Device for predicting foot pain risk and storage medium

Abstract

The invention discloses a foot pain risk prediction device and a storage medium. The apparatus comprises: a processor; and a memory in which a program instruction for predicting the risk of foot pain is stored, and when the program instruction is executed by the processor, the device implements the following operations: collecting a first predictive factor and a second predictive factor related to the foot pain of a detected person; constructing a first regression model and a second regression model based on the first prediction factor and the second prediction factor respectively; performing first risk prediction and second risk prediction of foot pain based on the first regression model and the second regression model respectively to obtain a corresponding first prediction result and a corresponding second prediction result; and obtaining a final prediction result according to the first prediction result and the second prediction result. By means of the scheme, the accuracy of foot pain risk prediction can be improved, and a reference is provided for preventing foot pain of a detected person.

Description

TECHNICAL FIELD

[0001] The present application generally relates to the technical field of foot pain risk prediction. More particularly, the present application relates to an apparatus for foot pain risk prediction and a computer readable storage medium. BACKGROUND

[0002] Foot pain is a common health problem for people who stand for long periods, walk or have specific occupations, such as nurses. Nurses are required to stand for long periods and walk frequently in their work, and factors such as shoe selection and previous foot injuries can affect foot health. The incidence of severe foot pain in nurses is significantly higher than that in other occupational groups. Severe foot pain not only reduces the work efficiency of nurses, but also can lead to chronic foot diseases, which has a long-term impact on their physical and mental health. Therefore, it is of great significance to accurately predict the risk of severe foot pain in nurses and other groups for early intervention and reducing the incidence of pain.

[0003] In existing foot pain risk prediction techniques, traditional prediction models have many limitations. On the one hand, single regression models output abstract regression coefficients or log odds ratios, which lack intuitiveness and are difficult for medical staff to quickly interpret and apply in practice. On the other hand, some models do not establish an efficient variable screening mechanism, either failing to eliminate redundant variables (such as nationality, marital status, and other factors with low correlation to foot pain) from high-dimensional data or relying on traditional stepwise regression with poor stability, resulting in inaccurate identification of key prediction factors (such as shoe comfort and foot injury history). In addition, existing models are not designed for multi-dimensional validation and joint prediction, and the reliability and generalization ability of single models are insufficient, making it difficult to adapt to individual differences in different groups of nurses. At the same time, the information collection dimension is redundant (such as including gender, occupation, and other non-core factors), which not only increases the collection cost but also easily introduces interference variables.

[0004] Therefore, there is an urgent need to provide a scheme for foot pain risk prediction to improve the accuracy of foot pain risk prediction and provide a reference for preventing foot pain in subjects. SUMMARY

[0005] To at least solve one or more of the above-mentioned technical problems, the present application provides a scheme for foot pain risk prediction in multiple aspects.

[0006] In a first aspect, the present application provides a device for foot pain risk prediction, comprising: a processor; and a memory having stored therein program instructions for foot pain risk prediction, which, when executed by the processor, cause the device to implement the following operations: collecting first and second predictors related to foot pain of a subject; constructing first and second regression models based on the first and second predictors, respectively; performing first and second risk predictions of foot pain based on the first and second regression models, respectively, to obtain corresponding first and second prediction results; and obtaining a final prediction result according to the first and second prediction results.

[0007] In some embodiments, the first predictors comprise at least age, education, work shoe material, work shoe comfort, foot injury history, and whether or not to include complications of the subject; and the second predictors comprise at least age, ethnicity, marriage, sleep, step count, work shoe type, work shoe shape, shoe sole thickness, self-perceived foot health level, whether or not to include complications, and foot injury history of the subject.

[0008] In some embodiments, the first regression model comprises a Logistic regression model, and the second regression model comprises a Lasso regression model.

[0009] In some embodiments, the device further comprises: performing a preprocessing operation on the first and second predictors, wherein the preprocessing operation comprises at least one or more of data cleaning, variable encoding processing, and L1 regularization.

[0010] In some embodiments, the program instructions, when executed by the processor, cause the device to further implement the following operations to construct the first regression model: screening, based on the first predictors, a predictor whose relevance to severe foot pain meets a preset threshold as a candidate predictor variable; constructing a linear relationship with a log odds ratio as an output according to the candidate predictor variable, wherein the log odds ratio is a natural logarithm of a ratio of a probability of occurrence of severe foot pain of a nurse to a probability of non-occurrence; and constructing the first regression model based on the linear relationship.

[0011] In some embodiments, the program instructions, when executed by the processor, cause the device to further implement the following operations to construct the linear relationship with the log odds ratio as the output: calculating a first regression coefficient of each candidate predictor variable and a first model intercept term by using a maximum likelihood estimation method to construct the linear relationship with the log odds ratio as the output.

[0012] In some embodiments, wherein the program instructions, when executed by the processor, cause the apparatus to further implement the following operations to construct the second regression model: calculating a target regularization parameter based on the second prediction factor; constructing a total loss function of the second regression model according to the target regularization parameter by increasing an L1 regularization penalty term; and determining a regression relationship according to the total loss function to construct the second regression model.

[0013] In some embodiments, wherein the program instructions, when executed by the processor, cause the apparatus to further implement the following operations to determine the regression relationship: calculating a minimum value of the total loss function to determine a second regression coefficient, a key prediction factor and a second model intercept term to determine the regression relationship.

[0014] In some embodiments, wherein the program instructions, when executed by the processor, cause the apparatus to further implement the following operations to obtain a final prediction result according to the first prediction result and the second prediction result: performing a weighted sum or an arithmetic mean on the first prediction result and the second prediction result to obtain the final prediction result.

[0015] In a second aspect, the present application provides a computer-readable storage medium having stored thereon computer-readable instructions for foot pain risk prediction, which, when executed by one or more processors, implement the operations performed by the apparatus of one or more embodiments of the preceding first aspect.

[0016] Through the scheme for foot pain risk prediction as provided above, the embodiments of the present application solve the problems of existing single model interpretation difficulty, inaccurate variable screening, low reliability and information redundancy by collecting two types of prediction factors related to the foot pain of a subject, constructing a first regression model and a second regression model respectively for double risk prediction, and obtaining a final prediction result by combining the double prediction results, thereby improving the accuracy, intuitiveness and generalization ability of foot pain risk prediction, efficiently adapting to individual differences of nurses and other populations, and providing precise technical support for early intervention and health management of foot pain. BRIEF DESCRIPTION OF DRAWINGS

[0017] The above and other objects, features and advantages of the exemplary embodiments of the present application will be more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which: Figure 1 is an exemplary structural block diagram illustrating a scheme for foot pain risk prediction according to an embodiment of the present application; Figure 2is an exemplary flow chart illustrating overall operations performed by an apparatus for foot pain risk prediction according to embodiments of the present application; Figure 3 is an exemplary schematic diagram illustrating an nomogram according to embodiments of the present application; Figure 4 is an exemplary schematic diagram illustrating a characteristic curve graph of a nomogram model according to embodiments of the present application; Figure 5 is an exemplary schematic diagram illustrating a calibration curve graph of a nomogram model according to embodiments of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of the present application.

[0019] It should be understood that the terms "comprise" and "include" used in the specification and claims of the present application indicate the presence of the described features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0020] It should also be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and claims of the present application, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should be further understood that the term "and / or" used in the specification and claims of the present application means any combination of one or more of the associated listed items and all possible combinations thereof, and includes these combinations.

[0021] As used in the specification and claims of the present application, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting" depending on the context. Similarly, the phrases "if it is determined" or "if [a described condition or event] is detected" can be interpreted as meaning "upon determining" or "in response to determining" or "upon detecting [a described condition or event]" or "in response to detecting [a described condition or event]" depending on the context.

[0022] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0023] Figure 1is an exemplary structural block diagram showing an apparatus 100 for foot pain risk prediction according to an embodiment of the present application. As shown in Figure 1 The apparatus 100 can include a processor 101 and a memory 102, as shown in the foregoing. The processor 101 can include, for example, a general-purpose processor (“CPU”) or a special-purpose graphics processor (“GPU”), and the memory 102 can store program instructions executable on the processor. In some embodiments, the memory 102 can include, but is not limited to, a resistive random access memory (“RRAM”), a dynamic random access memory (“DRAM”), a static random access memory (“SRAM”), and an enhanced dynamic random access memory (“EDRAM”).

[0024] Further, the memory 102 can store program instructions for foot pain risk prediction, which, when executed by the processor, cause the apparatus 100 to perform the following operations: collect first and second predictors related to the subject's foot pain; construct first and second regression models based on the first and second predictors, respectively; perform first and second risk predictions of foot pain based on the first and second regression models, respectively, to obtain corresponding first and second prediction results; and obtain a final prediction result according to the first and second prediction results.

[0025] In some embodiments, the first predictor can include at least the subject's age, education, work shoe material (e.g., natural leather, artificial leather, cotton, etc.), work shoe comfort, foot injury history, and whether or not to include complications (e.g., diabetes, arthritis, etc.). The second predictor can include at least the subject's age, ethnicity, marriage, sleep, step count, work shoe type, work shoe shape (e.g., round, square, or pointed, etc.), sole thickness, self-perceived foot health level, whether or not to include complications, foot injury history, etc. In some implementation scenarios, the first and second predictor information can be collected through electronic questionnaires, which are designed in a structured manner and include single-choice, multiple-choice, and numerical input, etc. Objective indicators such as step count and sleep duration can be automatically obtained by connecting smart wearable devices, and after the collection is completed, the data is automatically stored in the database in the memory and a unique identifier is generated and associated with the subject.

[0026] In some embodiments, when the program instructions are executed by the processor, the apparatus is caused to further implement the following operations: performing a pre-processing operation on the first and second predictors, wherein the pre-processing operation comprises at least one or more of data cleaning, variable encoding processing, and LI regularization. In particular, the LI regularization is performed on the second predictor.

[0027] It can be understood that data cleaning is a basic step of pre-processing, aiming to eliminate noise and outliers in the data and ensure data quality, which can include, for example, missing value processing (missing value detection, missing mechanism judgment or missing value filling) and / or outlier processing (outlier detection or processing). Among them, for missing value processing, the missing value can be identified by counting the missing proportion of each variable, for example, the variable whose missing proportion exceeds the preset value needs to be evaluated for importance and whether to be excluded. For normally distributed continuous variables (such as age, step count), the mean filling method is used, and the calculation formula is wherein represents the mean, n represents the number of non-missing values, represents the non-missing value. For non-normally distributed continuous variables and ordered categorical variables (such as the comfort level of work shoes), the median filling method is used, that is, the value at the middle position after sorting is taken as the filling value; for unordered categorical variables (such as the material of work shoes), the mode filling method is used, that is, the category with the highest frequency is taken as the filling value.

[0028] For outlier processing, for continuous variables, the Z-Score method is used for outlier detection, and the calculation formula is wherein x is the variable value, μ is the mean, and σ is the standard deviation. When | Z|>3, it is determined as an outlier. For categorical variables, outliers are identified through frequency analysis, that is, categories with a frequency of less than 1% are considered outliers. For confirmed outliers, first verify whether there is an error in the data collection process, and if it is a collection error, correct it. For outliers that cannot be corrected, the truncation method is used for processing, that is, for Z-Score>3, replace it with μ+3σ; for Z-Score<-3, replace it with μ-3σ.

[0029] Variable encoding can convert non-numeric data into numeric data recognizable by the model while retaining the semantic information of the category variable, avoiding model bias caused by incorrect encoding. In some implementation scenarios, variable encoding processing can include, for example, label encoding, one-hot encoding, and binary classification encoding. Among them, label encoding is used for, for example, label encoding, work shoe comfort level, self-perceived foot health level, whether to include complications, etc. Taking whether to include complications as an example, 0 is set to 0, 1 is set to 1, 2 is set to 2, and 3 and above are set to 3. One-hot encoding is used for, for example, work shoe material, work shoe type, etc. For example, natural leather corresponds to [1, 0, 0], artificial leather corresponds to [0, 1, 0], and cotton corresponds to [0, 0, 1]. Binary classification encoding is used for foot injury history, with no setting 0 and with setting 1.

[0030] For the second prediction factor, L1 regularization is added in the preprocessing, so that a penalty term of the absolute value of the parameter is added in the loss function in the subsequent, realizing parameter sparsification, that is, part of the parameters become 0, achieving the purpose of feature selection. In an implementation scenario, the L1 regularization term is: where λ is a regularization parameter that controls the penalty strength, is a model parameter. L1 regularization can effectively solve the overfitting problem of high-dimensional data, realize automatic selection of variables, simplify the model structure, and at the same time maintain a high prediction accuracy.

[0031] In some embodiments, when the program instructions are executed by the processor, the device implements the following operations: constructing a first regression model and a second regression model based on a first prediction factor and a second prediction factor, respectively. In some embodiments, the first regression model includes a Logistic regression model (i.e., a logistic regression model), and the second regression model includes a Lasso regression model (i.e., a least absolute shrinkage and selection operator regression model).

[0032] In some embodiments, when the program instructions are executed by the processor, the device further implements the following operations to construct the first regression model: selecting, based on the first prediction factor, a prediction factor whose relevance to foot severe pain meets a preset threshold as a candidate prediction variable, constructing a linear relationship with a log odds ratio as an output according to the candidate prediction variable, wherein the log odds is the natural logarithm of the ratio of the probability of occurrence of foot severe pain to the probability of non-occurrence of foot severe pain. Then, based on the linear relationship, the first regression model is constructed. In some embodiments, the maximum likelihood estimation method can be used to calculate the first regression coefficient of each candidate prediction variable and the first model intercept term to construct a linear relationship with a log odds ratio as an output.

[0033] Specifically, a method such as univariate analysis can be used to assess the association between each primary predictor and severe foot pain. For example, for continuous variables (e.g., age), an independent samples t-test is used to compare the means of the pain group and the non-pain group; for ordered categorical variables (e.g., work shoe comfort), ordered logistic regression analysis is used; and for unordered categorical variables (e.g., work shoe material) or binary variables (e.g., history of foot injury), a chi-square test is used to assess the association between each primary predictor and severe foot pain. Next, predictors whose association meets a preset threshold (e.g., significance level α = 0.05) are included as candidate predictors. If the association p-value is < 0.05, the predictor is considered to have a statistical association with severe foot pain and is included as a candidate predictor. Based on this, predictors with weak associations to foot pain are effectively eliminated, reducing the model input dimensions, lowering the computational complexity of subsequent multivariate analysis, and avoiding interference from irrelevant variables.

[0034] Furthermore, a linear relationship with the log-dot ratio as the output is constructed based on the candidate predictor variables to build the first regression model. Specifically, the first regression coefficients and the intercept term of the first model for each candidate predictor variable are calculated using the maximum likelihood estimation method to establish the linear relationship.

[0035] In some implementation scenarios, the aforementioned linear relationship can be represented as ,in Indicates candidate predictor variables, denoted as the first regression coefficient of each candidate predictor variable, and b as the intercept term.

[0036] In some implementation scenarios, the aforementioned logarithmic advantage can be expressed as follows:

[0037] From this, we can obtain:

[0038] Based on the above, the first regression coefficients of each candidate predictor variable are calculated using the maximum likelihood estimation method. Intercept term b of the first model, and thus obtain Thus, the first regression model is obtained.

[0039] In some embodiments, when the program instructions are executed by the processor, the apparatus further performs the following operations to construct a second regression model: calculates a target regularization parameter based on a second predictor, adds an L1 regularization penalty term to the target regularization parameter to construct the total loss function of the second regression model, and determines the regression relationship based on the total loss function to construct the second regression model. In some embodiments, the second regression coefficients, key predictor, and second model intercept term can be determined by calculating the minimum value of the total loss function to determine the regression relationship.

[0040] In some implementation scenarios, the Z-Score method can be used to standardize the second predictor to avoid excessive penalty for features with large units (such as age measured in "years," whose numerical range may be much larger than other binary features), ensuring a fair and reasonable regularization effect. Next, a target regularization parameter (i.e., the optimal regularization parameter λ) is selected through methods such as cross-validation. Specifically, the second predictor is divided into training and test sets. Five-fold cross-validation is performed on the training set, repeated five times, and the average deviation corresponding to different λ values ​​is calculated. The λ value that minimizes the validation deviation is selected as the target regularization parameter.

[0041] Next, an L1 regularization term can be introduced into the traditional logistic regression model to construct the total loss function of the second regression model. By minimizing the total loss function, the second regression coefficients, key predictors, and the second model intercept term can be obtained, thus constructing the second regression model. In some implementation scenarios, minimizing the total loss function can be expressed as follows:

[0042] in, Indicates the true label, Indicates the predicted probability. Includes intercept term Second regression coefficient N represents the sample size, and λ represents the optimal regularization parameter λ. Additionally, the key predictor factors correspond to... The corresponding features. The aforementioned second regression coefficient, key predictor factors, and second model intercept term determine the regression relationship. Thus, the second regression model was obtained.

[0043] Based on the constructed first and second regression models, the first and second risk predictions for foot pain are performed respectively, yielding the corresponding first and second prediction results. Specifically, in some implementation scenarios, the Sigmoid function can be used to separate the first and second regression models. Mapping to the target interval (e.g., 0~1) to obtain the predicted probability, thereby obtaining the first and second prediction results.

[0044] Further, a final prediction result is obtained based on the first prediction result and the second prediction result. In some embodiments, when the program instructions are executed by the processor, the device further performs the following operation to obtain the final prediction result based on the first prediction result and the second prediction result: performing a weighted sum or arithmetic average of the first prediction result and the second prediction result to obtain the final prediction result. The final prediction result corresponds to the probability of occurrence of severe foot pain; for example, a probability of occurrence < 0.2 is considered low risk, a probability of occurrence between 0.2 and 0.5 is considered medium risk, and a probability of occurrence > 0.5 is considered high risk.

[0045] As described above, this application's embodiments collect two types of predictive factors (the first type includes indicators such as age and work shoes, and the second type expands to include more indicators such as ethnicity and sleep). After data cleaning and variable encoding preprocessing, a Logistic regression model (screening factors associated with foot pain and outputting the first prediction result) and a Lasso regression model (selecting regularization parameters through cross-validation, using a loss function with L1 regularization to screen key factors and outputting the second prediction result) are constructed respectively. The results of the two models are then fused to obtain the final prediction result. At the same time, nomogram visualization and cross-validation are used to ensure model performance. Computer-readable storage media stores instructions to implement the above operations, solving the problem of low prediction accuracy in existing technologies and providing a tool for foot health management for people in occupations that require long periods of standing.

[0046] Figure 2 This is an exemplary flowchart illustrating the overall operation performed by a device for predicting the risk of foot pain according to an embodiment of this application. It should be understood that... Figure 2 The above Figure 1 One specific embodiment of the operation performed by the device 100, therefore the above regarding Figure 1 The description also applies to Figure 2 .

[0047] like Figure 2As shown, in step S201, a first predictor and a second predictor are collected. The first predictor may include, for example, age, education level, work shoe material, work shoe comfort, history of foot injury, and whether complications are present. The second predictor may include, for example, age, ethnicity, marital status, sleep duration, step count, type of work shoe, work shoe shape, sole thickness, self-perceived foot health level, whether complications are present, and history of foot injury. Next, in step S202, the first predictor is input into the constructed first regression model for a first risk prediction, and in step S203, a prediction result is obtained. In some embodiments, the first regression model may be, for example, a logistic regression model.

[0048] Further, in step S204, the second predictor is input into the constructed second regression model for the second risk prediction, and in S205, the second prediction result is obtained. In some embodiments, the second regression model may be, for example, a Lasso regression model. In some implementation scenarios, preprocessing operations may be performed on the first and second predictors before inputting them into the model, such as data cleaning, variable encoding, and L1 regularization, or one or more of these. For specific details regarding the construction of the first and second regression models, please refer to the foregoing. Figure 1 The description of the first and second prediction results is not repeated here. Finally, in step S206, the first prediction result and the second prediction result are weighted and summed or arithmetic averaged to obtain the final prediction result.

[0049] In some implementation scenarios, the aforementioned first regression analysis can be visualized graphically. This involves integrating multiple predictive indicators and plotting them on a plane using scaled line segments to represent the relationships between the variables in the first regression model. For example... Figure 3 The nomogram shown.

[0050] Figure 3 This is an exemplary schematic diagram illustrating a column line diagram according to an embodiment of this application. For example... Figure 3 The example illustrates several predictors: age, education level, work shoe material, work shoe comfort, number of complications, and history of foot injuries. For education level, each scale represents 1 = vocational school, 2 = junior college, 3 = bachelor's degree, and 4 = master's degree; for work shoe material, each scale represents 1 = natural leather, 2 = synthetic leather, 3 = cotton, and 4 = plastic; for the number of complications, each scale represents the actual value; and for the history of foot injuries, each scale represents 0 = none and 1 = yes.

[0051] The actual value of each predictor factor is projected upwards to obtain a score. The scores of all predictors are summed to obtain a total score. The total score is then projected downwards onto the corresponding value for foot pain, yielding the probability of experiencing severe foot pain. For example, a nurse, 40 years old, with a bachelor's degree, wearing leather work shoes, a work shoe comfort score of 4, no history of foot injury, and 0 other complications, has a probability of approximately 0.12 (56 + 22 + 14 + 0 + 0 + 0 = 92). In other words, based on the contribution of each influencing factor to the outcome variable (the magnitude of the regression coefficient), each value level of each influencing factor is assigned a score. These scores are then summed to obtain a total score. Finally, the predicted value of the individual's outcome event is calculated using the functional relationship between the total score and the probability of the outcome event. Similarly, the second regression model can also be represented by the aforementioned array diagram.

[0052] Figure 4 This is an exemplary schematic diagram illustrating the characteristic curves of a noctilinear graph model according to an embodiment of this application. Figure 5 This is an exemplary schematic diagram illustrating the calibration curve of a nodal plot model according to an embodiment of this application. Figure 4 As shown in the figure, the horizontal axis represents the false positive rate, and the vertical axis represents the true positive rate. The concordance index (C-index) is 0.706; a value greater than 0.70 indicates high accuracy. Figure 5 The figure shows that the standard curve and the calibration prediction curve fit well. The horizontal axis represents the predicted probability {pain = 1}, and the vertical axis represents the actual probability. The figure shows that the predicted value of nurse foot pain obtained according to the model has a good consistency with the observed value, and the accuracy of the prediction effect is high.

[0053] Based on the foregoing description in conjunction with the accompanying drawings, those skilled in the art will understand that the embodiments of this application can also be implemented by software programs. Therefore, this application also provides a computer-readable storage medium. This computer-readable storage medium stores computer-readable instructions thereon for predicting foot pain risk. When executed by one or more processors, these computer-readable instructions implement the embodiments of this application in conjunction with the accompanying drawings. Figure 1 The operation performed by the device described for predicting the risk of foot pain.

[0054] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0055] It should be noted that although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0056] It should be understood that when the terms "first," "second," "third," and "fourth," etc., are used in the claims, specification, and drawings of this application, they are used only to distinguish different objects and not to describe a specific order. The terms "comprising" and "including" as used in the specification and claims of this application indicate the presence of the described features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof.

[0057] It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this specification and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this specification and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0058] Although the embodiments of this application are described above, the content is merely an example adopted for the purpose of facilitating understanding of this application and is not intended to limit the scope and application scenarios of this application. Any person skilled in the art described in this application may make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in this application, but the scope of patent protection of this application shall still be determined by the scope defined in the appended claims.

Claims

1. A device for predicting the risk of foot pain, comprising: processor; as well as A memory storing program instructions for predicting the risk of foot pain, which, when executed by the processor, cause the device to perform the following operations: Collect the first and second predictors of foot pain associated with the subjects; A first regression model and a second regression model are constructed based on the first predictor and the second predictor, respectively. First and second risk predictions for foot pain are performed based on the first regression model and the second regression model, respectively, to obtain the corresponding first and second prediction results. The final prediction result is obtained based on the first prediction result and the second prediction result.

2. The device according to claim 1, wherein the first predictive factor includes at least the subject's age, education level, work shoe material, work shoe comfort level, history of foot injury, and whether it includes complications; the second predictive factor includes at least the subject's age, ethnicity, marital status, sleep, steps, type of work shoe, work shoe shape, sole thickness, self-perceived foot health level, whether it includes complications, and history of foot injury.

3. The apparatus according to claim 1, wherein the first regression model includes a Logistic regression model, and the second regression model includes a Lasso regression model.

4. The apparatus according to claim 1, further comprising: The first predictor and the second predictor are preprocessed, wherein the preprocessing operation includes at least one or more of data cleaning, variable encoding, and L1 regularization.

5. The apparatus of claim 1 or 4, wherein when the program instructions are executed by the processor, the apparatus further performs the following operations to construct the first regression model: Based on the first predictor, predictors whose association with severe foot pain meets a preset threshold are selected as candidate predictors. A linear relationship is constructed based on the candidate predictor variables, with the log-odds ratio as the output, where the log-odds ratio is the natural logarithm of the ratio of the probability of severe foot pain occurring to the probability of not occurring in nurses. Based on the linear relationship, the first regression model is constructed.

6. The apparatus of claim 5, wherein when the program instructions are executed by the processor, the apparatus further performs the following operation to construct a linear relationship with a logarithmic advantage ratio as the output: The first regression coefficient and the first model intercept term of each candidate predictor variable are calculated using the maximum likelihood estimation method, and a linear relationship is constructed with the log odds ratio as the output.

7. The apparatus of claim 1 or 4, wherein when the program instructions are executed by the processor, the apparatus further performs the following operations to construct the second regression model: Calculate the target regularization parameter based on the second predictor factor; The total loss function of the second regression model is constructed by adding an L1 regularization penalty term based on the target regularization parameter. The regression relationship is determined based on the total loss function to construct the second regression model.

8. The apparatus of claim 7, wherein when the program instructions are executed by the processor, the apparatus further performs the following operations to determine the regression relationship: The minimum value of the total loss function is calculated to determine the second regression coefficient, the key predictor, and the second model intercept term, thereby determining the regression relationship.

9. The apparatus of claim 1, wherein when the program instructions are executed by the processor, the apparatus further performs the following operation to obtain a final prediction result based on the first prediction result and the second prediction result: The first prediction result and the second prediction result are weighted and summed or arithmetically averaged to obtain the final prediction result.

10. A computer-readable storage medium storing computer-readable instructions for predicting foot pain risk, which, when executed by one or more processors, perform the operations performed by the apparatus as claimed in any one of claims 1-9.

Images