A sock customization design method based on intelligent optimization algorithms

Abstract

The present invention relates to the technical field of sock customization design, and specifically to a sock customization design method based on an intelligent optimization algorithm. The present invention obtains the basic information data, body shape data and historical purchase data of customers; inputs these three types of data into a sock preference prediction model to identify a sock preference feature set and calculate a sock preference degree set of customers, including multiple preference degree parameters such as color, size, pattern, function, material and thickness. Based on the sock preference degree set, a first sock design optimization objective function is constructed, and the particle swarm algorithm is used to optimize and analyze the objective function to obtain a first sock customization design parameter set. Sock design is carried out according to the design parameter set and recommended to customers, and the design parameters are adjusted according to the feedback data to obtain a second sock customization design parameter set. This method realizes precise sock customization design through an intelligent optimization algorithm, which helps to improve customer satisfaction.

Description

Technical Field

[0001] The present invention relates to the technical field of sock customization design, and specifically provides a sock customization design method based on an intelligent optimization algorithm. Background Art

[0002] Socks are an essential clothing item in daily life. However, with the continuous improvement of living standards, people's aesthetic standards are also constantly rising, and thus the emphasis on the styles of socks is increasing day by day. Therefore, socks should not only have the basic value of keeping warm, but also meet people's more style requirements.

[0003] Traditional sock manufacturers are difficult to meet the customization requirements of different people for elements such as size, color, and pattern in sock design, which is likely to lead to problems with customer satisfaction.

[0004] Therefore, a sock customization design method based on an intelligent optimization algorithm is proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a sock customization design method based on an intelligent optimization algorithm. The present invention relates to the technical field of sock customization design, and specifically provides a sock customization design method based on an intelligent optimization algorithm. The present invention obtains the basic information data, body posture data, and historical purchase data of customers; inputs the three types of data into a sock preference prediction model to identify a set of sock preference features and calculate a set of sock preference degrees of customers, including multiple preference degree parameters such as color, size, pattern, function, material, and thickness. Based on the set of sock preference degrees, a first sock design optimization objective function is constructed, and the particle swarm algorithm is used to optimize and analyze the objective function to obtain a first set of sock customization design parameters. Sock design is carried out according to the set of design parameters and recommended to customers, and the design parameters are adjusted according to the feedback data to obtain a second set of sock customization design parameters. This method realizes precise sock customization design through an intelligent optimization algorithm, which helps to improve customer satisfaction.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] A sock customization design method based on an intelligent optimization algorithm, including:

[0008] S10. Obtain the first basic information data, second body posture data, and first historical sock purchase data of customers;

[0009] S20. Input these three types of data into the sock preference prediction model, and successively pass through the preprocessing layer, preference feature extraction layer, preference degree analysis layer, and output layer of the sock preference prediction model; the preference feature extraction layer analyzes through a deep neural network to obtain a sock preference feature set; the preference degree analysis layer obtains a sock preference degree set of customers according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, and a thickness preference degree set;

[0010] S30. According to the sock preference degree set and the preset objective function optimization threshold, construct a first sock design optimization objective function, perform optimization analysis on the first sock design optimization objective function based on the particle swarm algorithm to obtain multiple groups of sock optimization parameters, and construct a first sock customized design parameter set according to the multiple groups of sock optimization parameters;

[0011] S40. Conduct sock design according to the first sock customized design parameter set, recommend the designed socks to customers, obtain customer feedback data on the recommended socks, obtain the first sock design satisfaction of customers according to the recommended sock feedback data, and adjust the first sock customized design parameter set based on the first sock design satisfaction and sock customized design constraint conditions to obtain a second sock customized design parameter set.

[0012] Preferably, the first basic information data includes age, gender, and occupation; the second body posture data includes shoe size, height, and body type; the first historical sock purchase data includes multi-element data of each sock purchased by the customer, including color, pattern, size, material, function, and thickness, as well as the evaluation data of the customer for each sock.

[0013] Preferably, the sock preference prediction model includes a preprocessing layer, a sock preference feature extraction layer, a preference degree analysis layer, and an output layer;

[0014] The preprocessing layer is used to preprocess the first basic information data, the second body posture data, and the first historical sock purchase data to obtain second basic information data, third body posture data, and second historical sock purchase data; the preprocessing includes data cleaning, normalization, and one-hot encoding;

[0015] The preference feature extraction layer is used to identify the second basic information data, the third body posture data, and the second historical sock purchase data through a deep neural network to obtain the sock preference feature set; the sock preference feature set includes a color preference feature set, a size preference feature set, a pattern preference feature set, a material preference feature set, a function preference feature set, and a thickness preference feature set;

[0016] The preference degree analysis layer is used to obtain the sock preference degree set of customers according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, a function preference degree set, and a thickness preference degree set;

[0017] The output layer is used to output the sock preference degree set.

[0018] Preferably, the sock preference feature is used to represent the correlation between the historical purchase data of customers and the basic information data and body data of customers; the sock preference feature set includes preference features, basic information influence coefficients, and body data influence coefficients; the sock preference degree set of each sock preference feature set is obtained according to the basic information influence coefficient and the body data influence coefficient.

[0019] Preferably, the sock preference feature set is:

[0020] ;

[0021] Among them, represents the sock preference feature set; represents the color preference feature set; represents the size preference feature set; represents the pattern preference feature set; represents the material preference feature set; represents the function preference feature set; represents the thickness preference feature set; represents the th color preference feature in the color preference feature set; represents the th size preference feature in the size preference feature set; represents the th pattern preference feature in the pattern preference feature set; represents the th material preference feature in the material preference feature set; represents the th function preference feature in the function preference feature set; represents the th thickness preference feature in the thickness preference feature set; represents the basic information data influence coefficient of the th color preference feature in the color preference feature set; represents the body data influence coefficient of the th color preference feature in the color preference feature set; represents the The influence coefficient of the basic information data of a size preference feature; Indicates the influence coefficient of the body shape data of the th size preference feature in the size preference feature set; Indicates the influence coefficient of the basic information data of the th pattern preference feature in the pattern preference feature set; Indicates the influence coefficient of the body shape data of the th pattern preference feature in the pattern preference feature set; Indicates the influence coefficient of the basic information data of the th material preference feature in the material preference feature set; Indicates the influence coefficient of the body shape data of the th material preference feature in the material preference feature set; Indicates the influence coefficient of the basic information data of the th function preference feature in the function preference feature set; Indicates the influence coefficient of the body shape data of the th function preference feature in the function preference feature set; Indicates the influence coefficient of the basic information data of the th thickness preference feature in the thickness preference feature set; Indicates the influence coefficient of the body shape data of the th thickness preference feature in the thickness preference feature set.

[0022] Preferably, the sock preference degree set is:

[0023] ;

[0024] Among them, Indicates the sock preference degree set; Indicates the color preference degree set; Indicates the preference degree of the th color preference feature in the color preference degree set; Indicates the size preference degree set; Indicates the preference degree of the th size preference feature in the size preference degree set; Indicates the pattern preference degree set; Indicates the preference degree of the th pattern preference feature in the pattern preference degree set; Indicates the material preference degree set; Indicates the preference degree of the th material preference feature in the material preference degree set; Indicates the function preference degree set; Indicates the preference degree of the Preference degree of a functional preference feature; Represents a set of thickness preference degrees; Represents the Preference degree of the thickness preference feature in the set of thickness preference degrees; Represents the exponential function with the natural constant

[0025] Preferably, the first sock design optimization objective function is:

[0026] ;

[0027] Wherein, Represents the first sock design optimization objective function; Represents the influence weight of color preference degree; Represents the Preference degree of the th color preference feature; Represents the influence weight of size preference degree; Represents the Preference degree of the th size preference feature; Represents the influence weight of pattern preference degree; Represents the influence weight of material preference degree; Represents the Preference degree of the th material preference feature; Represents the influence weight of functional preference degree; Represents the Preference degree of the th functional preference feature; Represents the influence weight of thickness preference degree; Represents the optimization threshold of the preset objective function.

[0028] Preferably, the recommended sock feedback data includes color satisfaction, pattern satisfaction, size satisfaction, functional satisfaction, thickness satisfaction, and material satisfaction.

[0029] Preferably, the first sock design satisfaction is:

[0030] ;

[0031] Wherein, Represents the first sock design satisfaction; Represents the number of parameter groups in the first sock customization design parameter set; Represents the customer's color satisfaction with the sock designed with the th group of parameters; Indicates the customer's satisfaction with the size of the socks designed with the group of parameters; Indicates the customer's satisfaction with the pattern of the socks designed with the group of parameters; Indicates the customer's satisfaction with the material of the socks designed with the group of parameters; Indicates the customer's satisfaction with the function of the socks designed with the group of parameters; Indicates the customer's satisfaction with the thickness of the socks designed with the group of parameters.

[0032] Preferably, the sock customization design constraint condition is:

[0033] ;

[0034] Among them, represents the sock customization design constraint condition; represents the first sock design satisfaction degree; represents the sock design satisfaction threshold; represents the first sock customization design parameter set any parameter group in .

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0036] 1. The present invention obtains the customer's first basic information data, second body posture data and first historical sock purchase data; inputs these three types of data into the sock preference prediction model, and passes through the preprocessing layer, preference feature extraction layer, preference degree analysis layer and output layer of the sock preference prediction model; the preference feature extraction layer analyzes through a deep neural network to obtain a sock preference feature set; the preference degree analysis layer obtains the customer's sock preference degree set according to the basic information data influence coefficient and body posture data influence coefficient in the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set and a thickness preference degree set; this method effectively interprets the customer's tendency degree for each preference feature through the fine analysis of the customer's basic information data and body posture data, provides a data basis for the subsequent optimization of sock customization design, and thus further improves the customer's satisfaction with sock customization design.

[0037] 2. The present invention optimizes the threshold according to the sock preference degree set and the preset objective function, constructs the first sock design optimization objective function, optimizes and analyzes the first sock design optimization objective function based on the particle swarm algorithm, obtains multiple groups of sock optimization parameters, and constructs the first sock customization design parameter set according to the multiple groups of sock optimization parameters. This method improves the effectiveness of sock customization design by constructing the first sock design optimization objective function and performing optimization design based on the particle swarm optimization algorithm, thereby further improving the customer satisfaction with sock customization design.

[0038] 3. The present invention designs socks according to the first sock customization design parameter set, recommends the designed socks to customers, obtains the recommended sock feedback data of customers, obtains the first sock design satisfaction of customers according to the recommended sock feedback data, and adjusts the first sock customization design parameter set based on the first sock design satisfaction and the sock customization design constraint conditions to obtain the second sock customization design parameter set. This method recommends sock design based on the optimized first sock customization design parameter set, reflects the customer sock satisfaction according to the feedback data, and makes effective adjustments based on this satisfaction, improving the comprehensiveness of sock customization design, thereby further improving the customer satisfaction with sock customization design. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a schematic flowchart of a sock customization design method based on an intelligent optimization algorithm provided by an embodiment of the present invention;

[0040] Figure 2 is a schematic structural diagram of a sock preference prediction model provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] Embodiment 1

[0043] A sock manufacturer applied a sock customization design method based on an intelligent optimization algorithm to improve customer satisfaction;

[0044] Referring to Figure 1 , which is a schematic flowchart of a sock customization design method based on an intelligent optimization algorithm provided by an embodiment of the present invention, specifically including:

[0045] S10. Obtain the first basic information data, the second body posture data, and the first historical sock purchase data of the customer;

[0046] Further, the first basic information data includes age, gender, and occupation; the second body posture data includes shoe size, height, and body type; the first historical sock purchase data includes multi-element data of each sock purchased by the customer, including color, pattern, size, material, function, and thickness, as well as the evaluation data of the customer for each sock; among them, the evaluation data is a score evaluation of the color, pattern, size, function, thickness, and material of the purchased sock; the full score is 10 points; among them, the pattern includes geometric pattern, text pattern, letter pattern, anime pattern, and no pattern, etc.; the function includes sweat absorption, warmth retention, and waterproofing, etc.; the material includes wool, cotton, and nylon, etc.; the body type is analyzed according to the IBM index; the shoe size is represented according to the Chinese size; the sock size is represented according to the length of the sock; part of the basic information data and body posture data are shown in Table 1; part of the historical purchase data is shown in Table 2;

[0047] Table 1 Part of the basic information data and body posture data table

[0048]

[0049] Table 2 Part of the historical purchase data table

[0050]

[0051] S20. Input these three types of data into the sock preference prediction model, and sequentially pass through the preprocessing layer, preference feature extraction layer, preference degree analysis layer, and output layer of the sock preference prediction model; the preference feature extraction layer is analyzed through a deep neural network to obtain a sock preference feature set; the preference degree analysis layer obtains a sock preference degree set of the customer according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, and a thickness preference degree set;

[0052] Further, the sock preference prediction model includes a preprocessing layer, a sock preference feature extraction layer, a preference degree analysis layer, and an output layer; the structural schematic diagram of the sock preference prediction model is as Figure 2 shown;

[0053] The preprocessing layer is used to preprocess the first basic information data, the second body posture data, and the first historical sock purchase data to obtain the second basic information data, the third body posture data, and the second historical sock purchase data; the preprocessing includes data cleaning, normalization, and one-hot encoding; specifically, one-hot encoding is performed on the gender, and normalization is performed on the age, height, shoe size, size, thickness, and the evaluation data of the customer for each sock;

[0054] The preference feature extraction layer is used to identify the second basic information data, the third body posture data, and the second historical sock purchase data through a deep neural network to obtain the sock preference feature set; the sock preference feature set includes a color preference feature set, a size preference feature set, a pattern preference feature set, a material preference feature set, a function preference feature set, and a thickness preference feature set;

[0055] The specific identification process of the sock preference feature set includes: first, converting the occupation in the basic information data, the body type in the body posture data, and the color, pattern, material, and function data in the multi-element data of each sock that the customer has purchased into vector representations; and splicing the preprocessed data into a comprehensive feature vector, and training through a deep neural network to learn the basic information data influence coefficient and the body posture data influence coefficient of each feature in each preference feature. The basic information data influence coefficient is used to represent the influence of the customer's basic information data on the preference feature; the body posture data influence coefficient is used to represent the influence of the customer's body posture data on the preference feature; and output the preference feature set.

[0056] The preference degree analysis layer is used to obtain the sock preference degree set of the customer according to the sock preference feature set. The sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, a function preference degree set, and a thickness preference degree set;

[0057] The output layer is used to output the sock preference degree set.

[0058] Further, the sock preference feature is used to represent the correlation between the customer's historical purchase data and the customer's basic information data and body posture data; the sock preference feature set includes preference features, basic information influence coefficients, and body posture data influence coefficients; the sock preference degree set of each sock preference feature set is obtained according to the basic information influence coefficient and the body posture data influence coefficient.

[0059] Further, the sock preference feature set is:

[0060] ;

[0061] Among them, represents the sock preference feature set; represents the color preference feature set; represents the size preference feature set; represents the pattern preference feature set; represents the material preference feature set; represents the function preference feature set; represents the thickness preference feature set; Represents the th color preference feature in the set of color preference features; Represents the th size preference feature in the set of size preference features; Represents the th pattern preference feature in the set of pattern preference features; Represents the th material preference feature in the set of material preference features; Represents the th function preference feature in the set of function preference features; Represents the th thickness preference feature in the set of thickness preference features; Represents the influence coefficient of the basic information data of the th color preference feature in the set of color preference features; influence coefficient of the body posture data of the th color preference feature in the set of color preference features; influence coefficient of the basic information data of the th size preference feature in the set of size preference features; influence coefficient of the body posture data of the th size preference feature in the set of size preference features; influence coefficient of the basic information data of the th pattern preference feature in the set of pattern preference features; influence coefficient of the body posture data of the th pattern preference feature in the set of pattern preference features; influence coefficient of the basic information data of the th material preference feature in the set of material preference features; influence coefficient of the body posture data of the th material preference feature in the set of material preference features; influence coefficient of the basic information data of the th function preference feature in the set of function preference features; influence coefficient of the body posture data of the th function preference feature in the set of function preference features; influence coefficient of the basic information data of the th thickness preference feature in the set of thickness preference features; influence coefficient of the body posture data of the

[0062] Furthermore, the sock preference degree set is:

[0063] ;

[0064] Among them, represents the sock preference degree set; represents the color preference degree set; represents the preference degree of the th color preference feature in the color preference degree set; represents the size preference degree set; represents the preference degree of the th size preference feature in the size preference degree set; represents the pattern preference degree set; represents the preference degree of the th pattern preference feature in the pattern preference degree set; represents the material preference degree set; represents the preference degree of the th material preference feature in the material preference degree set; represents the function preference degree set; represents the preference degree of the th function preference feature in the function preference degree set; represents the thickness preference degree set; represents the preference degree of the th thickness preference feature in the thickness preference degree set; represents the exponential function with the natural constant as the base.

[0065] In this embodiment, by obtaining the customer's first basic information data, second body posture data, and first historical sock purchase data; inputting these three types of data into the sock preference prediction model, through the preprocessing layer, preference feature extraction layer, preference degree analysis layer, and output layer of the sock preference prediction model; the preference feature extraction layer analyzes through a deep neural network to obtain the sock preference feature set; the preference degree analysis layer obtains the customer's sock preference degree set according to the basic information data influence coefficient and body posture data influence coefficient in the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, and a thickness preference degree set; this method effectively interprets the customer's inclination degree for each preference feature through a detailed analysis of the customer's basic information data and body posture data, provides a data basis for the subsequent optimization of sock customization design, and thus further improves the customer's satisfaction with sock customization design.

[0066] S30. Construct a first sock design optimization objective function according to the sock preference set and the preset objective function optimization threshold, perform optimization analysis on the first sock design optimization objective function based on the particle swarm algorithm, obtain multiple groups of sock optimization parameters, and construct a first sock customization design parameter set according to the multiple groups of sock optimization parameters;

[0067] Further, the particle swarm optimization parameters are constructed by combining the preference values of multiple preference features in the sock preference set, including various combination parameters such as color preference, pattern preference, size preference, function preference, thickness preference, and material preference; and construct a first sock customization design parameter set;

[0068] Further, the first sock design optimization objective function is:

[0069] ;

[0070] Wherein, represents the first sock design optimization objective function; represents the color preference influence weight; represents the th preference of the color preference feature; represents the size preference influence weight; represents the th preference of the size preference feature; represents the pattern preference influence weight; represents the th preference of the pattern preference feature; represents the material preference influence weight; represents the th preference of the material preference feature; represents the function preference influence weight; represents the th preference of the function preference feature; represents the thickness preference influence weight; represents the th preference of the thickness preference feature; represents the preset objective function optimization threshold.

[0071] According to the sock preference degree set and the preset objective function optimization threshold, this embodiment constructs the first sock design optimization objective function, optimizes and analyzes the first sock design optimization objective function based on the particle swarm algorithm, obtains multiple groups of sock optimization parameters, and constructs the first sock customized design parameter set according to the multiple groups of sock optimization parameters. This method improves the effectiveness of sock customized design by constructing the first sock design optimization objective function and performing optimization design based on the particle swarm optimization algorithm, thereby further improving the customer satisfaction with sock customized design.

[0072] S40. Perform sock design according to the first sock customized design parameter set, recommend the designed socks to customers, obtain the recommended sock feedback data from customers, obtain the first sock design satisfaction of customers according to the recommended sock feedback data, and adjust the first sock customized design parameter set based on the first sock design satisfaction and the sock customized design constraint conditions to obtain the second sock customized design parameter set.

[0073] Further, the recommended sock feedback data includes color satisfaction, pattern satisfaction, size satisfaction, function satisfaction, thickness satisfaction, and material satisfaction.

[0074] Further, the first sock design satisfaction is:

[0075] ;

[0076] Among them, represents the first sock design satisfaction; represents the number of parameter groups in the first sock customized design parameter set; represents the customer's color satisfaction with the socks designed with the th group of parameters; represents the customer's size satisfaction with the socks designed with the th group of parameters; represents the customer's pattern satisfaction with the socks designed with the th group of parameters; represents the customer's material satisfaction with the socks designed with the th group of parameters; represents the customer's function satisfaction with the socks designed with the th group of parameters;

[0077] Further, the sock customized design constraint conditions are:

[0078] ;

[0079] Among them, Represents the constraints for customized sock design; Represents the first sock design satisfaction level; Represents the threshold of sock design satisfaction level; Represents the set of the first customized sock design parameters Any parameter group .

[0080] If the constraints for the customized sock design are not satisfied, parameter adjustment is performed on the unsatisfied parameter group, mainly adjusting each characteristic parameter therein, including color, pattern, size, function, thickness, and material; until the constraints for the customized sock design are satisfied, and the second set of customized sock design parameters is obtained.

[0081] In this embodiment, sock design is performed according to the first set of customized sock design parameters, and the designed socks are recommended to customers. Customer feedback data on the recommended socks is obtained, and the first sock design satisfaction level of the customers is obtained based on the feedback data on the recommended socks. Based on the first sock design satisfaction level and the constraints for the customized sock design, the first set of customized sock design parameters is adjusted to obtain the second set of customized sock design parameters. This method recommends sock design based on the optimized first set of customized sock design parameters, reflects the sock satisfaction level of customers according to the feedback data, and makes effective adjustments based on this satisfaction level, improving the comprehensiveness of the customized sock design, and thus further improving the satisfaction level of customers with the customized sock design.

[0082] To compare the customer satisfaction of a sock customization design method based on an intelligent optimization algorithm proposed in this embodiment with that of a traditional sock design method, multiple comparison experiments are conducted, mainly through Method 1 and Method 2. Method 1 is a sock customization design method based on an intelligent optimization algorithm proposed in this embodiment, and Method 2 is a traditional sock design method. Satisfaction data feedback is respectively obtained from multiple customers, and the average satisfaction comparison data is shown in Table 3 as follows;

[0083] Table 3 Comparison table of customer satisfaction for different sock design methods

[0084]

[0085] As shown in Table 3, a sock customization design method based on an intelligent optimization algorithm proposed in this embodiment shows a certain effectiveness in sock customization design.

[0086] The present invention obtains the basic information data, body data, and historical purchase data of customers; inputs these three types of data into a sock preference prediction model to identify a set of sock preference characteristics and calculates a set of sock preference degrees for customers, including multiple preference degree parameters such as color, size, pattern, function, material, and thickness. Based on the set of sock preference degrees, a first sock design optimization objective function is constructed, and the particle swarm algorithm is used to optimize and analyze the objective function to obtain a first set of sock customization design parameters. Sock design is carried out according to the set of design parameters and recommended to customers, and the design parameters are adjusted according to the feedback data to obtain a second set of sock customization design parameters. This method realizes precise sock customization design through intelligent optimization algorithms, which helps to improve customer satisfaction.

[0087] Embodiment 2

[0088] Sock manufacturer B applied a sock customization design method based on intelligent optimization algorithms to improve customer satisfaction;

[0089] Refer to Figure 1 , which is a schematic flow chart of a sock customization design method provided by an embodiment of the present invention, specifically including:

[0090] S10. Obtain the first basic information data, second body data, and first historical sock purchase data of customers;

[0091] Furthermore, the first basic information data includes age, gender, and occupation; the second body data includes shoe size, height, and body type; the first historical sock purchase data includes multi-element data of each sock purchased by the customer, including color, pattern, size, material, function, and thickness, as well as the evaluation data of the customer for each sock; among them, the evaluation data is a score evaluation of the color, pattern, size, function, thickness, and material of the purchased sock; the full score is 10 points; the body type is analyzed according to the IBM index; the pattern includes geometric patterns, text patterns, letter patterns, anime patterns, and no pattern, etc.; the function includes sweat absorption, warmth retention, and waterproofing, etc.; the material includes wool, cotton, and nylon, etc.; the body type is analyzed according to the IBM index; the shoe size is represented according to the Chinese size; the sock size is represented according to the length of the sock; part of the basic information data and body data are shown in Table 4;

[0092] Table 4 Partial Basic Information Data and Body Data Table

[0093]

[0094] Part of the historical purchase data is shown in Table 5;

[0095] Table 5 Partial Historical Purchase Data Table

[0096]

[0097] S20. Input these three types of data into the sock preference prediction model, and successively pass through the preprocessing layer, preference feature extraction layer, preference degree analysis layer, and output layer of the sock preference prediction model; the preference feature extraction layer analyzes through a deep neural network to obtain a sock preference feature set; the preference degree analysis layer obtains a sock preference degree set of customers according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, and a thickness preference degree set;

[0098] Further, the sock preference prediction model includes a preprocessing layer, a sock preference feature extraction layer, a preference degree analysis layer, and an output layer; the structural schematic diagram of the sock preference prediction model is as Figure 2 shown;

[0099] The preprocessing layer is used to preprocess the first basic information data, the second body state data, and the first historical sock purchase data to obtain second basic information data, third body state data, and second historical sock purchase data; the preprocessing includes data cleaning, normalization, and one-hot encoding; specifically, one-hot encoding is performed on gender, and normalization is performed on age, height, shoe size and size, thickness, and the evaluation data of each sock by the customer;

[0100] The preference feature extraction layer is used to identify the second basic information data, the third body state data, and the second historical sock purchase data through a deep neural network to obtain the sock preference feature set; the sock preference feature set includes a color preference feature set, a size preference feature set, a pattern preference feature set, a material preference feature set, a function preference feature set, and a thickness preference feature set;

[0101] The specific identification process of the sock preference feature set includes: first, convert the occupation in the basic information data, the body type in the body state data, and the color, pattern, material, and function data in the multi-element data of each sock purchased by the customer into vector representations; and splice the preprocessed data into a comprehensive feature vector, and train through a deep neural network to learn the basic information data influence coefficient and body state data influence coefficient of each feature in each preference feature, where the basic information data influence coefficient is used to represent the influence of the customer's basic information data on the preference feature; the body state data influence coefficient is used to represent the influence of the customer's body state data on the preference feature; and output the preference feature set.

[0102] The preference degree analysis layer is used to obtain the sock preference degree set of customers according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, a function preference degree set, and a thickness preference degree set;

[0103] The output layer is used to output the sock preference degree set.

[0104] Furthermore, the sock preference features are used to represent the correlation between the historical purchase data of customers and the basic information data and body data of customers; the sock preference feature set includes preference features, basic information influence coefficients, and body data influence coefficients; the sock preference degree set of each sock preference feature set is obtained according to the basic information influence coefficient and the body data influence coefficient.

[0105] Furthermore, the sock preference feature set is:

[0106] ;

[0107] Among them, represents the sock preference feature set; represents the color preference feature set; represents the size preference feature set; represents the pattern preference feature set; represents the material preference feature set; represents the function preference feature set; represents the thickness preference feature set; represents the th color preference feature in the color preference feature set; represents the th size preference feature in the size preference feature set; represents the th pattern preference feature in the pattern preference feature set; represents the th material preference feature in the material preference feature set; represents the th function preference feature in the function preference feature set; represents the th thickness preference feature in the thickness preference feature set; represents the basic information data influence coefficient of the th color preference feature in the color preference feature set; represents the body data influence coefficient of the th color preference feature in the color preference feature set; represents the Coefficient of influence of basic information data of a size preference feature; Denotes the coefficient of influence of body posture data of the th size preference feature in the set of size preference features; Denotes the coefficient of influence of basic information data of the th pattern preference feature in the set of pattern preference features; Denotes the coefficient of influence of body posture data of the th pattern preference feature in the set of pattern preference features; Denotes the coefficient of influence of basic information data of the th material preference feature in the set of material preference features; Denotes the coefficient of influence of body posture data of the th material preference feature in the set of material preference features; Denotes the coefficient of influence of basic information data of the th function preference feature in the set of function preference features; Denotes the coefficient of influence of body posture data of the th function preference feature in the set of function preference features; Denotes the coefficient of influence of basic information data of the th thickness preference feature in the set of thickness preference features; Denotes the coefficient of influence of body posture data of the th thickness preference feature in the set of thickness preference features.

[0108] Furthermore, the sock preference degree set is:

[0109] ;

[0110] wherein, Denotes the sock preference degree set; Denotes the color preference degree set; Denotes the preference degree of the th color preference feature in the color preference degree set; Denotes the size preference degree set; Denotes the preference degree of the th size preference feature in the size preference degree set; Denotes the pattern preference degree set; Denotes the preference degree of the th pattern preference feature in the pattern preference degree set; Denotes the material preference degree set; Denotes the preference degree of the th material preference feature in the material preference degree set; Denotes the function preference degree set; Denotes the preference degree of the Preference degree of a functional preference feature Represents a set of thickness preference degrees Represents the Preference degree of the nth thickness preference feature in the set of thickness preference degrees Represents the exponential function with the natural constant as the base

[0111] S30. Construct a first sock design optimization objective function according to the sock preference degree set and the preset objective function optimization threshold, perform optimization analysis on the first sock design optimization objective function based on the particle swarm algorithm, obtain multiple groups of sock optimization parameters, and construct a first sock customization design parameter set according to the multiple groups of sock optimization parameters

[0112] Furthermore, the particle swarm optimization parameters are constructed by combining the preference degree values of multiple preference features in the sock preference degree set, including various combination parameters such as color preference degree, pattern preference degree, size preference degree, functional preference degree, thickness preference degree, and material preference degree; and construct a first sock customization design parameter set

[0113] Furthermore, the first sock design optimization objective function is

[0114] ;

[0115] wherein represents the first sock design optimization objective function represents the color preference degree influence weight represents the nth color preference feature preference degree represents the size preference degree influence weight represents the nth size preference feature preference degree represents the pattern preference degree influence weight represents the nth pattern preference feature preference degree represents the material preference degree influence weight represents the nth material preference feature preference degree represents the functional preference degree influence weight represents the nth functional preference feature preference degree represents the thickness preference degree influence weight represents the nth thickness preference feature preference degree represents the preset objective function optimization threshold

[0116] S40. Design socks according to the first set of sock customization design parameters, recommend the designed socks to customers, obtain the feedback data of the recommended socks from customers, obtain the first sock design satisfaction degree of customers according to the feedback data of the recommended socks, and adjust the first set of sock customization design parameters based on the first sock design satisfaction degree and the sock customization design constraint conditions to obtain the second set of sock customization design parameters.

[0117] Further, the feedback data of the recommended socks includes color satisfaction, pattern satisfaction, size satisfaction, function satisfaction, thickness satisfaction, and material satisfaction.

[0118] Further, the first sock design satisfaction degree is:

[0119] ;

[0120] Among them, represents the first sock design satisfaction degree; represents the number of parameter groups in the first set of sock customization design parameters; represents the color satisfaction of the socks designed with the th group of parameters by the customer; represents the size satisfaction of the socks designed with the th group of parameters by the customer; represents the pattern satisfaction of the socks designed with the th group of parameters by the customer; represents the material satisfaction of the socks designed with the th group of parameters by the customer; represents the function satisfaction of the socks designed with the th group of parameters by the customer; represents the thickness satisfaction of the socks designed with the th group of parameters by the customer.

[0121] Further, the sock customization design constraint conditions are:

[0122] ;

[0123] Among them, represents the sock customization design constraint conditions; represents the first sock design satisfaction degree; represents the sock design satisfaction threshold; represents the first set of sock customization design parameters any parameter group in .

[0124] If the sock customization design constraint conditions are not met, parameter adjustment is performed on the parameter groups that do not meet the requirements. Mainly, each characteristic parameter is adjusted, including color, pattern, size, function, thickness, and material; until the sock customization design constraint conditions are met, a second sock customization design parameter set is obtained.

[0125] In order to compare the customer satisfaction of a sock customization design method based on an intelligent optimization algorithm proposed in this embodiment and a traditional sock design method, multiple comparison experiments are conducted. Mainly through Method 1 and Method 2, Method 1 is a sock customization design method based on an intelligent optimization algorithm proposed in this embodiment, and Method 2 is a traditional sock design method. The satisfaction data feedback is collected from multiple customers respectively, and the average satisfaction comparison data is shown in Table 6.

[0126] Table 6 Comparison Table of Customer Satisfaction of Different Sock Design Methods

[0127]

[0128] As shown in Table 6, a sock customization design method based on an intelligent optimization algorithm proposed in this embodiment shows certain effectiveness in sock customization design.

[0129] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sock customization design method based on an intelligent optimization algorithm, characterized in that Including: S10. Obtain the first basic information data, the second body data, and the first historical sock purchase data of the customer; S20. Input these three types of data into the sock preference prediction model, and successively pass through the preprocessing layer, preference feature extraction layer, preference degree analysis layer, and output layer of the sock preference prediction model; The preference feature extraction layer analyzes through a deep neural network to obtain a sock preference feature set; the sock preference features are used to represent the correlation between the customer's historical purchase data and the customer's basic information data and body data; the sock preference feature set includes preference features, basic information influence coefficients, and body data influence coefficients; The preference degree analysis layer obtains the sock preference degree set of the customer according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, and a thickness preference degree set; S30. Optimize the threshold according to the sock preference degree set and the preset objective function, construct the first sock design optimization objective function, perform optimization analysis on the first sock design optimization objective function based on the particle swarm algorithm, obtain multiple groups of sock optimization parameters, and construct the first sock customization design parameter set according to the multiple groups of sock optimization parameters; S40. Perform sock design according to the first sock customization design parameter set, recommend the designed socks to the customer, obtain the recommended sock feedback data of the customer, obtain the first sock design satisfaction of the customer according to the recommended sock feedback data, and adjust the first sock customization design parameter set based on the first sock design satisfaction and the sock customization design constraint conditions to obtain the second sock customization design parameter set.

2. The sock customization design method based on an intelligent optimization algorithm according to claim 1, characterized in that: The first basic information data includes age, gender, and occupation; the second body data includes shoe size, height, and body type; the first historical sock purchase data includes multi-element data of each sock purchased by the customer, including color, pattern, size, material, function, and thickness, as well as the evaluation data of the customer for each sock.

3. The sock customization design method based on an intelligent optimization algorithm according to claim 1, characterized in that: The sock preference prediction model includes a preprocessing layer, a sock preference feature extraction layer, a preference degree analysis layer, and an output layer; The preprocessing layer is used to preprocess the first basic information data, the second body data, and the first historical sock purchase data to obtain the second basic information data, the third body data, and the second historical sock purchase data; the preprocessing includes data cleaning, normalization, and one-hot encoding; The preference feature extraction layer is used to identify the second basic information data, the third body data, and the second historical sock purchase data through a deep neural network to obtain the sock preference feature set; the sock preference feature set includes a color preference feature set, a size preference feature set, a pattern preference feature set, a material preference feature set, a function preference feature set, and a thickness preference feature set; The preference degree analysis layer is used to obtain the sock preference degree set of customers according to the sock preference feature set, and the sock preference degree set includes a color preference degree set, a size preference degree set, a pattern preference degree set, a material preference degree set, a function preference degree set, and a thickness preference degree set; The output layer is used to output the sock preference degree set.

4. The method for custom sock design based on an intelligent optimization algorithm according to claim 1, wherein: The sock preference degree sets of each sock preference feature set are obtained according to the basic information influence coefficient and the body posture data influence coefficient.

5. The method for custom sock design based on an intelligent optimization algorithm according to claim 1, wherein: The sock preference feature set is: Among them, represents the set of sock preference features; represents the set of color preference features; represents the set of size preference features; represents the set of pattern preference features; represents the set of material preference features; represents the set of function preference features; represents the set of thickness preference features; represents the th color preference feature in the set of color preference features; represents the th size preference feature in the set of size preference features; represents the th pattern preference feature in the set of pattern preference features; represents the th material preference feature in the set of material preference features; represents the th function preference feature in the set of function preference features; represents the th thickness preference feature in the set of thickness preference features; represents the basic information data influence coefficient of the th color preference feature in the set of color preference features; represents the body posture data influence coefficient of the th color preference feature in the set of color preference features; represents the basic information data influence coefficient of the th size preference feature in the set of size preference features; represents the th size preference feature in the set of size preference features; represents the basic information data influence coefficient of the th pattern preference feature in the set of pattern preference features; represents the body posture data influence coefficient of the th pattern preference feature in the set of pattern preference features; represents the basic information data influence coefficient of the th material preference feature in the set of material preference features; represents the th material preference feature in the set of material preference features; represents the basic information data influence coefficient of the th function preference feature in the set of function preference features; represents the The influence coefficient of body posture data for a functional preference feature; Indicates the basic information data influence coefficient for the th thickness preference feature in the thickness preference feature set; Indicates the body posture data influence coefficient for the th thickness preference feature in the thickness preference feature set.

6. The method for custom sock design based on an intelligent optimization algorithm according to claim 5, wherein: The sock preference degree set is: ； Among them, represents the sock preference set; represents the color preference set; represents the preference degree of the th color preference feature in the color preference set; represents the size preference set; represents the preference degree of the th size preference feature in the size preference set; represents the pattern preference set; represents the preference degree of the th pattern preference feature in the pattern preference set; represents the material preference set; represents the preference degree of the th material preference feature in the material preference set; represents the function preference set; represents the preference degree of the th function preference feature in the function preference set; represents the thickness preference set; represents the preference degree of the th thickness preference feature in the thickness preference set; represents the exponential function with the natural constant as the base.

7. The method for custom sock design based on an intelligent optimization algorithm according to claim 1, wherein: The first sock design optimization objective function is: ； Among them, represents the first sock product design optimization objective function; represents the influence weight of color preference degree; represents the preference degree of the th color preference feature; represents the influence weight of size preference degree; represents the preference degree of the th size preference feature; represents the influence weight of pattern preference degree; represents the preference degree of the th pattern preference feature; represents the influence weight of material preference degree; represents the preference degree of the th material preference feature; represents the influence weight of thickness preference degree; represents the preference degree of the th thickness preference feature; represents the optimization threshold of the preset objective function.

8. The method for custom sock design based on an intelligent optimization algorithm according to claim 1, wherein: The recommended sock feedback data includes color satisfaction, pattern satisfaction, size satisfaction, function satisfaction, thickness satisfaction, and material satisfaction.

9. The method for custom sock design based on an intelligent optimization algorithm according to claim 1, wherein: The first sock design satisfaction is: ； Among them, represents the first sock design satisfaction; represents the number of parameters in the first sock customization design parameter set; represents the customer's satisfaction with the color of the socks designed with the th group of parameters; represents the customer's satisfaction with the size of the socks designed with the th group of parameters; represents the customer's satisfaction with the pattern of the socks designed with the th group of parameters; represents the customer's satisfaction with the material of the socks designed with the th group of parameters; represents the customer's satisfaction with the function of the socks designed with the th group of parameters; represents the customer's satisfaction with the thickness of the socks designed with the th group of parameters.

10. The method for custom sock design based on an intelligent optimization algorithm according to claim 9, wherein: The sock custom design constraint conditions are: ； Among them, represents the sock customization design constraint conditions; represents the first sock design satisfaction degree; represents the sock design satisfaction threshold; represents any parameter group in the first sock customization design parameter set .

Images