An AI-driven skin care solution management system and method

Abstract

This invention discloses an AI-driven skincare solution management system and method, relating to the field of dynamic skincare solution adjustment technology. The method includes the following steps: collecting historical and real-time skincare data; performing data preprocessing; using the preprocessed data as a benchmark, and combining it with preset initial skincare indicator weights, calculating an initial skin condition assessment value; dynamically updating the contribution of skincare indicators based on skin condition changes over several periods and the corresponding initial skin condition assessment value, thereby updating the weights of the skincare indicators and calculating the current skin condition assessment value in real time; comparing with preset solution adjustment trigger conditions, recommending product adjustments when the trigger conditions are exceeded, and adjusting the usage frequency based on the change in the skin condition assessment value. This method employs a dynamic weight update mechanism to continuously optimize the impact of each indicator on the skin condition assessment value, achieving personalized and iterative skincare condition modeling.

Description

Technical Field

[0001] This invention relates to the field of dynamic skincare regimen adjustment technology, specifically an AI-driven skincare regimen management system and method. Background Technology

[0002] As people's living standards improve, skincare is receiving increasing attention. Traditional skincare solutions are often based on experience or general skin type classifications, lacking personalization and precision. However, everyone's skin condition and needs are unique, influenced by a combination of factors;

[0003] Meanwhile, continuous technological advancements have brought new opportunities to the skincare field. Various skincare devices can measure basic physiological indicators of the skin, and the widespread use of smartphones and various apps allows users to easily record and upload facial photos, lifestyle data, and other information. However, traditional methods lack integrated analysis of multi-source data such as lifestyle habits and environmental factors, making it difficult to comprehensively reflect the dynamic changes in skin condition. Existing assessment models mostly use fixed weights to calculate skin condition, failing to dynamically adjust the importance of indicators based on individual differences or the evolution of skin condition, resulting in significant discrepancies between assessment results and actual needs. Most systems rely on historical data to generate one-time skincare recommendations, without incorporating real-time device data for dynamic adjustments, making it difficult to address the impact of environmental or behavioral changes on the skin. Summary of the Invention

[0004] The purpose of this invention is to provide an AI-driven skincare solution management system and method to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an AI-driven skincare plan management method, wherein the skincare plan management method specifically includes the following steps:

[0006] Collect historical and real-time skincare data;

[0007] The skincare history data includes basic skincare data and user feedback data;

[0008] The real-time skincare data includes data from one or more skincare devices.

[0009] Regularly update basic skincare data and user feedback data;

[0010] Real-time data collection from one or more of the aforementioned skincare devices;

[0011] The collected historical and real-time skincare data are preprocessed.

[0012] Based on the preprocessed historical skincare data and real-time skincare data, and combined with the preset initial weights of each skincare indicator, an initial skin condition assessment value is calculated.

[0013] Based on the changes in skin condition over several cycles and the corresponding initial skin condition assessment value, the contribution of each skin care indicator is dynamically updated, and the weight of each skin care indicator is updated accordingly, and the current skin condition assessment value is calculated in real time.

[0014] Among them, the contribution and weight of each skin care indicator are positively correlated;

[0015] Based on the real-time calculated current skin condition assessment value, the system compares it with the preset adjustment trigger conditions. If the skin condition is exceeded, a product adjustment recommendation is made.

[0016] The frequency of use is adjusted based on changes in skin condition assessment values.

[0017] The collection of historical and real-time skincare data specifically includes the following skincare indicators:

[0018] Skin moisture content: Using a high-precision skin moisture meter, the moisture content of the stratum corneum of the skin is measured through capacitive or resistive principles;

[0019] As a preferred option, users can use a portable device at home, attaching the probe to the skin surface, and the device will automatically record data and synchronize it to the APP;

[0020] Sebum secretion: A sebum meter is used to measure the amount of sebum secreted on the skin surface by utilizing the reflection or absorption characteristics of sebum to specific wavelengths of light; measurements are taken at a certain time after cleansing, and data from different areas are recorded periodically;

[0021] Skin pH level: Using pH test strips or a professional pH meter, gently touch the test strip to the skin or use the meter probe to touch the skin to understand the skin's acid-base balance. Periodic measurements are required.

[0022] Facial image data: Users take facial photos through a specific shooting interface of the mobile APP under fixed light source and fixed distance. The APP can provide shooting guidance, such as prompting users to keep their face clean and their expression natural to ensure photo quality.

[0023] Lifestyle data: By connecting with health management apps, we can obtain data such as users' daily sleep time, exercise duration, and dietary intake;

[0024] Environmental data: Access the meteorological department's public data interface to obtain real-time ambient temperature, humidity, and ultraviolet intensity data;

[0025] The collected historical and real-time skincare data are preprocessed, specifically as follows:

[0026] The data preprocessing includes data cleaning and data standardization;

[0027] Preferably, the data cleaning includes outlier handling, specifically the following steps:

[0028] Step S2-1-1: Perform outlier processing on the collected skincare history data and sort the data of the same skincare indicator in ascending order; calculate the upper quartile, median, and lower quartile of the ascending sequence; the upper quartile, median, and lower quartile of the ascending sequence are the data values ​​corresponding to the upper quartile Q1, median Q2, and lower quartile Q3, respectively. ; ; Where m represents the number of data points in the ascending sequence, and m is a positive integer; Q1 and Q2 are rounded up; Q3 is rounded down;

[0029] Step S2-1-2: Calculate the difference between the upper quartile and the lower quartile; denote the absolute value of the difference between the upper quartile and the lower quartile as IQR;

[0030] Step S2-1-3: Using the upper quartile as the upper limit and the lower quartile as the lower limit, data that are greater than the upper quartile plus 1.5 times the IQR and data that are less than the lower quartile minus 1.5 times the IQR are classified as outliers.

[0031] Step S2-1-4: Replace outliers greater than the upper quartile plus 1.5 times the IQR with the upper quartile plus 1.5 times the IQR, and replace outliers less than the lower quartile minus 1.5 times the IQR with the lower quartile minus 1.5 times the IQR.

[0032] The data standardization specifically refers to:

[0033] For numerical data, data standardization is performed using the Max-Min method, x norm =(xx min ) / (x max -x min ), where x norm x represents the standardized data value; x represents the original data value; x max and x min These represent the maximum and minimum values ​​of the same skincare indicator data, respectively.

[0034] For categorical data, one-hot encoding is used to convert it into a numerical vector;

[0035] Based on historical data, establish a mapping relationship between the numerical vector and the score;

[0036] Based on preprocessed historical and real-time skincare data, and combined with pre-defined initial weights for each skincare indicator, an initial skin condition assessment value is calculated, specifically:

[0037] ;

[0038] Where S represents the skin condition assessment value; n represents the number of skin care indicators, where n is a positive integer; w i (0) represents the preset initial weights of each skincare indicator, and i represents the category label of the skincare indicator; x norm This represents the standardized data value, where, for categorical data, x norm This represents the mapping score corresponding to the numerical vector;

[0039] The weights of the preset initial skincare indicators are allocated equally.

[0040] Based on historical skincare data showing changes in skin condition over several periods and the corresponding initial skin condition assessment values, the contribution of each skincare indicator is dynamically updated, thereby updating the weight of each skincare indicator and calculating the current skin condition assessment value in real time. Specifically:

[0041] Step S1: Using association analysis, calculate the correlation coefficient between changes in skincare indicators and skin condition assessment values ​​over several periods. Preferably, the Pearson correlation coefficient formula is used.

[0042] ;

[0043] Where, r i This represents the correlation coefficient between changes in skincare indicators and skin condition assessment values; m represents the number of historical data points, where m is a positive integer; j represents the number of historical data points and their labels. This represents the change in the i-th skincare indicator within the j-th historical data set. S represents the average value of the change in the i-th skincare indicator; j This represents the skin condition assessment value in the j-th historical data; This represents the average value of the skin condition assessment.

[0044] Step S2: Based on the correlation coefficient r i Calculate the contribution C of each skincare indicator. i The characterization formula is:

[0045] ;

[0046] Among them, C i Indicates the contribution of skincare indicators;

[0047] Step S3: Based on the calculated contribution C of each skincare indicator i The weights of each skincare indicator are updated, and the iterative formula for weight updates is as follows:

[0048] ;

[0049] Among them, w i (t) represents the weight of each skincare indicator after t iterations; w i (t-1) represents the weight of each skincare indicator in iteration t-1; γ represents the adjustment coefficient; C i (t-1) represents the contribution of the skin care index at iteration t-1;

[0050] When the contribution of a certain skincare indicator increases, its weight increases accordingly; conversely, when the contribution decreases, its weight decreases. By continuously updating the weights based on the new contribution, the calculation method of the skin condition assessment value can more accurately reflect the impact of each skincare indicator on the overall skin condition, thereby providing users with more precise skincare adjustment suggestions.

[0051] Based on the real-time calculated current skin condition assessment value, the system compares it with the preset adjustment trigger conditions. If the skin condition is exceeded, a product adjustment recommendation is made, specifically:

[0052] Step S1: Set a skin condition assessment threshold. When the change in the skin condition assessment value S, ΔS, exceeds the skin condition assessment threshold T, or the change in a single skin care indicator exceeds the corresponding threshold T... i When |ΔS| > T or |ΔS i |>T i At that time, product adjustment recommendations will be initiated;

[0053] Step S2: Establish a correlation matrix M between the efficacy of historically used products and skincare indicators. The elements M in the correlation matrix M... pi This represents the impact of product p on skincare metric i, calculated from historical data on the values ​​x after using the product. norm The change in value is expressed as a measure;

[0054] Step S3: Based on the changes in skin condition assessment values ​​and the correlation matrix M, calculate the correlation between the changes in skincare indicator assessment values ​​and the product's impact on the skincare indicator values, and make product adjustment recommendations. The formula is:

[0055] ;

[0056] Among them, P best Indicates recommended historical products; ΔS iThis represents the change in the evaluation value of the i-th skincare indicator; p represents the product number.

[0057] Preferably, the frequency of use is adjusted periodically based on changes in skin condition assessment values, specifically as follows:

[0058] ;

[0059] Among them, f new Indicates the adjusted usage frequency; f old Indicates the usage frequency before adjustment; k represents the adjustment coefficient; S max This represents the maximum value of the skin condition assessment.

[0060] An AI-driven skincare solution management system, comprising a data acquisition module, a data preprocessing module, a simulation calculation module, and an AI-driven module;

[0061] The data acquisition module is responsible for collecting historical skincare data and real-time skincare data. The historical skincare data includes basic skincare data and user feedback data; the real-time skincare data includes data from several skincare devices.

[0062] The data from the skincare devices is collected and stored via an app;

[0063] Regularly update basic skincare data and user feedback data;

[0064] Real-time updates of skincare device data;

[0065] The data preprocessing module is responsible for performing data preprocessing; the data preprocessing includes data cleaning and data standardization; for numerical data, data standardization is performed using the Max-Min method; for categorical data, one-hot encoding is used to convert it into a numerical vector.

[0066] The simulation calculation module is used to calculate the initial skin condition assessment value based on the preprocessed skin care historical data and real-time skin care data, combined with the preset weights of the initial skin care indicators; and dynamically updates the contribution of the skin care indicators based on the changes in skin condition over several periods of skin care historical data and the corresponding initial skin condition assessment value, thereby updating the weights of the skin care indicators and calculating the current skin condition assessment value in real time.

[0067] The AI-driven module is used to compare the current skin condition assessment value calculated in real time with the preset scheme adjustment trigger conditions. When the scheme adjustment trigger conditions are exceeded, product adjustment recommendations are made, and the usage frequency is adjusted according to the change in the skin condition assessment value.

[0068] The simulation calculation module includes a data storage unit, a skin condition assessment unit, and an index weight iteration unit;

[0069] The data storage unit is used to store the data that has undergone data preprocessing;

[0070] The skin condition assessment unit is used to calculate the skin condition assessment value based on the preprocessed historical skin care data and real-time skin care data.

[0071] The indicator weight iteration unit is used to dynamically update the contribution of skin care indicators based on the changes in skin condition over several periods of skin care historical data and the corresponding initial skin condition assessment value, thereby updating the weight of skin care indicators.

[0072] The AI-driven module includes a threshold control unit, an association matrix unit, a product recommendation unit, and a frequency update unit;

[0073] The threshold control unit is used to obtain the skin condition assessment threshold based on historical data statistics;

[0074] The association matrix unit is used to construct an association matrix M based on the historical efficacy of products used and skincare indicators;

[0075] The product recommendation unit is used to make product adjustment recommendations by calculating the correlation between the change in the skin care index evaluation value and the product's impact on the skin care index value.

[0076] The frequency update unit is used to adjust the usage frequency based on the change in skin condition assessment values.

[0077] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention constructs a comprehensive skin condition assessment system by integrating historical and real-time skincare data, combined with multi-dimensional skincare indicators such as moisture, oil, pH value, images, lifestyle habits, and environment. This method improves data quality through data preprocessing and employs a dynamic weight update mechanism, combined with Pearson correlation coefficient and contribution calculation, to continuously optimize the impact of each indicator on the skin condition assessment value, achieving personalized and iterative skincare condition modeling. Based on changes in skin condition and combined with a historical product effect matrix, the system intelligently recommends suitable products and dynamically adjusts the usage frequency, thereby improving the accuracy and effectiveness of skincare solutions. Simultaneously, this method supports simultaneous data collection via portable devices and an app, enhancing user convenience, and has the ability to interface with health management and environmental data, exhibiting strong scalability and suitability for building intelligent, closed-loop personalized skincare management systems. Attached Figure Description

[0078] Figure 1 This is a schematic diagram illustrating the steps of an AI-driven skincare solution management method according to the present invention.

[0079] Figure 2 This is a structural diagram of an AI-driven skincare solution management system according to the present invention. Detailed Implementation

[0080] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0081] Example: Figure 1-Figure 2 As shown, this invention provides a technical solution: an AI-driven skincare plan management method, which specifically includes the following steps:

[0082] Collect historical and real-time skincare data;

[0083] Skincare history data includes basic skincare data and user feedback data;

[0084] Real-time skincare data includes data from one or more skincare devices;

[0085] Regularly update basic skincare data and user feedback data;

[0086] Real-time data collection from one or more skincare devices;

[0087] Data preprocessing is performed on the collected historical and real-time skincare data;

[0088] Based on the preprocessed historical skincare data and real-time skincare data, and combined with the preset initial weights of each skincare indicator, an initial skin condition assessment value is calculated.

[0089] Based on the changes in skin condition over several cycles and the corresponding initial skin condition assessment values ​​based on historical skin care data, the contribution of each skin care indicator is dynamically updated, and the weight of each skin care indicator is updated accordingly, and the current skin condition assessment value is calculated in real time.

[0090] Among them, the contribution and weight of each skin care indicator show a positive correlation;

[0091] Based on the real-time calculated current skin condition assessment value, the system compares the preset plan adjustment trigger conditions and recommends product adjustments when the conditions are exceeded.

[0092] The frequency of use is adjusted based on changes in skin condition assessment values.

[0093] This includes collecting historical and real-time skincare data, specifically the following skincare indicators:

[0094] Skin moisture content: Using a high-precision skin moisture meter, the moisture content of the stratum corneum of the skin is measured through capacitive or resistive principles;

[0095] As a preferred option, users can use a portable device at home, attaching the probe to the skin surface, and the device will automatically record data and synchronize it to the APP;

[0096] Sebum secretion: A sebum meter is used to measure the amount of sebum secreted on the skin surface by utilizing the reflection or absorption characteristics of sebum to specific wavelengths of light; measurements are taken at a certain time (2 hours) after cleansing, and data from different areas are recorded periodically;

[0097] Skin pH level: Using pH test strips or a professional pH meter, gently touch the test strip to the skin or use the meter probe to touch the skin to understand the skin's acid-base balance. Periodic measurements are required.

[0098] Facial image data: Users take facial photos through a specific shooting interface of the mobile APP under fixed light source and fixed distance. The APP can provide shooting guidance, such as prompting users to keep their face clean and their expression natural to ensure photo quality.

[0099] Lifestyle data: By connecting with health management apps (such as fitness trackers and sleep monitoring apps), data such as users' daily sleep time, exercise duration, and dietary intake (such as sugar, fat, and vitamin intake) can be obtained.

[0100] Environmental data: Access the meteorological department's public data interface to obtain real-time ambient temperature, humidity, and ultraviolet intensity data;

[0101] The collected historical and real-time skincare data are preprocessed, specifically as follows:

[0102] Data preprocessing includes data cleaning and data standardization;

[0103] Data cleaning includes outlier handling, and the specific steps are as follows:

[0104] Step S2-1-1: Perform outlier processing on the collected skincare history data and sort the data of the same skincare indicator in ascending order; calculate the upper quartile, median, and lower quartile of the ascending sequence; the upper quartile, median, and lower quartile of the ascending sequence are the data values ​​corresponding to the upper quartile Q1, median Q2, and lower quartile Q3, respectively. ; ; Where m represents the number of data points in the ascending sequence, and m is a positive integer; Q1 and Q2 are rounded up; Q3 is rounded down;

[0105] Step S2-1-2: Calculate the difference between the upper quartile and the lower quartile; denote the absolute value of the difference between the upper quartile and the lower quartile as IQR;

[0106] Step S2-1-3: Using the upper quartile as the upper limit and the lower quartile as the lower limit, data that are greater than the upper quartile plus 1.5 times the IQR and data that are less than the lower quartile minus 1.5 times the IQR are classified as outliers.

[0107] Step S2-1-4: Replace outliers greater than the upper quartile plus 1.5 times the IQR with the upper quartile plus 1.5 times the IQR, and replace outliers less than the lower quartile minus 1.5 times the IQR with the lower quartile minus 1.5 times the IQR.

[0108] Data standardization specifically refers to:

[0109] For numerical data, data standardization is performed using the Max-Min method, x norm =(xx min ) / (x max -x min ), where x norm x represents the standardized data value; x represents the original data value; x max and x min These represent the maximum and minimum values ​​of the same skincare indicator data, respectively.

[0110] For categorical data, one-hot encoding is used to convert it into a numerical vector;

[0111] Based on historical data, establish a mapping relationship between numerical vectors and scores;

[0112] One-hot encoding of categorical data converts non-numerical categorical data (skin type) into numerical vectors, making it easier for machine learning models to process.

[0113] The original category data are: ["dry", "oily", "combination", "sensitive"];

[0114] After one-hot encoding, it is converted into a vector:

[0115] Dry: [1,0,0,0]; Oily: [0,1,0,0];

[0116] Mixedness: [0,0,1,0]; Sensitivity: [0,0,0,1];

[0117] Establishing the mapping relationship between numerical vectors and scores:

[0118] By analyzing historical data, we can understand the impact of category features on skin condition assessment scores and assign corresponding scoring weights to different categories.

[0119] Based on statistical analysis of user feedback data, a mapping relationship between numerical vectors and ratings is established, and the ratings are normalized to the [0,1] interval;

[0120]

[0121] Table 1

[0122] Using machine learning to solve for the regression coefficients corresponding to different skin types:

[0123]

[0124] Table 2

[0125] In Table 2, if the user has dry skin, the weight of the skin moisture index is increased to 0.72 times the currently detected moisture value.

[0126] For users with oily skin, the sebum secretion weight is 0.55 times the current sebum value;

[0127] Based on preprocessed historical and real-time skincare data, and combined with pre-defined initial weights for each skincare indicator, an initial skin condition assessment value is calculated, specifically:

[0128] ;

[0129] Where S represents the skin condition assessment value; n represents the number of skin care indicators, where n is a positive integer; w i (0) represents the preset initial weights of each skincare indicator, and i represents the category label of the skincare indicator; x norm This represents the standardized data value, where, for categorical data, x norm This represents the mapping score corresponding to the numerical vector;

[0130] The weights of the initial skincare indicators are allocated equally.

[0131] Based on historical skincare data showing changes in skin condition over several periods and corresponding initial skin condition assessment values, the contribution of each skincare indicator is dynamically updated, thereby updating the weight of each skincare indicator and calculating the current skin condition assessment value in real time. Specifically:

[0132] Step S1: Using association analysis, calculate the correlation coefficient between changes in skincare indicators and skin condition assessment values ​​over several periods. Preferably, the Pearson correlation coefficient formula is used.

[0133] ;

[0134] Where, r i This represents the correlation coefficient between changes in skincare indicators and skin condition assessment values; m represents the number of historical data points, where m is a positive integer; j represents the number of historical data points and their labels. This represents the change in the i-th skincare indicator within the j-th historical data set. S represents the average value of the change in the i-th skincare indicator; j This represents the skin condition assessment value in the j-th historical data; This represents the average value of the skin condition assessment.

[0135] Step S2: Based on the correlation coefficient r i Calculate the contribution C of each skincare indicator. i The characterization formula is:

[0136] ;

[0137] Among them, C i Indicates the contribution of skincare indicators;

[0138] Step S3: Based on the calculated contribution C of each skincare indicator i The weights of each skincare indicator are updated, and the iterative formula for weight updates is as follows:

[0139] ;

[0140] Among them, w i (t) represents the weight of each skincare indicator after t iterations; w i (t-1) represents the weight of each skincare indicator in iteration t-1; γ represents the adjustment coefficient; C i (t-1) represents the contribution of the skin care index at iteration t-1;

[0141] When the contribution of a certain skincare indicator increases, its weight increases accordingly; conversely, when the contribution decreases, its weight decreases. By continuously updating the weights based on the new contribution, the calculation method of the skin condition assessment value can more accurately reflect the impact of each skincare indicator on the overall skin condition, thereby providing users with more precise skincare adjustment suggestions.

[0142] Based on the real-time calculated assessment value of the current skin condition, the system compares this value with the preset adjustment trigger conditions. If the conditions are exceeded, product adjustment recommendations are made, specifically:

[0143] Step S1: Set a skin condition assessment threshold. When the change in the skin condition assessment value S, ΔS, exceeds the skin condition assessment threshold T, or the change in a single skin care indicator exceeds the corresponding threshold T... iWhen |ΔS| > T or |ΔS i |>T i At that time, product adjustment recommendations will be initiated;

[0144] Historical data statistics: Collect data on the changes in the overall skin condition score of a large number of users during normal skin care processes over a period of time (one month); analyze this data and calculate its mean μ and standard deviation σ;

[0145] Determine the threshold range: Based on historical statistical patterns, 95% of the data falls within the range of the mean 2σ, and approximately 99.7% of the data falls within the range of the mean 3σ;

[0146] Set the threshold T range to μ+3σ; if the value is exceeded, the scheme adjustment process will be initiated.

[0147] The threshold for a specific skincare indicator was obtained using the same method.

[0148] Step S2: Establish a correlation matrix M between the efficacy of historically used products and skincare indicators. The elements M in the correlation matrix M... pi This represents the impact of product p on skincare metric i, calculated from historical data on the values ​​x after using the product. norm The change in value is expressed as a measure;

[0149] Step S3: Based on the changes in skin condition assessment values ​​and the correlation matrix M, calculate the correlation between the changes in skincare indicator assessment values ​​and the product's impact on these indicators, and make product adjustment recommendations. The formula is:

[0150] ;

[0151] Among them, P best Indicates recommended historical products; ΔS i This represents the change in the evaluation value of the i-th skincare indicator; p represents the product number.

[0152] Preferably, the frequency of use is adjusted periodically based on changes in skin condition assessment values, specifically as follows:

[0153] ;

[0154] Among them, f new Indicates the adjusted usage frequency; f old Indicates the usage frequency before adjustment; k represents the adjustment coefficient; S max This represents the maximum value of the skin condition assessment.

[0155] An AI-driven skincare solution management system, comprising a data acquisition module, a data preprocessing module, a simulation calculation module, and an AI-driven module;

[0156] The data acquisition module is responsible for collecting historical and real-time skincare data. Historical skincare data includes basic skincare data and user feedback data; real-time skincare data includes data from several skincare devices.

[0157] Skincare device data is collected and stored via an app;

[0158] Regularly update basic skincare data and user feedback data;

[0159] Real-time updates of skincare device data;

[0160] The data preprocessing module is responsible for performing data preprocessing, which includes data cleaning and data standardization. For numerical data, data standardization is performed using the Max-Min method. For categorical data, one-hot encoding is used to convert it into numerical vectors.

[0161] The simulation calculation module is used to calculate the initial skin condition assessment value based on the preprocessed skin care historical data and real-time skin care data, combined with the preset weights of the initial skin care indicators; and based on the changes in skin condition over several periods of skin care historical data and the corresponding initial skin condition assessment value, it dynamically updates the contribution of skin care indicators, thereby updating the weights of skin care indicators and calculating the current skin condition assessment value in real time.

[0162] The AI-driven module is used to compare the current skin condition assessment value calculated in real time with the preset plan adjustment trigger conditions. When the plan adjustment trigger conditions are exceeded, product adjustment recommendations are made, and the usage frequency is adjusted according to the change in the skin condition assessment value.

[0163] The simulation calculation module includes a data storage unit, a skin condition assessment unit, and an indicator weight iteration unit;

[0164] The data storage unit is used to store the data that has undergone preprocessing.

[0165] The skin condition assessment unit is used to calculate the skin condition assessment value based on the preprocessed historical skin care data and real-time skin care data.

[0166] The indicator weight iteration unit is used to dynamically update the contribution of skin care indicators based on the changes in skin condition over several periods and the corresponding initial skin condition assessment values, thereby updating the weight of skin care indicators.

[0167] The AI-driven module includes a threshold control unit, an association matrix unit, a product recommendation unit, and a frequency update unit;

[0168] The threshold control unit is used to obtain the skin condition assessment threshold based on historical data statistics;

[0169] The association matrix unit is used to construct an association matrix M based on the historical efficacy of products used and skincare indicators;

[0170] The product recommendation unit is used to make product adjustment recommendations by calculating the correlation between the change in the skin care index evaluation value and the product's impact on the skin care index value.

[0171] The frequency update unit is used to adjust the usage frequency based on changes in skin condition assessment values.

[0172] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An AI-driven skincare regimen management method, characterized in that: The skincare regimen management method specifically includes the following steps: Collect historical and real-time skincare data; The collected historical and real-time skincare data are preprocessed; the data preprocessing includes data cleaning and data standardization. The data cleaning includes outlier handling, and the specific steps are as follows: Step S2-1-1: Perform outlier processing on the collected skincare history data and sort the data of the same skincare indicator in ascending order; calculate the upper quartile, median, and lower quartile of the ascending sequence; the upper quartile, median, and lower quartile of the ascending sequence are the data values ​​corresponding to the upper quartile Q1, median Q2, and lower quartile Q3, respectively. ; ; Where m represents the number of data points in the ascending sequence, and m is a positive integer; Q1 and Q2 are rounded up; Q3 is rounded down; Step S2-1-2: Calculate the difference between the upper quartile and the lower quartile; denote the absolute value of the difference between the upper quartile and the lower quartile as IQR; Step S2-1-3: Using the upper quartile as the upper limit and the lower quartile as the lower limit, data that are greater than the upper quartile plus 1.5 times the IQR and data that are less than the lower quartile minus 1.5 times the IQR are classified as outliers. Step S2-1-4: Replace outliers greater than the upper quartile plus 1.5 times the IQR with the upper quartile plus 1.5 times the IQR, and replace outliers less than the lower quartile minus 1.5 times the IQR with the lower quartile minus 1.5 times the IQR. The data standardization specifically refers to: For numerical data, data standardization is performed using the Max-Min method, x norm =(xx min ) / (x max -x min ), where x norm x represents the standardized data value; x represents the original data value; x max and x min These represent the maximum and minimum values ​​of the same skincare indicator data, respectively. For categorical data, one-hot encoding is used to convert it into a numerical vector; Based on preprocessed historical and real-time skincare data, and combined with the weights of preset initial skincare indicators, an initial skin condition assessment value is calculated, specifically: ； Where S represents the skin condition assessment value; n represents the number of skin care indicators, where n is a positive integer; w i (0) represents the preset initial weights of each skincare indicator, and i represents the category label of the skincare indicator; x norm This represents the standardized data value, where, for categorical data, x norm This represents the mapping score corresponding to the numerical vector; Based on historical skincare data showing changes in skin condition over several periods and the corresponding initial skin condition assessment values, the contribution of skincare indicators is dynamically updated, thereby updating the weights of these indicators. The current skin condition assessment value is calculated in real time. Specifically: Step S6-1: Using association analysis, calculate the correlation coefficient between changes in skincare indicators and skin condition assessment values ​​over several periods, using the Pearson correlation coefficient formula: ； Where, r i This represents the correlation coefficient between changes in skincare indicators and skin condition assessment values; m represents the number of historical data points, where m is a positive integer; j represents the number of historical data points and their labels. This represents the change in the i-th skincare indicator within the j-th historical data set. S represents the average value of the change in the i-th skincare indicator; j This represents the skin condition assessment value in the j-th historical data; This represents the average value of the skin condition assessment. Step S6-2: Based on the correlation coefficient r i Calculate the contribution C of each skincare indicator. i The characterization formula is: ； Among them, C i Indicates the contribution of skincare indicators; Step S6-3: Based on the calculated contribution C of each skincare indicator i The weights of each skincare indicator are updated, and the iterative formula for weight updates is as follows: ； Among them, w i (t) represents the weight of each skincare indicator after t iterations; w i (t-1) represents the weight of each skincare indicator in iteration t-1; γ represents the adjustment coefficient; C i (t-1) represents the contribution of the skin care index at iteration t-1; Based on real-time calculated skin condition assessment values, the system compares these values ​​with preset adjustment trigger conditions. If the skin condition assessment conditions are exceeded, product adjustment recommendations are made, and the usage frequency is adjusted according to the change in skin condition assessment values. Specifically: Step S7-1: Set a skin condition assessment threshold. When the change in the skin condition assessment value S, ΔS, exceeds the skin condition assessment threshold T, or the change in a single skin care indicator exceeds the corresponding threshold T... i When |ΔS| > T or |ΔS i |>T i At that time, product adjustment recommendations will be initiated; Step S7-2: Establish a correlation matrix M between the efficacy of historically used products and skincare indicators. The elements M in the correlation matrix M... pi This represents the impact of product p on skincare metric i, calculated from historical data on the values ​​x after using the product. norm The change in value is expressed as a measure; Step S7-3: Based on the changes in skin condition assessment values ​​and the aforementioned correlation matrix M, calculate the correlation between the changes in skincare indicator assessment values ​​and the product's impact on these indicators, and then make product adjustment recommendations. The formula is: ； Among them, P best Indicates recommended historical products; ΔS i This represents the change in the evaluation value of the i-th skincare indicator; p represents the product number. The frequency of use should be adjusted periodically based on changes in skin condition assessment values, specifically as follows: ； Among them, f new Indicates the adjusted usage frequency; f old Indicates the usage frequency before adjustment; k represents the adjustment coefficient; S max This represents the maximum value of the skin condition assessment.

2. The AI-driven skincare management method according to claim 1, characterized in that: Collect historical and real-time skincare data, specifically: The skincare history data includes basic skincare data and user feedback data; The real-time skincare data includes skincare device data; Regularly update basic skincare data and user feedback data; Data from the skincare device is collected in real time.

3. The AI-driven skincare management method according to claim 2, characterized in that: Specifically: Based on the preprocessed historical skincare data and real-time skincare data, and combined with the weights of the preset initial skincare indicators, the initial skin condition assessment value is calculated. The weights of the preset initial skincare indicators are distributed equally.

4. An AI-driven skincare solution management system, employing the AI-driven skincare solution management method as described in any one of claims 1-3, characterized in that: The skincare solution management system includes a data acquisition module, a data preprocessing module, a simulation calculation module, and an AI-driven module; The data acquisition module is responsible for collecting historical skincare data and real-time skincare data. The historical skincare data includes basic skincare data and user feedback data; the real-time skincare data includes data from several skincare devices. The data from the skincare devices is collected and stored via an app; Regularly update basic skincare data and user feedback data; Real-time updates of skincare device data; The data preprocessing module is responsible for performing data preprocessing; the data preprocessing includes data cleaning and data standardization. For numerical data, data standardization is performed using the Max-Min method; for categorical data, one-hot encoding is used to convert it into numerical vectors. The simulation calculation module is used to calculate the initial skin condition assessment value based on the preprocessed skin care historical data and real-time skin care data, combined with the preset weights of the initial skin care indicators; and dynamically updates the contribution of the skin care indicators based on the changes in skin condition over several periods of skin care historical data and the corresponding initial skin condition assessment value, thereby updating the weights of the skin care indicators and calculating the current skin condition assessment value in real time. The AI-driven module is used to compare the current skin condition assessment value calculated in real time with the preset scheme adjustment trigger conditions. When the scheme adjustment trigger conditions are exceeded, product adjustment recommendations are made, and the usage frequency is adjusted according to the change in the skin condition assessment value.

5. The AI-driven skincare solution management system according to claim 4, characterized in that: The simulation calculation module includes a data storage unit, a skin condition assessment unit, and an index weight iteration unit; The data storage unit is used to store the data that has undergone data preprocessing; The skin condition assessment unit is used to calculate the skin condition assessment value based on the preprocessed historical skin care data and real-time skin care data. The indicator weight iteration unit is used to dynamically update the contribution of skin care indicators based on the changes in skin condition over several periods of skin care historical data and the corresponding initial skin condition assessment value, thereby updating the weight of skin care indicators.

6. The AI-driven skincare solution management system according to claim 5, characterized in that: The AI-driven module includes a threshold control unit, an association matrix unit, a product recommendation unit, and a frequency update unit; The threshold control unit is used to obtain the skin condition assessment threshold based on historical data statistics; The association matrix unit is used to construct an association matrix M based on the historical efficacy of products used and skincare indicators; The product recommendation unit is used to make product adjustment recommendations by calculating the correlation between the change in the skin care index evaluation value and the product's impact on the skin care index value. The frequency update unit is used to adjust the usage frequency based on the change in skin condition assessment values.

Images