Footwear oxidation age prediction method, program product, electronic device and storage medium

Abstract

The invention provides a footwear oxidation production age prediction method, a program product, electronic equipment and a storage medium. The method comprises the following steps: obtaining footwear characteristics of a shoe product; inputting the footwear features into a pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used for representing whether the shoes are oxidized or not; inputting the shoe features of the shoes with oxidation classification results to a pre-trained oxidation age prediction model to obtain oxidation age results; the oxidation age result is used for representing the time interval from the production time to the oxidation of the shoes. The automatic prediction process reduces the dependence on artificial experience, and improves the efficiency and accuracy of oxidation age identification. A staged model is adopted, shoes which are possibly oxidized are screened out at the first stage, detailed oxidation age prediction is carried out on the shoes at the second stage, the prediction process is more refined, and the oxidation age prediction accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of footwear testing, and in particular to a method for predicting the oxidation age of footwear, a program product, an electronic device and a storage medium. Background Art

[0002] Shoes exposed to the air for a long time are prone to oxidation and yellowing. Reducing the oxidation and yellowing of shoes due to long-term storage is an important part of reducing losses. Existing solutions mainly rely on manual labor to regularly check whether there are signs of oxidation in shoes that have been stored for a long time. If oxidation is about to occur or has already occurred, appropriate measures will be taken. However, this method relies on manual experience, resulting in inaccurate detection. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a method, program product, electronic device and storage medium for predicting the oxidation age of shoes, which use a machine learning algorithm to automatically analyze the impact of shoe characteristics on the oxidation age, reduce dependence on manual experience, and improve the accuracy of predicting the oxidation age.

[0004] In a first aspect, an embodiment of the present application provides a method for predicting the oxidation age of shoes, comprising: obtaining shoe features of shoes; inputting the shoe features into a pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used to characterize whether the shoes will undergo oxidation; inputting the shoe features of shoes whose oxidation classification results indicate that oxidation will occur into a pre-trained oxidation age prediction model to obtain an oxidation age result; the oxidation age result is used to characterize the time interval between the production time and the occurrence of oxidation of the shoes.

[0005] In the above implementation process, the oxidation age results are predicted through the neural network model to help managers better plan inventory and reduce losses. The automated prediction process reduces the reliance on manual experience and improves the efficiency and accuracy of oxidation age identification. In addition, a phased model is adopted. In the first phase, the shoes that may be oxidized are screened out, and in the second phase, detailed oxidation age predictions are made for these shoes, making the prediction process more refined and improving the accuracy of oxidation age predictions.

[0006] Optionally, in an embodiment of the present application, after obtaining the oxidation classification result, the method also includes: if the oxidation classification result indicates that oxidation of the shoe occurs, obtaining record information of the shoe; the record information includes the recording time and the record content corresponding to the recording time; the recording time includes at least one of the production date, the warehousing date, the inspection date or the maintenance date; the record content includes at least one of the image data, material performance characteristics or storage conditions of the shoe; based on the recording time and the record content corresponding to the recording time, adding corresponding status information to the time point in the recording time to generate a time series feature of the shoe; the status information is obtained through the record content corresponding to the recording time; the time series feature of the shoe is input into a preset time series model to obtain an oxidation trend analysis result of the shoe; the trend analysis result is used to characterize the potential oxidation degree of the shoe at different time points.

[0007] In the above implementation process, the use of the model's automated prediction process reduces the need for manual inspection and evaluation, improving efficiency. The degree and time of oxidation can be accurately predicted, which is conducive to taking timely measures according to the degree of oxidation, reducing the risk of loss caused by oxidation, and making the management of footwear more refined. The time series model can provide accurate predictions of the potential degree of oxidation, provide data support for decision-making, and increase the scientificity and accuracy of decision-making.

[0008] Optionally, in an embodiment of the present application, the shoe features of the shoes whose oxidation classification results are that they will oxidize are input into a pre-trained oxidation age prediction model to obtain an oxidation age result, including: obtaining input features based on the shoe features of the shoes whose oxidation classification results are that they will oxidize, and the oxidation trend analysis results; inputting the input features into the oxidation age prediction model to obtain the oxidation age result.

[0009] In the above implementation process, the oxidation age prediction model combines static footwear features and dynamic oxidation trend analysis results to improve the accuracy of oxidation age prediction by providing more comprehensive input information. In addition, combining different types of features can improve the generalization ability of the model, so that it can perform well in different data sets and conditions.

[0010] Optionally, in an embodiment of the present application, before inputting the footwear features into a pre-trained oxidation classification model to obtain the oxidation classification results, the method also includes: obtaining sample footwear features of sample shoes; the sample footwear features include brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, process, environmental characteristics or at least one of supply chain characteristics; adding oxidation classification labels to the sample shoes to obtain classification training samples; the oxidation classification labels include oxidation occurring or not oxidation occurring; and using the classification training samples to train the classification model to obtain an oxidation classification model.

[0011] In the above implementation process, by building an oxidation classification model, it is possible to determine whether the footwear is prone to oxidation, and to conduct further oxidation age analysis on the footwear that is prone to oxidation, thereby improving the accuracy of oxidation age prediction. After knowing whether oxidation is likely, preventive measures can be taken in advance, such as improving storage conditions, adjusting materials, or promoting products in advance.

[0012] Optionally, in an embodiment of the present application, before inputting the shoe features of the shoe products whose oxidation classification results are that oxidation will occur into a pre-trained oxidation age prediction model to obtain the oxidation age results, the method also includes: obtaining sample shoe features of the sample shoes that have undergone oxidation; the sample shoe features include at least one of brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, color matching, auxiliary color, process, environmental characteristics or supply chain characteristics; adding the sample shoe features to the oxidation age label corresponding to the sample shoe to obtain an age training sample; the oxidation age label includes the time interval from the production time to the occurrence of oxidation of the sample shoe; and training the initial model using the age training sample to obtain an oxidation age prediction model.

[0013] In the above implementation process, an oxidation age prediction model is established to predict the occurrence time of oxidation yellowing of shoes. The oxidation age prediction model is trained by using the sample shoe features of the sample shoes, which improves the problem of using the full amount of goods as training data, resulting in a decrease in model performance and insufficient accuracy in predicting oxidation age, thereby improving the prediction accuracy of age.

[0014] Optionally, in an embodiment of the present application, the sample footwear features include a feature matrix; obtaining the sample footwear features of the sample footwear that has undergone oxidation includes: obtaining original feature information of the sample footwear; preprocessing the original feature information to obtain processed footwear information; associating the processed footwear information according to the unique identifier of the sample footwear to obtain a feature matrix; a row of features in the feature matrix represents the sample footwear features of the same sample shoe.

[0015] In the above implementation process, by constructing a feature matrix, the features of different data sources are effectively integrated and aligned to form a feature set with consistent data and rich information, laying a solid foundation for model training and analysis. In addition, using the unique identifier of the sample shoe for feature alignment makes it easier to check for missing feature values ​​and improve the quality of model training.

[0016] Optionally, in an embodiment of the present application, after obtaining the oxidation age result, the method also includes: obtaining the actual age of the shoes; determining whether the shoes are high-risk oxidation shoes based on the difference between the actual age and the oxidation age result; testing the high-risk oxidation shoes to determine whether oxidation has occurred and obtain actual test results; generating supplementary data based on the actual test results, and adjusting the oxidation age prediction model using the supplementary data to obtain an updated oxidation age prediction model, so that subsequent shoes can be predicted using the updated oxidation age prediction model.

[0017] In the above implementation process, by continuously feeding back the actual test results, the model can learn more accurate oxidation age prediction and improve prediction accuracy. It can accurately identify high-risk oxidation shoes, effectively manage memory, and reduce losses.

[0018] In a second aspect, an embodiment of the present application further provides a device for predicting the oxidation age of shoes, including: a feature acquisition module, used to acquire shoe features of shoes; an oxidation classification module, used to input the shoe features into a pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used to characterize whether the shoes will undergo oxidation, and an oxidation age module, used to input the shoe features of shoes whose oxidation classification results indicate that oxidation will occur into a pre-trained oxidation age prediction model to obtain an oxidation age result; the oxidation age result is used to characterize the time interval between the production time and the occurrence of oxidation of the shoes.

[0019] In a third aspect, an embodiment of the present application further provides a computer program product, including computer program instructions, which, when executed by a processor, execute the method provided by the first aspect or any one of the implementations of the first aspect.

[0020] In a fourth aspect, an embodiment of the present application further provides an electronic device, comprising: a processor and a memory, the memory storing computer program instructions, and the computer program instructions, when executed by the processor, execute the method provided by the first aspect or any one of the implementations of the first aspect.

[0021] In a fifth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method provided by the first aspect or any one of the implementations of the first aspect is executed.

[0022] The present application provides a method, program product, electronic device and storage medium for predicting the oxidation age of shoes, and a neural network model is used to predict the oxidation age results, helping managers to better plan inventory and reduce losses. The automated prediction process reduces the reliance on manual experience and improves the efficiency and accuracy of oxidation age identification. A phased model is used, in which the first phase screens out shoes that may be oxidized, and in the second phase, detailed oxidation age predictions are made for these shoes, making the prediction process more refined and improving the accuracy of oxidation age predictions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A schematic diagram of a flow chart of a method for predicting the oxidation age of footwear provided in an embodiment of the present application;

[0025] Figure 2 A schematic diagram of the structure of a device for predicting the oxidation age of shoes provided in an embodiment of the present application;

[0026] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following embodiments of the technical solution of the present application are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0029] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0030] Shoes that have been stored for too long will oxidize and turn yellow, which will lead to asset loss and waste of resources. The existing solution mainly relies on manual work, which is to regularly check whether the shoes that have been stored for a long time have signs of oxidation. If oxidation is about to occur or has already occurred, appropriate measures will be taken. However, this method cannot predict when yellowing will occur in advance. Yellowing may have already occurred when manual inspection is carried out, resulting in losses.

[0031] In order to predict when yellowing will occur, experienced testers will summarize some patterns that are prone to oxidation yellowing. For example, white mesh shoes are prone to yellowing, and the oxidation age is about 2-3 years. Through these summarized patterns, the shoe inventory that meets these patterns is matched from the warehouse, and the corresponding oxidation age of each shoe is given. For example, the oxidation age of white mesh shoes is marked as 2-3 years, so that the oxidation age of different types of shoes can be obtained.

[0032] However, the above method still relies heavily on manual experience. The model summarized by the manual model may not be comprehensive, and it is difficult to match some shoes with complex styles or complex ingredients with simple models, and thus the oxidation age cannot be calculated, or the calculation of the oxidation age is inaccurate.

[0033] In response to the above problems, the embodiments of the present application provide a method, program product, electronic device and storage medium for predicting the oxidation age of shoes. The oxidation age results provided by the neural network model for shoes can help managers better plan inventory and reduce losses. The automated prediction process reduces the reliance on manual experience and improves the efficiency and accuracy of age recognition. A phased model is adopted. In the first phase, shoes that may be oxidized are screened out, and in the second phase, detailed oxidation age predictions are made for these shoes, making the prediction process more refined and improving the accuracy of oxidation age predictions.

[0034] See also Figure 1 The flowchart of a method for predicting the oxidation age of shoes provided in an embodiment of the present application is shown. The method for predicting the oxidation age of shoes provided in an embodiment of the present application can be applied to electronic devices, which may include physical devices such as servers, PCs, tablet computers, or smart phones, or virtual devices such as virtual machines or containers. The electronic device may be a single device, or a combination of multiple devices or a cluster of a large number of devices. The method for predicting the oxidation age of shoes may include:

[0035] Step S110: Acquire shoe characteristics of the shoe product.

[0036] Step S120: inputting the shoe features into a pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used to characterize whether the shoe will be oxidized.

[0037] Step S130: Input the shoe features of the shoes classified as those that will oxidize into a pre-trained oxidation age prediction model to obtain an oxidation age result; the oxidation age result is used to characterize the time interval from the production time to the occurrence of oxidation of the shoes.

[0038] In step S110, detailed information of the footwear is collected, including at least one of brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, and craftsmanship; and footwear features of the footwear are constructed based on the collected data. The footwear features of the footwear may include text features (such as brand, material description), numerical features (such as price, material hardness), classification features (such as color coding), etc.

[0039] As an implementation method, considering that the factors affecting the oxidation age of shoes also include the environment in which the shoes are located and the shoe supply chain, the shoe characteristics of the shoes can be environmental characteristics extracted based on the environmental information of the environment in which the shoes are located, and supply chain characteristics extracted based on the supply chain information of the shoes; wherein the environmental characteristics include at least one of temperature, humidity, storage conditions or lighting conditions; and the supply chain characteristics include at least one of material source, production batch, and transportation.

[0040] Among them, brand, shoe category, applicable population, and functionality reflect the design purpose and target market of the shoe, and may use more oxidation-resistant materials or coatings. Price may indirectly reflect the quality of materials and production processes used to make shoes. Shoes with higher prices may use more oxidation-resistant materials or more complex anti-oxidation treatments. The chemical composition and structure of the material directly affect its antioxidant properties. For example, some rubber and plastic materials are more susceptible to oxidation, resulting in yellowing and hardening. Color may affect the oxidation rate of shoes, as some dyes are more likely to fade in the sun, which indirectly affects the oxidation rate of materials. Production processes, such as vulcanization, injection molding, or bonding, may affect the sealing of shoes and the durability of the protective layer, thereby affecting the oxidation rate. Environmental factors have a significant impact on the oxidation rate. High temperature and high humidity environments accelerate the oxidation process. Storage conditions, such as storage away from light, can slow down oxidation. Supply chain characteristics can reflect the environmental exposure of shoes during production and transportation. For example, the production environment in certain regions may be more susceptible to material oxidation, or environmental factors during transportation may also have an impact on oxidation.

[0041] Footwear characteristics are important factors that affect the yellowing of shoes due to oxidation. By analyzing these characteristics, we can better understand and predict the oxidation behavior of shoes, so that appropriate measures can be taken to slow down the oxidation process.

[0042] Of course, it is understandable that the shoe features of the shoes acquired in step S110 should be consistent with the sample shoe features of the sample shoes collected during the training of the oxidation classification model.

[0043] In step S120, for an optional solution, due to reasons such as the material and color of the shoes, not every type of shoe will be oxidized and yellowed, and in the actual processing process, the oxidized and yellowed products only account for a small part of the total products, that is, the ratio of oxidized and yellowed products is unbalanced. Of course, in other solutions, there may be a small difference in the ratio of yellowed and non-yellowing shoes, or even a situation where the ratio of yellowed shoes exceeds that of non-yellowing shoes.

[0044] If the oxidation age prediction model is directly trained using all products (including shoes that will oxidize and those that will not) as training data, then for the training data that will oxidize and yellow, the oxidation age label is the actual oxidation age (for example, 1 year, 2 years, etc.); but for the training data that will not oxidize and yellow, the oxidation age label may be close to positive infinity, or a preset value (should be large enough). However, assigning a fixed, very large label value to products that will not oxidize may introduce noise, because this very large label value does not represent the actual oxidation age, so the model may mistake this noise for valid information, resulting in learning the wrong pattern. In the subsequent prediction process, an oxidation age between the actual label value and the very large label value (for example, 100 years) may appear, and such an oxidation age result is not of reference significance.

[0045] And because the number of products that will not oxidize and turn yellow may be much larger than those that will oxidize, this will cause the model to be biased towards the majority class. The model may overfit samples assigned large label values ​​(shoes that will not oxidize) and fail to effectively learn and predict the minority class (i.e., shoes that will oxidize), resulting in decreased model performance and insufficient accuracy in predicting oxidation age.

[0046] Furthermore, if only the shoes that will oxidize among all the goods are used as training data to train the oxidation age prediction model, then although the oxidation age labels of the training data are relatively uniform, they are all real oxidation ages (for example, 1 year, 2 years, etc.). However, in the process of predicting the oxidation age of shoes, since it is not known in advance whether each shoe will oxidize, it is necessary to input the shoe characteristics of each shoe into the oxidation age prediction model to obtain the oxidation age result. Then, for shoes that will not oxidize, a numerical value of oxidation age will also be predicted, which is obviously inaccurate. This method will mistakenly mark shoes that will not oxidize as those that will oxidize within a preset time, resulting in uncontrolled inventory handling.

[0047] In summary, if the shoe characteristics of the shoe product are directly input into the oxidation age prediction model to obtain the corresponding oxidation age results, no matter whether the oxidation age prediction model uses all products as training data or uses only shoes that will undergo oxidation as training data, inaccurate predictions will occur.

[0048] Therefore, before identifying the oxidation age of the shoes, the embodiment of the present application first uses the oxidation classification model to identify whether the shoes will be oxidized. The shoe features of the shoes that are classified as shoes that will be oxidized are input into the pre-trained oxidation age prediction model to further predict the oxidation age results. In this way, the label setting of the oxidation age prediction model is more scientific (there will be no values ​​close to positive infinity or very large values), which improves the performance of the model; at the same time, it effectively reduces the situation where the oxidation age of shoes that will not be oxidized is predicted.

[0049] The following is an introduction to the oxidation classification model: The oxidation classification model can be built by selecting a machine learning model suitable for the classification task, such as logistic regression, random forest, support vector machine or neural network, and then the machine learning model is trained using the sample shoe features of the sample shoes to obtain a trained oxidation classification model. The detailed training process will be described later.

[0050] After the oxidation classification model is trained, the footwear features obtained in step S110 are input into the pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used to characterize whether the footwear will oxidize; for example, the oxidation classification result may be a probability value or a category label.

[0051] In the embodiment of the present application, oxidation refers to the phenomenon of color change (e.g., yellowing, darkening), material hardening, or surface cracks of the shoes. If the oxidation classification result is that oxidation will occur, it means that the shoes belong to the category of oxidation, but oxidation has not necessarily occurred. That is, the oxidation classification result is a prediction of whether oxidation will occur in the future. Of course, if the shoes have already been oxidized, the oxidation classification result is also that oxidation will occur.

[0052] In step S130, first, based on the oxidation classification result, it is determined that the oxidation classification result is a shoe that will oxidize, and the shoe characteristics of the shoe are obtained, and the characteristics are input into a pre-trained oxidation age prediction model to obtain an oxidation age result. The oxidation age result refers to an oxidation phenomenon that can be observed or detected. For example, if the oxidation age result is 2 years, then after 2 years, there is a certain probability that the shoe will turn yellow and hard.

[0053] In the above implementation process, the oxidation age results are predicted through the neural network model to help managers better plan inventory and reduce losses. The automated prediction process reduces the reliance on manual experience and improves the efficiency and accuracy of oxidation age identification. A phased model is adopted. In the first phase, the shoes that may be oxidized are screened out, and in the second phase, detailed oxidation age predictions are made for these shoes, making the prediction process more refined and improving the accuracy of oxidation age predictions.

[0054] Optionally, in the embodiment of the present application, as mentioned above, the oxidation age result refers to the observable oxidation phenomenon, because in the process of training the oxidation age prediction model, the label of the training data is the time when oxidation yellowing is observed or detected. For example, if the shoes turn yellow 3 years after the production date of the shoes, the oxidation age label of the shoes is set to 3 years.

[0055] But in fact, within 3 years, for example, after 2 years, the shoes may have already undergone some potential oxidation reactions. As time goes by, these potential oxidation levels gradually increase until yellowing occurs after 3 years. The potential oxidation levels at these different time points are not easy to observe. If the potential oxidation levels of the shoes at different time points can be mastered, it will be very effective for inventory management. For example, the inflection point of increased oxidation (sudden increase in oxidation rate) can be determined based on the potential oxidation levels of the shoes at different time points, and timely measures can be taken before the inflection point arrives to avoid losses. Knowing the potential oxidation levels of the shoes at different time points can enable more refined management of the shoes based on the oxidation age. The implementation process of this embodiment is described below.

[0056] After obtaining the oxidation classification result, the method further includes: if the oxidation classification result indicates that the shoe is oxidized, obtaining the record information of the shoe; the record information includes the record time and the record content corresponding to the record time; the record time includes at least one of the production date, storage date, inspection date or maintenance date; the record content includes at least one of the image data of the shoe, material performance characteristics or storage conditions. The inspection date can be fixed or random, and each record time has corresponding record content.

[0057] For example, when shoes are put into storage, the date of entry, photos of the shoes, material performance characteristics of the first softness, and storage conditions of the first warehouse (temperature, humidity and other environmental information are also recorded) can be recorded. During a certain inspection process, the inspection date is recorded, and image data at the time of the inspection is taken, and the material performance characteristics are recorded as soft (or slightly hard), and the storage conditions are the second warehouse (temperature, humidity and other environmental information are also recorded). By integrating this information, the recorded content of the shoes at each time point can be obtained.

[0058] Based on the recording time and the recording content corresponding to the recording time, the corresponding state information is added to the time point in the recording time to generate the shoe product time series feature; the state information is obtained through the recording content corresponding to the recording time.

[0059] For example, select key time points in the life cycle of shoes, such as production date, warehousing date, inspection date, maintenance date, etc. Create a timestamp for each key time point, and record multiple time points in time to form a time axis of time series features. The state information is generated by: using image processing technology (such as color analysis, texture analysis) to extract signs of oxidation yellowing from image data; and / or, calculating material performance change indicators, such as hardness change rate, elastic modulus change, etc.; and / or, analyzing changes in storage conditions, such as fluctuations in temperature and humidity, and the impact of these changes on the oxidation of shoes.

[0060] According to the state information corresponding to the time point, a corresponding feature vector is constructed for each time point. The state information includes at least one of the signs of oxidation and yellowing, material performance indicators, and storage conditions extracted from the image data. For example, for numerical state information (such as humidity, temperature, and material performance indicators), it is directly extracted as a feature vector of the time point. For non-numerical information such as image data or text descriptions, image processing or natural language processing technology is used to extract features, such as using image recognition technology to extract color change and texture change features, or using text analysis to extract keywords, which are used as feature vectors corresponding to the time point. If the features are classified, one-hot encoding or label encoding can be performed.

[0061] The feature vectors of all time points are arranged in chronological order to form the shoe time series features corresponding to the shoe.

[0062] Input the time series characteristics of the shoes into the preset time series model to obtain the oxidation trend analysis results of the shoes; the trend analysis results are used to characterize the potential oxidation degree of the shoes at different time points. The potential oxidation degree refers to the oxidation percentage, oxidation rate or oxidation category (such as high, medium or low oxidation degree) of the shoes in the future. Assuming that the oxidation age of a shoe is 3 years, the time series model can predict the oxidation degree at each time point from now to the next 3 years. The time series model may predict: the oxidation degree will be 5% after 10 months (for example, the material hardness index will change slightly); the oxidation degree will be 10% after 1 year; until 3 years later, directly observable oxidation phenomena will appear.

[0063] Training process of time series model: Select appropriate time series original analysis model, such as ARIMA, exponential smoothing, machine learning model, etc. Use historical time series data to build historical time series features, and then use historical time series features to train the time series original analysis model to obtain the time series model. The method of building historical time series features is as mentioned above. The time series model can effectively capture the trend of the oxidation state of shoes over time and provide accurate prediction of the potential oxidation degree.

[0064] As an implementation method, the potential oxidation degree of the footwear at different time points may be plotted as a graph, such as a line graph, a bar graph, etc., and the user may visually observe an upward trend, a downward trend or a cyclical pattern through the graph.

[0065] In the above implementation process, the automated prediction process of the model reduces the process of manual inspection and evaluation, and improves efficiency. It can also accurately predict the degree and time of oxidation, which is conducive to taking timely measures according to the degree of oxidation, reducing the risk of loss caused by oxidation, and making the management of footwear more refined. The time series model can provide accurate predictions of the potential degree of oxidation, provide data support for decision-making, and increase the scientificity and accuracy of decision-making.

[0066] Optionally, in the embodiment of the present application, the shoe features of the shoe that is classified as likely to be oxidized are input into a pre-trained oxidation age prediction model to obtain an oxidation age result, including:

[0067] According to the oxidation classification result, the shoe characteristics of the shoe that will be oxidized, and the oxidation trend analysis result, the input features are obtained. The time series data can be converted into a set of features, such as the maximum value, minimum value, average value, and change rate of the oxidation degree. Then, the converted features and the shoe characteristics of the shoe are spliced ​​to obtain the input features. Of course, the input features can also be processed by feature encoding or feature scaling.

[0068] Inputting the input features into the oxidation age prediction model to obtain the oxidation age result. It can be understood that the oxidation age prediction model in this embodiment is also obtained by training the input features composed of the sample shoe features of the sample shoes and the oxidation trend analysis results of the sample shoes.

[0069] In the implementation process of the above embodiment: the oxidation age prediction model combines static footwear features and dynamic oxidation trend analysis results, and improves the accuracy of oxidation age prediction by providing more comprehensive input information. In addition, combining different types of features can improve the generalization ability of the model, so that it can perform well in different data sets and conditions.

[0070] Optionally, in an embodiment of the present application, before inputting the footwear features into a pre-trained oxidation classification model to obtain the oxidation classification results, the method further includes: obtaining sample footwear features of sample footwear; the sample footwear features include at least one of brand, footwear category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, process, environmental features, or supply chain features. Features that may be helpful for oxidation classification may be selected. These features may directly affect the oxidation rate or possibility of footwear.

[0071] The collected feature data is then cleaned and formatted, missing values ​​and outliers are processed, and non-numeric data is encoded, such as using one-hot encoding or label encoding.

[0072] Add oxidation classification labels to sample shoes to obtain classification training samples; the oxidation classification labels include oxidation occurring or not occurring; the oxidation classification labels can be based on actual oxidation conditions in historical data or obtained through detection tests.

[0073] The classification model is trained using the classification training samples to obtain an oxidation classification model. First, a suitable machine learning algorithm is selected, such as logistic regression, random forest, support vector machine, or neural network. The model is trained using the classification training samples (including footwear features and oxidation classification labels). The training process includes adjusting model parameters, iterative learning of training data, and optimization of model performance.

[0074] You can also evaluate the performance of the model through cross-validation, independent test sets, or other cross-validation methods to improve the generalization ability of the model.

[0075] In the implementation process of the above embodiment: by constructing an oxidation classification model, it is possible to determine whether the footwear is prone to oxidation, and further analyze the oxidation age of the footwear that is prone to oxidation, thereby improving the accuracy of oxidation age prediction. After knowing whether oxidation is likely, preventive measures can be taken in advance, such as improving storage conditions, adjusting materials, or promoting products in advance.

[0076] Optionally, in the embodiment of the present application, before inputting the shoe features of the shoe that is classified as likely to be oxidized into a pre-trained oxidation age prediction model to obtain the oxidation age result, the method further includes:

[0077] Obtain sample footwear features of sample footwear that has undergone oxidation; the sample footwear features include at least one of brand, footwear category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, process, environmental features, or supply chain features. Collect detailed information on footwear that has been marked as oxidized from an oxidation classification model, database, or data warehouse.

[0078] The sample shoe features are added with the oxidation age label corresponding to the sample shoe to obtain the age training sample; the oxidation age label includes the time interval between the production time and the oxidation of the sample shoe. By combining the sample shoe features with the oxidation age label, the training sample for training the oxidation age prediction model is obtained.

[0079] The initial model was trained using birth age training samples to obtain the oxidation birth age prediction model.

[0080] In the implementation process of the above embodiment: an oxidation age prediction model is established to predict the occurrence time of oxidation yellowing of shoes. The oxidation age prediction model is trained by using the sample shoe features of the sample shoes, thereby improving the problem of using all the goods as training data, which causes the model performance to decline and the prediction accuracy of oxidation age to be insufficient, and improving the prediction accuracy of age.

[0081] Optionally, in the embodiment of the present application, the sample footwear features include a feature matrix; obtaining the sample footwear features of the sample footwear that has been oxidized includes:

[0082] The original characteristic information of the sample shoes is obtained; the original characteristic information is preprocessed to obtain the processed shoe information. The original characteristic information of the sample shoes is obtained by a database, a data warehouse or other data sources, and the data from different sources are integrated together.

[0083] Preprocessing includes at least one of data cleaning, feature encoding, feature scaling, and feature selection. Data cleaning: Check for missing values ​​and outliers in the data and process them. Missing values ​​can be processed by interpolation, deletion, or filling; outliers can be corrected or deleted. Feature encoding: Encode categorical features, such as using One-Hot Encoding or Label Encoding. Feature scaling: Standardize or normalize numerical features to eliminate the impact of different values, making model training more stable and efficient. Feature selection: Select the most influential features through correlation analysis, feature importance evaluation, and other methods to remove redundant or irrelevant features.

[0084] According to the unique identifier of the sample shoe, the processed shoe information is associated to obtain a feature matrix; a row of features in the feature matrix represents the sample shoe features of the same sample shoe.

[0085] The processed feature information is aligned using the unique identifier of the sample shoe (such as product ID, model or serial number, etc.), and the aligned feature information is organized into a feature matrix, where each row represents a sample shoe feature of a sample shoe and each column represents a feature.

[0086] You can verify whether the data in the feature matrix is ​​consistent with the processed shoe information to verify consistency. You can also decide whether to delete the corresponding sample rows or fill them with the mean, median, or other methods for missing values ​​found during the construction process.

[0087] The constructed feature matrix is ​​then imported into the machine learning model training process to train the oxidation age prediction model.

[0088] In the implementation process of the above embodiment: by constructing a feature matrix, the features of different data sources are effectively integrated and aligned to form a feature set with consistent data and rich information, laying a solid foundation for model training and analysis. In addition, the unique identification of the sample shoes is used for feature alignment, which makes it easier to check for missing feature values ​​and improves the quality of model training.

[0089] Optionally, in the embodiment of the present application, after obtaining the oxidation age result, the method further includes:

[0090] Get the actual production age of the shoes; for example, the actual production date of each pair of shoes can be collected from production records, sales records, and inventory management systems. For each pair of shoes, calculate the time interval from the production date to the current date as the actual production age.

[0091] Determine whether the shoe is a high-risk oxidized shoe based on the difference between the actual age and the oxidized age. Determine a time difference threshold, and shoes that exceed this time difference threshold are considered high-risk oxidized shoes; the time difference threshold can be set according to actual needs. Compare the oxidized age predicted by the model with the actual age and calculate the difference; if the difference between the actual age and the predicted age exceeds the set threshold, the shoe is marked as a high-risk oxidized shoe. High-risk oxidized shoes refer to footwear products with a higher risk of oxidation reactions.

[0092] Test high-risk oxidative footwear to determine whether it has oxidized and obtain actual test results. Through laboratory testing or on-site inspection, determine whether these high-risk oxidative footwear have oxidized and obtain actual test results. The actual test results can be that the high-risk oxidative footwear has oxidized or that the high-risk oxidative footwear has not oxidized.

[0093] Supplementary data is generated based on the actual test results, and the oxidation age prediction model is adjusted using the supplementary data to obtain an updated oxidation age prediction model, so that subsequent footwear products can be predicted using the updated oxidation age prediction model.

[0094] Retrain or fine-tune the existing oxidation age prediction model using supplementary data to update model parameters. Evaluate the updated model performance through cross-validation or independent test sets to ensure that the model adjustments are effective.

[0095] In an optional embodiment, after the model is deployed, its performance in actual applications is continuously monitored. Appropriate evaluation indicators such as accuracy, recall, F1 score or mean square error are used to measure model performance. A data collection and management mechanism can be established to ensure that new data can be effectively collected and integrated into existing data sets. This may include regularly acquiring data from different data sources, as well as cleaning and preprocessing new data. The above-mentioned supplementary data can also be regarded as new data.

[0096] Based on the new data, you can make a plan for model retraining, including the frequency and trigger conditions of retraining. For example, you can set the model to be retrained every six months or every year, or trigger retraining when the amount of new data reaches a certain level.

[0097] Retrain the model using the latest dataset. This may include using the same model architecture or adjusting the model architecture based on the latest data characteristics. During the retraining process, it may be necessary to reselect features, adjust model parameters, or optimize the model structure.

[0098] Deploy the retrained model with improved performance to the production environment to replace the old model. Ensure a smooth transition of the model to avoid affecting existing business processes during the deployment process.

[0099] As the data set continues to expand and update, the model can learn more information and the latest trends, thereby improving the accuracy of predictions. Regularly updating the model ensures that the model adapts to new data trends and patterns and avoids performance degradation caused by outdated data.

[0100] In the implementation process of the above embodiment: by continuously feeding back the actual test results, the model can learn more accurate oxidation age prediction, improve prediction accuracy, accurately identify high-risk oxidation shoes, effectively manage memory, and reduce losses.

[0101] See also Figure 2 The schematic diagram of the structure of the device for predicting the oxidation age of shoes provided in the embodiment of the present application is shown; the embodiment of the present application provides a device 200 for predicting the oxidation age of shoes, comprising:

[0102] A feature acquisition module 210 is used to acquire footwear features of footwear;

[0103] The oxidation classification module 220 is used to input the footwear features into a pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used to characterize whether the footwear will be oxidized.

[0104] The oxidation age module 230 is used to input the shoe features of the shoes classified as those that will oxidize into a pre-trained oxidation age prediction model to obtain an oxidation age result; the oxidation age result is used to characterize the time interval from the production time to the occurrence of oxidation of the shoes.

[0105] Optionally, in an embodiment of the present application, the shoe oxidation age prediction device 200 and the oxidation degree analysis module are used to obtain the record information of the shoe if the oxidation classification result indicates that the shoe has been oxidized; the record information includes the record time and the record content corresponding to the record time; the record time includes at least one of the production date, the warehousing date, the inspection date or the maintenance date; the record content includes at least one of the image data, material performance characteristics or storage conditions of the shoe; based on the record time and the record content corresponding to the record time, the corresponding status information is added to the time point in the record time to generate the shoe time series characteristics; the status information is obtained through the record content corresponding to the record time; the shoe time series characteristics are input into a preset time series model to obtain the oxidation trend analysis results of the shoe; the trend analysis results are used to characterize the potential oxidation degree of the shoe at different time points.

[0106] Optionally, in an embodiment of the present application, the shoe oxidation age prediction device 200 and the oxidation age module 230 are specifically used to obtain input features based on the oxidation classification results of shoe features that may undergo oxidation, and the oxidation trend analysis results; and input the input features into the oxidation age prediction model to obtain the oxidation age results.

[0107] Optionally, in an embodiment of the present application, the shoe oxidation age prediction device 200 and the oxidation classification model training module are used to obtain sample shoe characteristics of sample shoes; the sample shoe characteristics include brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, process, environmental characteristics or at least one of supply chain characteristics; oxidation classification labels are added to the sample shoes to obtain classification training samples; the oxidation classification labels include oxidation occurring or not oxidation occurring; the classification model is trained using the classification training samples to obtain an oxidation classification model.

[0108] Optionally, in an embodiment of the present application, the shoe oxidation age prediction device 200 and the oxidation age model training module are used to obtain sample shoe features of sample shoes that have undergone oxidation; the sample shoe features include brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, process, environmental characteristics or at least one of supply chain characteristics; the sample shoe features are added with the oxidation age label corresponding to the sample shoe to obtain an age training sample; the oxidation age label includes the time interval from the production time to the occurrence of oxidation of the sample shoe; the initial model is trained using the age training sample to obtain an oxidation age prediction model.

[0109] Optionally, in an embodiment of the present application, in the shoe oxidation age prediction device 200, the sample shoe features include a feature matrix; the oxidation age model training module is also used to obtain original feature information of the sample shoes; the original feature information is preprocessed to obtain processed shoe information; according to the unique identifier of the sample shoes, the processed shoe information is associated to obtain a feature matrix; a row of features in the feature matrix represents the sample shoe features of the same sample shoe.

[0110] Optionally, in an embodiment of the present application, the shoe oxidation age prediction device 200, the model adjustment module, is used to obtain the actual age of the shoes; determine whether the shoes are high-risk oxidation shoes based on the difference between the actual age and the oxidation age result; test the high-risk oxidation shoes to determine whether they are oxidized and obtain actual test results; generate supplementary data based on the actual test results, and use the supplementary data to adjust the oxidation age prediction model to obtain an updated oxidation age prediction model, so that subsequent shoes can be predicted using the updated oxidation age prediction model.

[0111] It should be understood that the device corresponds to the above-mentioned footwear oxidation age prediction method embodiment and can execute the various steps involved in the above-mentioned method embodiment. The specific functions of the device can be found in the above description. To avoid repetition, the detailed description is appropriately omitted here. The device includes at least one software function module that can be stored in a memory in the form of software or firmware or fixed in the operating system (OS) of the device.

[0112] See also Figure 3 The electronic device 300 provided in the embodiment of the present application includes: a processor 310 and a memory 320, wherein the memory 320 stores machine-readable instructions executable by the processor 310, and when the machine-readable instructions are executed by the processor 310, the above method is executed.

[0113] Figure 3 Each component shown in can be implemented by hardware, software or a combination thereof. The electronic device 300 may be a physical device, such as a server, a PC, etc., or a virtual device, such as a virtual machine, a virtualized container, etc. Moreover, the electronic device 300 is not limited to a single device, but may also be a combination of multiple devices or a cluster consisting of a large number of devices.

[0114] An embodiment of the present application further provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above method is executed.

[0115] Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0116] In several embodiments provided by the embodiments of the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to the multiple embodiments of the embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and a part of the module, program segment or 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 can also occur in a different order than the order marked in the accompanying drawings. For example, two consecutive boxes 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 and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0117] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0118] The above description is only an optional implementation manner of the embodiments of the present application, but the protection scope of the embodiments of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed in the embodiments of the present application, which should be covered within the protection scope of the embodiments of the present application.

Claims

1. A method for predicting the oxidation age of footwear, characterized in that: include: Obtaining footwear characteristics of footwear; Inputting the footwear features into a pre-trained oxidation classification model to obtain an oxidation classification result; the oxidation classification result is used to characterize whether the footwear will be oxidized; The shoe features of the shoe whose oxidation classification result is that it will oxidize are input into a pre-trained oxidation age prediction model to obtain an oxidation age result; the oxidation age result is used to characterize the time interval between the production time of the shoe and the occurrence of oxidation.

2. The method according to claim 1, characterized in that After obtaining the oxidation classification result, the method further includes: If the oxidation classification result indicates that the shoe product has been oxidized, then obtaining the record information of the shoe product; the record information includes the record time and the record content corresponding to the record time; the record time includes at least one of the production date, the storage date, the inspection date or the maintenance date; the record content includes at least one of the image data of the shoe product, the material performance characteristics or the storage conditions; Based on the recording time and the recording content corresponding to the recording time, adding corresponding state information to the time point in the recording time to generate a shoe product time sequence feature; the state information is obtained through the recording content corresponding to the recording time; The time series characteristics of the shoes are input into a preset time series model to obtain oxidation trend analysis results of the shoes; the trend analysis results are used to characterize the potential oxidation degree of the shoes at different time points.

3. The method according to claim 2, characterized in that The step of inputting the footwear features of the footwear that is classified as footwear that will oxidize into a pre-trained oxidation age prediction model to obtain an oxidation age result includes: Obtaining input features according to the footwear features of the footwear that is susceptible to oxidation as a result of the oxidation classification and the oxidation trend analysis result; The input features are input into the oxidation age prediction model to obtain an oxidation age result.

4. The method according to claim 1, characterized in that: Before inputting the footwear features into a pre-trained oxidation classification model to obtain an oxidation classification result, the method further includes: Obtaining sample shoe features of the sample shoe products; the sample shoe features include at least one of brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, matching color, secondary color, craftsmanship, environmental features or supply chain features; Adding an oxidation classification label to the sample footwear to obtain a classification training sample; the oxidation classification label includes oxidation or non-oxidation; The classification model is trained using the classification training samples to obtain the oxidation classification model.

5. The method according to claim 1, characterized in that Before inputting the footwear features of the footwear classified as the footwear that will oxidize into a pre-trained oxidation age prediction model to obtain an oxidation age result, the method further includes: Obtaining sample shoe characteristics of the sample shoe that has been oxidized; the sample shoe characteristics include at least one of brand, shoe category, applicable population, functionality, price, sole material, upper material, primary color, color matching, secondary color, craftsmanship, environmental characteristics, or supply chain characteristics; Adding the oxidation age label corresponding to the sample shoe to the sample shoe feature to obtain an age training sample; the oxidation age label includes the time interval from the production time to the oxidation of the sample shoe; The initial model is trained using the birth age training samples to obtain the oxidation birth age prediction model.

6. The method according to claim 5, characterized in that The sample footwear features include a feature matrix; The step of obtaining the sample footwear characteristics of the sample footwear that has been oxidized comprises: Obtaining original feature information of sample shoes; preprocessing the original feature information to obtain processed shoe information; The processed footwear information is associated according to the unique identifier of the sample shoe to obtain the feature matrix; a row of features in the feature matrix represents the sample footwear features of the same sample shoe.

7. The method according to any one of claims 1 to 6, characterized in that: After obtaining the oxidation age result, the method further comprises: Obtaining the actual production age of the shoe product; Determining whether the shoe is a high-risk oxidation shoe according to the difference between the actual age and the oxidation age result; Testing the high-risk oxidation footwear to determine whether oxidation has occurred and obtaining actual test results; Supplementary data is generated based on the actual detection results, and the oxidation age prediction model is adjusted using the supplementary data to obtain an updated oxidation age prediction model, so that subsequent shoes can be predicted using the updated oxidation age prediction model.

8. A computer program product, characterized in that The method comprises computer program instructions, and when the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is executed.

9. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method according to any one of claims 1 to 7 is executed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is executed.

Images