights. Thus, the solutions described in these embodiments can quickly determine the model weights of the base model. Figure 3

[0072] Figure 4 Further referring to Figure 4 , which shows the process 400 of some embodiments of the article turnover volume prediction method. The process 400 of the article turnover volume prediction method includes the following steps:

[0073] Step 401, obtaining a historical article turnover volume sequence of the target article within a preset historical time period as the target historical article turnover volume sequence.

[0074] In some embodiments, the execution subject (such as a computing device) on which the article turnover volume prediction method runs can obtain the historical article turnover volume sequence of the target article as the target historical article turnover volume sequence from the database through a wired connection method or a wireless connection method. Among them, the above historical article turnover volume sequence can be each historical article turnover volume within a preset historical time period. Here, there is no limitation on the specific setting of the preset historical time period. For example, the preset historical time period can be the historical time period of the previous week before the current time. The target article can be the same article as the Figure 2 and 3 target article in the corresponding embodiments.

[0075] It should be noted that the above wireless connection method can include but is not limited to 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other currently known or future-developed wireless connection methods.

[0076] Step 402, inputting the target historical article turnover volume sequence into the pre-constructed article turnover volume prediction model corresponding to the target article to obtain the predicted article turnover volume.

[0077] In some embodiments, the above execution subject can input the above target historical article turnover volume sequence into the pre-constructed article turnover volume prediction model to obtain the predicted article turnover volume. Among them, the above article turnover volume prediction model can be obtained through Figure 2It is constructed by the steps in the embodiments corresponding to or 3. The predicted item turnover volume mentioned above can be the weighted sum of the output results of each trained item turnover volume prediction base model in the item turnover volume prediction model. The weighting coefficient of the output result of each trained item turnover volume prediction base model is the model weight corresponding to the trained item turnover volume prediction base model.

[0078] Optionally, the execution subject may also obtain the remaining inventory of the target item. In practice, the execution subject may obtain the remaining inventory of the target item from a database. Secondly, in response to determining that the predicted item turnover volume is greater than the remaining inventory, a replenishment quantity corresponding to the target item may be generated based on the remaining inventory and the predicted item turnover volume. In practice, the execution subject may determine the shortage quantity as the difference between the predicted item turnover volume and the remaining inventory. Then, a replenishment quantity corresponding to the target item may be generated based on the shortage quantity. Specifically, the execution subject may determine the shortage quantity as the replenishment quantity. The execution subject may also determine the sum of the shortage quantity and a preset value as the replenishment quantity. Here, the determination of the preset value is not limited. Finally, the associated item scheduling device may be controlled to schedule the target item with the replenishment quantity to the inventory location of the target item. Among them, the item scheduling device may be a device for scheduling items. For example, the item scheduling device may be an automated guided vehicle. The inventory location may be the location in the warehouse where the target item is stored. Thus, automatic replenishment can be achieved through the item scheduling device.

[0079] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: Through the item turnover volume prediction method of some embodiments of the present disclosure, the time complexity and the number of failures of the item turnover volume prediction task are reduced, and the accuracy of the predicted item turnover volume is improved. The item turnover volume prediction method of some embodiments of the present disclosure adopts an item turnover volume prediction model constructed by the steps in the embodiments corresponding to or 3 to predict the item turnover volume. Also, because the base model pool includes an extreme learning machine model and a moving average model, directly using a simple model to predict the item turnover volume can be avoided, and the extreme learning machine model can effectively learn the information contained in the time series data, improving the accuracy of the predicted item turnover volume. Also, because the time complexity of the extreme learning machine is relatively low compared to time series or machine learning models, the time complexity can be reduced, and further, the number of failures of the item turnover volume prediction task can be reduced. Thus, the time complexity and the number of failures of the item turnover volume prediction task are reduced, and the accuracy of the predicted item turnover volume is improved. Figure 2 It is constructed by the steps in the embodiments corresponding to or 3. The predicted item turnover volume mentioned above can be the weighted sum of the output results of each trained item turnover volume prediction base model in the item turnover volume prediction model. The weighting coefficient of the output result of each trained item turnover volume prediction base model is the model weight corresponding to the trained item turnover volume prediction base model.

[0080] Further reference Figure 5, as an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of an apparatus for constructing an article throughput prediction model. These apparatus embodiments correspond to Figure 2 the method embodiments shown, and the apparatus can be specifically applied to various electronic devices.

[0081] As Figure 5 shown, the article throughput prediction model construction apparatus 500 of some embodiments includes: a training unit 501, a verification unit 502, a generation unit 503, and an integration unit 504. Among them, the training unit 501 is configured to train each article throughput prediction base model in the base model pool according to the training set included in the training data corresponding to the target article, and obtain each trained article throughput prediction base model. Among them, the above training data includes the sample historical article throughput sequences of each article in the article set. The article throughput prediction base models in the above base model pool correspond to model types and model parameter information. Each of the model types corresponding to each of the article throughput prediction base models includes an extreme learning machine and a moving average; the verification unit 502 is configured to perform a verification process on each of the trained article throughput prediction base models through the verification set included in the above training data, and obtain each verification result. Among them, the verification result in each of the above verification results corresponds to the article throughput prediction base model in each of the trained article throughput prediction base models; the generation unit 503 is configured to generate each model weight corresponding to each of the trained article throughput prediction base models according to each of the above verification results; the integration unit 504 is configured to integrate each of the trained article throughput prediction base models according to the determined each model weight, and obtain an integrated model as the article throughput prediction model.

[0082] Optionally, the verification unit 502 can be further configured to: for each trained article throughput prediction base model, generate a prediction error of the trained article throughput prediction base model through the above verification set; perform a normalization process on each of the obtained prediction errors, and obtain each normalized prediction error as each verification result.

[0083] Optionally, the generation unit 503 can be further configured to: for each trained article throughput prediction base model, determine the reciprocal of the verification result corresponding to the trained article throughput prediction base model as the model weight.

[0084] Optionally, the generating unit 503 may be further configured to: extract features from the above training set to obtain a feature vector set; select, from the above-mentioned various trained item throughput prediction base models, an item throughput prediction base model whose corresponding verification result meets a preset optimization condition as a sample label; train a multi-classification model according to the above feature vector set and the above sample label; input the above feature vector set into the above multi-classification model to obtain respective probability values as respective model weights, where the probability value among the above respective probability values corresponds to the item throughput prediction base model among the above various trained item throughput prediction base models.

[0085] Optionally, the number of each item throughput prediction base model corresponding to the extreme learning machine in the above base model pool is a first number, the number of each item throughput prediction base model corresponding to the moving average in the above base model pool is a second number, and the first number is greater than the second number.

[0086] It can be understood that the various units described in the apparatus 500 correspond to the respective steps in the method described with reference to Figure 2 Therefore, the operations, features, and beneficial effects described above for the method also apply to the apparatus 500 and the units included therein, and will not be elaborated herein.

[0087] Further referring to Figure 6 , as an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of an item throughput prediction apparatus. These apparatus embodiments correspond to Figure 4 the method embodiments shown, and the apparatus can be specifically applied to various electronic devices.

[0088] As shown in Figure 6 , an item throughput prediction apparatus 600 in some embodiments includes: an acquisition unit 601 and an input unit 602. Among them, the acquisition unit 601 is configured to acquire a historical item throughput sequence of a target item within a preset historical time period as a target historical item throughput sequence; the input unit 602 is configured to input the above target historical item throughput sequence into a pre-constructed item throughput prediction model corresponding to the above target item to obtain a predicted item throughput, where the above item throughput prediction model is constructed by the method embodiments shown in Figure 2 or 3, and the above predicted item throughput is the weighted sum of the output results of each trained item throughput prediction base model in the above item throughput prediction model, and the weighting coefficient of the output result of each trained item throughput prediction base model is the model weight corresponding to the trained item throughput prediction base model.

[0089] Optionally, the article turnover prediction device 600 may further include: a remaining inventory quantity acquisition unit, a replenishment quantity generation unit, and a control unit (not shown in the figure). Among them, the remaining inventory quantity acquisition unit is configured to acquire the remaining inventory quantity of the above-mentioned target article; the replenishment quantity generation unit is configured to, in response to determining that the predicted article turnover quantity is greater than the remaining inventory quantity, generate a replenishment quantity corresponding to the above-mentioned target article according to the remaining inventory quantity and the predicted article turnover quantity; the control unit is configured to control an associated article scheduling device to schedule the above-mentioned target article with the above-mentioned replenishment quantity to the inventory location of the above-mentioned target article.

[0090] It can be understood that the various units described in the device 600 correspond to the respective steps in the method described with reference to Figure 4 Therefore, the operations, features, and beneficial effects described above for the method also apply to the device 600 and the units included therein, and will not be elaborated herein.

[0091] Next, with reference to Figure 7 , which shows a schematic structural diagram of an electronic device 700 suitable for implementing some embodiments of the present disclosure. Figure 7 The electronic device shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0092] As Figure 7 shown, the electronic device 700 may include a processing device 701 (such as a central processing unit, a graphics processing unit, etc.), which may perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage device 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 are also stored. The processing device 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0093] Generally, the following devices may be connected to the I / O interface 705: an input device 706 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 707 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 708 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 709. The communication device 709 may allow the electronic device 700 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 7 shows the electronic device 700 having various devices, it should be understood that it is not required to implement or have all the shown devices. Instead, more or fewer devices may be implemented or had. Figure 7Each block shown may represent a device or, as required, multiple devices.

[0094] In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product that includes a computer program carried on a computer-readable medium, the computer program including program code for performing the methods shown in the flowcharts. In such some embodiments, the computer program can be downloaded and installed from the network via a communication device 709, or installed from a storage device 708, or installed from a ROM 702. When the computer program is executed by a processing device 701, the above functions defined in the methods of some embodiments of the present disclosure are performed.

[0095] It should be noted that the computer-readable medium described in some embodiments of the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In some embodiments of the present disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. And in some embodiments of the present disclosure, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which computer-readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable signal medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0096] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (HyperText Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks ("LAN"), wide area networks ("WAN"), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0097] The above computer-readable medium can be included in the above electronic device; or can exist separately without being assembled into the electronic device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by the electronic device, the electronic device is caused to: train each item throughput prediction base model in the base model pool according to the training set included in the training data corresponding to the target item, to obtain each trained item throughput prediction base model, where the above training data includes the sample historical item throughput sequence of the target item, the item throughput prediction base models in the base model pool correspond to model types and model parameter information, and the model types corresponding to each of the above item throughput prediction base models include extreme learning machines and moving averages; perform a verification process on each of the above trained item throughput prediction base models through the verification set included in the above training data, to obtain each verification result, where the verification result in each of the above verification results corresponds to the item throughput prediction base model in each of the above trained item throughput prediction base models; generate each model weight corresponding to each of the above trained item throughput prediction base models according to each of the above verification results; and integrate each of the above trained item throughput prediction base models according to the determined each model weight, to obtain an integrated model as the item throughput prediction model. Or cause the electronic device to: obtain the historical item throughput sequence of the target item within a preset historical time period as the target historical item throughput sequence; input the above target historical item throughput sequence into the pre-constructed item throughput prediction model corresponding to the target item, to obtain a predicted item throughput, where the above item throughput prediction model is constructed by the method described in any implementation manner of the first aspect, and the above predicted item throughput is the weighted sum of the output results of each trained item throughput prediction base model in the above item throughput prediction model, and the weighting coefficient of the output result of each trained item throughput prediction base model is the model weight corresponding to the trained item throughput prediction base model.

[0098] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., connected through the Internet using an Internet service provider).

[0099] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system for performing the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0100] The units described in some embodiments of the present disclosure may be implemented in software or in hardware. The described units may also be provided in a processor. For example, it may be described as: a processor includes a training unit, a verification unit, a generation unit, and an integration unit. Among them, the names of these units do not constitute a limitation on the unit itself in some cases. For example, the training unit may also be described as "the unit that trains each item turnover prediction base model in the base model pool according to the training set included in the training data corresponding to the target item to obtain each trained item turnover prediction base model".

[0101] The functions described above in this text can be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include: Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), Systems on Chip (SOCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0102] Some embodiments of the present disclosure also provide a computer program product, including a computer program which, when executed by a processor, implements any one of the above-described methods for constructing an item throughput prediction model or the item throughput prediction method.

[0103] The above description is only some preferred embodiments of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, technical solutions formed by mutually replacing the above features with technical features having similar functions (but not limited to) disclosed in the embodiments of the present disclosure.

Claims

1. A method for constructing an item turnover prediction model, comprising: Train each item turnover prediction base model in the base model pool according to the training set included in the training data corresponding to the target item, to obtain each trained item turnover prediction base model, where the training data includes the sample historical item turnover volume sequence of the target item, and the item turnover prediction base models in the base model pool correspond to model types and model parameter information, and the model types corresponding to the item turnover prediction base models include extreme learning machines and moving averages; Perform verification processing on each trained item turnover prediction base model through the verification set included in the training data, to obtain each verification result, where the verification result in each verification result corresponds to the item turnover prediction base model in each trained item turnover prediction base model; Generate each model weight corresponding to each trained item turnover prediction base model according to the respective verification results; Integrate each trained item turnover prediction base model according to the determined model weights, to obtain an integrated model as the item turnover prediction model.

2. The method according to claim 1, wherein, The performing verification processing on each trained item turnover prediction base model through the verification set included in the training data to obtain each verification result includes: For each trained item turnover prediction base model, generate the prediction error of the trained item turnover prediction base model through the verification set; Perform standardization processing on the obtained prediction errors to obtain the standardized prediction errors as each verification result.

3. The method according to claim 2, wherein, The generating each model weight corresponding to each trained item turnover prediction base model according to the respective verification results includes: For each trained item turnover prediction base model, determine the reciprocal of the verification result corresponding to the trained item turnover prediction base model as the model weight.

4. The method according to claim 1, wherein, The generating each model weight corresponding to each trained item turnover prediction base model according to the respective verification results includes: Extract features from the training set to obtain a feature vector set; Select, from each of the trained item turnover prediction base models, the item turnover prediction base model whose corresponding verification result meets the preset optimization condition as the sample label; Train a multi-classification model according to the feature vector set and the sample label; Input the feature vector set into the multi-classification model to obtain each probability value as each model weight, where the probability value in each probability value corresponds to the item turnover prediction base model in each trained item turnover prediction base model.

5. The method according to any one of claims 1-4, wherein, The number of item turnover prediction base models in the base model pool corresponding to the model type of extreme learning machine is the first number, and the number of item turnover prediction base models in the base model pool corresponding to the model type of moving average is the second number, and the first number is greater than the second number.

6. An item turnover prediction method, comprising: Obtain the historical item turnover volume sequence of the target item within a preset historical time period as the target historical item turnover volume sequence; Input the target historical item turnover volume sequence into an item turnover volume prediction model pre-constructed for the target item to obtain a predicted item turnover volume, where the item turnover volume prediction model is constructed by the method described in any one of claims 1-5, and the predicted item turnover volume is the weighted sum of the output results of each trained item turnover volume prediction base model in the item turnover volume prediction model, and the weighting coefficient of the output result of each trained item turnover volume prediction base model is the model weight corresponding to the trained item turnover volume prediction base model.

7. The method according to claim 6, wherein, The method further includes: Obtain the remaining inventory quantity of the target item; In response to determining that the predicted item turnover volume is greater than the remaining inventory quantity, generate a replenishment quantity for the target item according to the remaining inventory quantity and the predicted item turnover volume; Control an associated item scheduling device to schedule the target item with the replenishment quantity to the inventory location of the target item.

8. An item turnover prediction model construction device, comprising: A training unit configured to train each item turnover volume prediction base model in a base model pool according to a training set included in training data corresponding to a target item to obtain each trained item turnover volume prediction base model, where the training data includes sample historical item turnover volume sequences of each item in an item set, the item turnover volume prediction base models in the base model pool correspond to model types and model parameter information, and the respective model types corresponding to each item turnover volume prediction base model include extreme learning machines and moving averages; A verification unit configured to perform verification processing on each trained item turnover volume prediction base model through a verification set included in the training data to obtain respective verification results, where the verification results in the respective verification results correspond to the item turnover volume prediction base models in each trained item turnover volume prediction base model; A generation unit configured to generate respective model weights corresponding to each trained item turnover volume prediction base model according to the respective verification results; An integration unit configured to integrate each trained item turnover volume prediction base model according to the determined respective model weights to obtain an integrated model as the item turnover volume prediction model.

9. An item turnover prediction device, comprising: An acquisition unit configured to acquire a historical item turnover volume sequence of a target item within a preset historical time period as a target historical item turnover volume sequence; An input unit configured to input the target historical item turnover volume sequence into an item turnover volume prediction model pre-constructed for the target item to obtain a predicted item turnover volume, where the item turnover volume prediction model is constructed by the method described in any one of claims 1-5, and the predicted item turnover volume is the weighted sum of the output results of each trained item turnover volume prediction base model in the item turnover volume prediction model, and the weighting coefficient of the output result of each trained item turnover volume prediction base model is the model weight corresponding to the trained item turnover volume prediction base model.

10. An electronic device, comprising: One or more processors; A storage device having one or more programs stored thereon, 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 1-5 or 6-7.

11. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by a processor, the method according to any one of claims 1-5 or 6-7 is implemented.

12. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-5 or 6-7.

Images