Cosmetic component database-driven cosmetic adverse reaction prediction method and system

Abstract

The invention discloses a cosmetic adverse reaction prediction method and system driven by a cosmetic component database, and relates to the technical field of computers, and the method comprises the steps: obtaining the attribute and component information of a target cosmetic, predicting the audience characteristics of the target cosmetic according to the information, and forming a multi-cluster audience characteristic flow. And performing registration optimization on the cosmetic component database by using the cosmetic attribute and component information, and determining a registration component database. A plurality of adverse reaction prediction learners and a loss analyzer are used for learning, and instant and delayed adverse reaction prediction dual channels are established. And inputting the cosmetic components and the audience characteristics into double channels to generate a cosmetic adverse reaction prediction report. The technical problem that personalized and accurate prediction cannot be provided in existing cosmetic adverse reaction prediction is solved, and the technical effect of improving prediction accuracy is achieved by achieving personalized and accurate prediction of cosmetic adverse reactions.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a method and system for predicting cosmetic adverse reactions driven by a cosmetic ingredient database. Background Art

[0002] With the increasing attention to the safety of cosmetics, the prediction of cosmetic adverse reactions has become an important issue in the cosmetics industry. Traditional prediction methods mainly rely on ingredient comparison or the general reactions of a wide population, which cannot accurately reflect individual differences and result in non-personalized and inaccurate prediction results. At the same time, existing technologies often only focus on the prediction of immediate reactions and ignore the delayed reactions after long-term use, resulting in insufficient comprehensiveness of the prediction. With the development of big data and artificial intelligence technologies, more accurate analysis can be performed based on the ingredient information of cosmetics and user characteristics, providing more personalized prediction and evaluation. However, the current technology is still difficult to achieve truly personalized and accurate prediction and cannot fully consider the different needs and potential risks of each user. Therefore, how to provide more accurate and personalized prediction results in the prediction of cosmetic adverse reactions remains the main problem faced by the current technology.

[0003] In the related technologies at the present stage, there are technical problems in the prediction of cosmetic adverse reactions that cannot provide personalized and accurate predictions. Summary of the Invention

[0004] This application solves the technical problem in the existing prediction of cosmetic adverse reactions that cannot provide personalized and accurate predictions by providing a method and system for predicting cosmetic adverse reactions driven by a cosmetic ingredient database.

[0005] This application provides a method for predicting cosmetic adverse reactions driven by a cosmetic ingredient database, including:

[0006] Obtaining a cosmetic adverse reaction prediction instruction, where the cosmetic adverse reaction prediction instruction includes cosmetic attribute information and cosmetic ingredient information corresponding to a target cosmetic; predicting the audience characteristics of the target cosmetic according to the cosmetic attribute information to obtain multiple clusters of audience characteristic streams; performing registration optimization on a cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information to determine a registered cosmetic ingredient library; introducing multiple adverse reaction prediction learners and an adverse reaction prediction loss resolver to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, and building a dual-channel for predicting cosmetic adverse reactions, where the dual-channel for predicting cosmetic adverse reactions includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel; inputting the cosmetic ingredient information and the multiple clusters of audience characteristic streams into the dual-channel for predicting cosmetic adverse reactions to generate a cosmetic adverse reaction prediction report.

[0007] The present application provides a cosmetic adverse reaction prediction system driven by a cosmetic ingredient database, including:

[0008] A prediction instruction acquisition module, which is used to obtain a cosmetic adverse reaction prediction instruction. Among them, the cosmetic adverse reaction prediction instruction includes cosmetic attribute information and cosmetic ingredient information corresponding to the target cosmetic; an audience characteristic prediction module, which is used to predict the audience characteristics of the target cosmetic according to the cosmetic attribute information to obtain multiple clusters of audience characteristic streams; a database registration optimization module, which is used to perform registration optimization on the cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information to determine a registered cosmetic ingredient library; a prediction dual-channel construction module, which is used to introduce multiple adverse reaction prediction learners and adverse reaction prediction loss analyzers to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library to construct a cosmetic adverse reaction prediction dual-channel. Among them, the cosmetic adverse reaction prediction dual-channel includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel; an adverse reaction prediction report generation module, which is used to input the cosmetic ingredient information and the multiple clusters of audience characteristic streams into the cosmetic adverse reaction prediction dual-channel to generate a cosmetic adverse reaction prediction report.

[0009] It is intended to propose a cosmetic adverse reaction prediction method and system driven by a cosmetic ingredient database according to the present application. First, obtain the attributes and ingredient information of the target cosmetic, and then predict the audience characteristics of the target cosmetic according to the information to form multiple clusters of audience characteristic streams. Use the cosmetic attribute and ingredient information to perform registration optimization on the cosmetic ingredient database to determine the registered ingredient library. Use multiple adverse reaction prediction learners and loss analyzers for learning to construct an immediate and delayed adverse reaction prediction dual-channel. Input the cosmetic ingredients and audience characteristics into the dual-channel to generate a cosmetic adverse reaction prediction report, achieving the technical effect of improving the prediction accuracy by realizing the personalized and accurate prediction of cosmetic adverse reactions. Brief Description of the Drawings

[0010] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the operations described above or below do not necessarily need to be executed precisely in order. On the contrary, according to the need, they can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several operations can be removed from these processes.

[0011] Figure 1Schematic flowchart of the method for predicting cosmetic adverse reactions driven by a cosmetic ingredient database provided by an embodiment of the present application;

[0012] Figure 2 Schematic structural diagram of the system for predicting cosmetic adverse reactions driven by a cosmetic ingredient database provided by an embodiment of the present application.

[0013] Explanation of reference numerals: Prediction instruction acquisition module 10, Audience characteristic prediction module 20, Database registration optimization module 30, Prediction dual-channel construction module 40, Adverse reaction prediction report generation module 50. Detailed implementation manners

[0014] The above description is only an overview of the technical solution of the present application. In order to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically gives the detailed implementation manners of the present application.

[0015] In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present application.

[0016] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The terms "first" and "second" involved are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0017] An embodiment of the present application provides a method for predicting cosmetic adverse reactions driven by a cosmetic ingredient database, as Figure 1 shown, the method includes:

[0018] Step S100, obtain a cosmetic adverse reaction prediction instruction, where the cosmetic adverse reaction prediction instruction includes cosmetic attribute information and cosmetic ingredient information corresponding to the target cosmetic. Specifically, obtain a cosmetic adverse reaction prediction instruction, which contains the attribute information and ingredient information of the target cosmetic, providing basic data for subsequent analysis. The cosmetic attribute information includes product category, function, applicable skin type, brand, usage method, etc., which is mainly used to match similar products in the database and predict the target user group, improving the accuracy of personalized adverse reaction assessment. The cosmetic ingredient information covers specific ingredient names, ingredient categories, contents, risk levels, and function descriptions. By matching with the cosmetic ingredient database, the system can find similar products, calculate the ingredient similarity, and evaluate whether the cosmetic contains known high-risk ingredients, such as sensitizers or irritating preservatives. To ensure data consistency, the system standardizes the cosmetic attribute and ingredient information, including unified naming, matching CAS numbers, concentration normalization, and risk level annotation, thereby constructing a structured input data set. Finally, this data set is stored in the database and can be called through an API interface or a machine learning model, providing accurate support for subsequent audience characteristic prediction, ingredient optimization matching, and adverse reaction analysis, thus improving the reliability and scientificity of the prediction system.

[0019] Step S200, predict the audience characteristics of the target cosmetic according to the cosmetic attribute information, and obtain multiple clusters of audience characteristic streams. Specifically, identify its potential user group through the attribute information of the target cosmetic and segment the audience characteristics, so that subsequent adverse reaction predictions can be analyzed individually for different populations. First, the system matches historical data and queries the market user portrait based on information such as product category, function, and applicable skin type to determine the target audience pool. Subsequently, core characteristics such as age, gender, and skin type are collected from the target audience, and data such as geographical environment, lifestyle, and skin allergy history are combined to construct a complete audience characteristic library. Then, the system classifies the target audience by age-gender and skin type. For example, women aged 18-25 prefer light moisturizing products, oily skin people are more inclined to oil-control cosmetics, and sensitive skin users are more likely to be stimulated by certain ingredients. Based on the classification results, the system constructs multiple clusters of audience characteristic streams, converting the characteristics of different populations into structured data streams, such as "audience cluster 1: 18-25 years old, female, oily skin" "audience cluster 2: 26-35 years old, female, dry skin", etc., enabling them to be used as data inputs for parallel computing, improving the adaptability and accuracy of the prediction system for different groups. Finally, this multiple clusters of audience characteristic streams provide data support for subsequent adverse reaction predictions, enabling the model to conduct customized analysis based on the characteristics of different populations, improving the scientificity and reliability of the prediction.

[0020] In a possible implementation, based on the cosmetic attribute information, the target cosmetic is predicted for audience characteristics, and multiple clusters of audience characteristic streams are obtained. Step S200 further includes step S210 of predicting the target cosmetic for the audience according to the cosmetic attribute information to determine the target audience population. Specifically, the audience prediction determines the target audience population of the target cosmetic by analyzing the attribute information of the cosmetic. First, the system predicts the potential user group of the cosmetic according to the attributes such as the product category, functional characteristics, and applicable skin type of the product. For example, an oil-control liquid foundation is suitable for young people with oily skin, and an anti-aging essence is suitable for women over 40 years old. Then, the system collects the characteristic information such as the age, gender, and skin type of the target audience, constructs a target audience characteristic library, and further subdivides the group through age and gender classification and skin type classification, such as women with oily skin aged 18-25, men with sensitive skin aged 30-40, etc. Finally, the system integrates the classification results into multiple clusters of audience characteristic streams, providing accurate data support for subsequent personalized adverse reaction prediction, thereby improving the accuracy and pertinence of the prediction.

[0021] Step S220, collect the age information, gender information, and skin type information of the target audience population to obtain a target audience characteristic library. Specifically, a target audience characteristic library is constructed by collecting the age, gender, and skin type information of the target audience. First, the system predicts the target audience population according to the attribute information of the cosmetic (such as product category, function, applicable skin type, etc.), and collects the age, gender, and skin type data of the audience through questionnaire surveys, e-commerce platform analysis, and social media data. For the age information, the system divides the audience into different age groups (such as 18-25 years old, 26-35 years old, etc.), and further refines the audience classification according to gender and skin type (such as oily, dry, combination, sensitive skin). By integrating the above data, the system establishes a database containing audience characteristics, ensuring that subsequent adverse reaction prediction and personalized recommendation can be analyzed based on accurate user data. The above process provides important support for the safety assessment of cosmetics and improves the accuracy of product matching and adverse reaction prediction.

[0022] Step S230: Perform age and gender classification based on the target audience feature library to obtain the initial classification results of multiple clusters of audiences. Specifically, based on the target audience feature library, the system first conducts age and gender classification to determine the target audience group. The system predicts the audience through the attribute information of the target cosmetics (such as product category, efficacy, and applicable skin type). Taking age as a dimension, the system divides the audience into multiple age groups. For example, the young population aged 18 - 25 usually needs oil-control, moisturizing, and basic skin care products, such as oil-control liquid foundation, moisturizing spray, etc.; young women aged 26 - 35 are more concerned about whitening and anti-aging effects and are suitable for using whitening essences containing niacinamide and vitamin C; while the female population aged 36 - 45 begins to pay attention to anti-aging and firming skin care products, such as anti-aging essence and firming cream; the population over 46 years old usually has loose and dry skin and prefers to use moisturizing and anti-aging repair products. Next, the system further refines the classification according to gender information. For women, the demands are usually more diverse, with high demands for basic skin care, beauty products, anti-aging essences, etc.; while men usually prefer basic care products, such as oil-control, cleansing, and sunscreen products. For example, men usually prefer simple facial cleansers and moisturizing creams, while women choose products containing fragrances or more beauty effects. Through the above methods, the system divides the target audience into multiple groups. For example, "18 - 25 years old, female, oily skin" may tend to choose fresh moisturizing and oil-control products, while "36 - 45 years old, male, sensitive skin" prefers mild repair skin care products. Finally, the system classifies the segmented audience groups into the initial classification results of multiple clusters of audiences. Each group has similar characteristics and demands, which is convenient for subsequent personalized recommendation and adverse reaction prediction. Classification not only helps to determine the potential market of cosmetics but also provides accurate data support for subsequent precise prediction and safety assessment.

[0023] Step S240: Conduct skin type classification based on the initial multi-cluster audience classification results to obtain the multi-cluster audience feature stream. Specifically, the system segments the target audience according to the skin type information of each audience (such as oily, dry, combination, sensitive), ensuring that different skin type groups can receive the most suitable skincare recommendations. For example, audiences with oily skin usually face problems such as oiliness, enlarged pores, and acne, so they tend to choose oil-control and refreshing skincare products, such as oil-control toners, light foundations, and refreshing creams. The system groups these people into a cluster, such as "18-25-year-old women, oily skin", whose needs focus on oil control and moisturization. For audiences with dry skin, their skin lacks moisture and is prone to tightness and dryness, and they usually need more moisturizing products, such as creams and serums containing ingredients like hyaluronic acid and ceramides. The system classifies users into the cluster of "26-35-year-old women, dry skin" and gives priority to high-moisturizing products when recommending products. For people with combination skin, the T-zone is usually oily while the U-zone is dry, and this group prefers products that balance oil and moisture, such as light moisturizing lotions and oil-control serums. For example, for the group of "36-45-year-old men, combination skin", balanced products are recommended to address skin problems in different areas. Finally, the group with sensitive skin reacts strongly to external stimuli and is prone to redness and allergies, so they need mild and non-irritating skincare products, such as repair lotions and creams with low-sensitivity ingredients. These users are classified into the cluster of "men over 45 years old, sensitive skin", and when recommending products, emphasis is placed on mild repair and products with strong repair properties. Through skin type classification, the system segments the audience into different clusters, provides personalized product recommendations for each group, and generates a multi-cluster audience feature stream, providing an accurate basis for subsequent adverse reaction prediction, ensuring that the skincare needs of each group are met, and at the same time improving the accuracy of cosmetic adverse reaction prediction.

[0024] Step S300: Perform registration optimization on the cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information to determine a registered cosmetic ingredient library. Specifically, the registration optimization of the cosmetic ingredient database is to optimize and register the existing cosmetic ingredient database by analyzing the attribute information and ingredient information of the target cosmetic, so as to ensure the accuracy of the ingredient data. First, the system filters the ingredient database according to the attribute information of the cosmetic, such as product category (such as sunscreen, anti-aging essence, cream, etc.), efficacy (such as whitening, anti-aging, moisturizing, etc.), and applicable skin type (such as dry skin, oily skin, sensitive skin, etc.). For example, for a moisturizing cream, the system first screens out ingredients with moisturizing functions, such as hyaluronic acid, glycerin, ceramide, etc. Then, based on the cosmetic ingredient information, the system evaluates the twin degree of each ingredient and the target cosmetic ingredient, that is, calculates the similarity of their structures, chemical properties, and efficacy. For example, hyaluronic acid and glycerin usually have a high similarity in moisturizing effect, so their registration coefficients are relatively high and are suitable for use in moisturizing products. Next, the system screens out the ingredients with registration coefficients greater than the threshold by setting a registration threshold for ingredients. For example, if the target cosmetic is an anti-aging essence, the system may set the registration coefficient to 0.8, and only ingredients with registration coefficients greater than 0.8, such as retinol and peptides, will be selected into the final formula. Finally, the system generates a registered cosmetic ingredient library, which contains all the optimized and screened ingredients, not only having a high matching degree but also meeting the functional requirements of the product. For example, for an anti-aging essence, the registered ingredient library may include hyaluronic acid, retinol, and peptide ingredients, which have significant effects in anti-aging and skin repair. The registered cosmetic ingredient library will serve as an important reference for subsequent adverse reaction prediction, personalized recommendation, and product development, providing data support for ensuring the safety and effectiveness of cosmetics.

[0025] In a possible implementation, the cosmetic ingredient database is registered and optimized according to the cosmetic attribute information and the cosmetic ingredient information to determine a registered cosmetic ingredient library. Step S300 further includes step S310 of performing associated screening on the cosmetic ingredient database according to the cosmetic attribute information to obtain a cosmetic ingredient library with the same attributes. Specifically, the cosmetic ingredient database is associated and screened according to the attribute information of the target cosmetic to screen out a set of ingredients that match the function, category, and applicable skin type of the target cosmetic. For example, if the target cosmetic is a moisturizing cream, the system will screen out ingredients with moisturizing effects in the ingredient database according to the moisturizing function, such as hyaluronic acid, glycerin, ceramides, etc. The above ingredients usually have good water-locking ability and are suitable for dry and sensitive skin. If the target cosmetic is an anti-aging essence, the system will screen out ingredients with anti-aging effects, such as retinol, peptides, etc., which help stimulate skin regeneration and reduce fine lines and wrinkles. In addition, if the target cosmetic is a sunscreen, ingredients with ultraviolet protection effects, such as titanium dioxide, zinc oxide, avobenzone, etc., will be screened out, which can effectively block ultraviolet rays and reduce the risk of skin aging and sunburn. During the screening process, the system will further refine the screening according to the applicable skin type of the target product to ensure that the ingredients meet the needs of different skin types. For example, oily skin is suitable for using oil-control ingredients such as salicylic acid and tea tree oil, while dry skin requires more moisturizing ingredients such as squalene and glycerin. Finally, the system screens out all ingredients that meet the attribute requirements of the target cosmetic to form a cosmetic ingredient library with the same attributes, which contains all ingredients suitable for the target product. The above ingredients will provide a basis for subsequent formula optimization and adverse reaction prediction to ensure the function and safety of the target cosmetic.

[0026] Step S320: Evaluate the twin degree of each same - attribute cosmetic ingredient data in the same - attribute cosmetic ingredient library according to the cosmetic ingredient information, and obtain multiple ingredient registration coefficients. Specifically, use the ingredient information of the target cosmetic to evaluate the twin degree of each ingredient in the same - attribute cosmetic ingredient library to calculate the registration coefficient of each ingredient. It involves a comprehensive evaluation of the chemical structure, functionality, and skin reactivity of each ingredient. Taking a moisturizing cream as an example, hyaluronic acid, as a common moisturizing ingredient, is often evaluated as highly matched due to its simple structure and effective moisturizing ability, and may obtain a relatively high registration coefficient, such as 0.9; while glycerol, also a moisturizing ingredient, although it has good moisturizing effect, its efficacy may be slightly weaker in some high - efficiency formulations, and the registration coefficient may be 0.85. In addition, niacinamide, as an ingredient with functions of whitening, anti - aging, and anti - inflammation, although it has good moisturizing effect, its main function is not moisturizing compared with hyaluronic acid, so it may obtain a relatively low registration coefficient, such as 0.7. The system screens out the ingredients that best meet the requirements of the target product according to the registration coefficients of the above - mentioned ingredients, and preferentially selects the ingredients with high registration coefficients into the final formula. For example, in this moisturizing cream, hyaluronic acid and glycerol may be selected, while niacinamide may be excluded. In this way, the system can ensure a high degree of matching in function and effect of the selected ingredients, and at the same time provide a scientific basis for subsequent formula optimization and adverse reaction prediction, ensuring the efficacy and safety of the product.

[0027] Step S330: Optimize and select the multiple ingredient registration coefficients according to the ingredient registration threshold to obtain an ingredient registration optimization distribution greater than or equal to the ingredient registration threshold. Specifically, first screen the registration coefficients of each ingredient according to the set ingredient registration threshold to ensure that only the ingredients that highly match the target cosmetic are selected. For example, when developing an anti - aging essence, the functional requirements of the target product include anti - aging and skin repair, so the system sets an ingredient registration threshold of 0.8. Suppose that after the twin - degree evaluation, the registration coefficient of hyaluronic acid is 0.9, the registration coefficient of retinol is 0.85, the registration coefficient of glycerol is 0.75, and the registration coefficient of the fragrance ingredient is 0.3. The system will add the registration coefficients of hyaluronic acid and retinol to the ingredient registration optimization distribution because their registration coefficients are greater than the set threshold of 0.8, while glycerol and the fragrance ingredient will be excluded because their registration coefficients are lower than the threshold. Finally, the generated ingredient registration optimization distribution contains hyaluronic acid and retinol, and their registration coefficients are 0.9 and 0.85 respectively, indicating that they highly match the functional requirements of the target anti - aging essence and can provide the required moisturizing and anti - aging effects. In this way, the system can effectively screen out the ingredients that best meet the requirements of the target cosmetic and provide strong support for product formula optimization.

[0028] Step S340: Screen the same-attribute cosmetic ingredient library according to the ingredient registration optimization distribution to generate the registered cosmetic ingredient library. Specifically, first, screen the registration coefficients of each ingredient according to the ingredient registration threshold to ensure that the selected ingredients can highly match the functional requirements of the target cosmetic. Specifically, when developing an anti-aging essence, the system sets the registration threshold at 0.8 according to the requirements of the target product. For example, after the previous twin-degree evaluation, the system obtains that the registration coefficient of hyaluronic acid is 0.9, retinol is 0.85, the registration coefficient of glycerin is 0.75, and the registration coefficient of the fragrance ingredient is 0.3. Since the registration coefficients of hyaluronic acid and retinol are both greater than 0.8, meeting the moisturizing and anti-aging functional requirements of the target product, they will be retained, while glycerin and fragrance ingredients with registration coefficients lower than 0.8 will be excluded. Finally, the system adds the registration coefficients of hyaluronic acid and retinol to the ingredient registration optimization distribution to form an optimized library containing highly matching ingredients. Through the above screening process, the system ensures that the final registered cosmetic ingredient library only contains those ingredients that highly fit the target product. Ingredients such as hyaluronic acid and retinol can not only effectively provide anti-aging and moisturizing effects but also ensure the efficacy and safety of the product.

[0029] Step S400: Introduce multiple adverse reaction prediction learners and adverse reaction prediction loss analyzers to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, and build a dual-channel for cosmetic adverse reaction prediction. The dual-channel for cosmetic adverse reaction prediction includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel. Specifically, by introducing multiple adverse reaction prediction learners and adverse reaction prediction loss analyzers, adverse reaction prediction loss optimization learning is performed on the ingredient library of the target cosmetic, and a dual-channel for cosmetic adverse reaction prediction is built. The dual-channel includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel. For example, for an anti-aging essence, the system first analyzes its ingredients such as retinol and fragrance ingredients through the immediate adverse reaction prediction channel to predict whether these ingredients will cause immediate reactions such as skin stinging and allergies in a short time. If the system discovers through historical data that the fragrance ingredient has a high probability of allergic reactions, it will immediately give a warning and may suggest reducing the proportion of the fragrance ingredient. On the other hand, the system evaluates the long-term effects of these ingredients through the delayed adverse reaction prediction channel. For example, retinol may cause skin dryness or irritation after long-term use. Therefore, the system will make predictions in the delayed channel to remind users to monitor their skin conditions or adjust the usage frequency. Through the coordinated work of these two channels, the system can predict the safety during the use of cosmetics in real-time and in the long term, and provide guidance for the optimization and adjustment of the formula to ensure that the final product is both effective and safe in the market.

[0030] In a possible implementation, multiple adverse reaction prediction learners and an adverse reaction prediction loss parser are introduced to perform optimization learning on the adverse reaction prediction loss of the registered cosmetic ingredient library, and a dual-channel for cosmetic adverse reaction prediction is built. Among them, the dual-channel for cosmetic adverse reaction prediction includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel. Step S400 further includes step S410, which retrieves adverse reaction records according to the registered cosmetic ingredient library to obtain an immediate adverse reaction record library and a delayed adverse reaction record library. Specifically, first, adverse reaction records are retrieved based on the registered cosmetic ingredient library to obtain immediate and delayed adverse reaction records for each ingredient. First, the system establishes an immediate adverse reaction record library based on the historical data of ingredients, consumer feedback, and clinical trial data, recording the reactions that ingredients may cause in the short term (such as within a few minutes to a few hours). For example, fragrance ingredients are often reported to cause immediate reactions such as skin allergies, redness, and stinging, and all these data will be collected and stored in the immediate adverse reaction record library. Then, the system establishes a delayed adverse reaction record library based on long-term usage data and product trial records to record the reactions that cosmetic ingredients may cause during long-term use, such as skin dryness, pigmentation, or peeling. For example, although retinol can effectively anti-aging, long-term use may cause skin dryness or peeling, and the above reactions will be recorded in the delayed adverse reaction record library. Through the comprehensive retrieval of the immediate and delayed adverse reactions of the above ingredients, the system can better understand the risks that each ingredient may bring at different usage stages, thus providing a scientific basis for the optimization of product formulations and safety assessments.

[0031] Step S420: Collect audience characteristics based on the immediate adverse reaction record library to obtain the first audience characteristics record library. Specifically, first extract relevant audience characteristics from the immediate adverse reaction record library, and establish the first audience characteristics record library by collecting information such as the age, gender, skin type, and allergy history of users. For example, assume that a skin care product containing fragrance ingredients has caused an allergic reaction in the group of women with oily skin aged 20 - 30. The system will extract the characteristic information (age, gender, skin type, reaction type) of this group and store it in the first audience characteristics record library. Similarly, if some users have immediate reactions such as skin dryness or stinging when using an anti-aging essence containing retinol, the system will record the above immediate reactions and associate the data with the corresponding audience characteristics (such as women with dry skin aged 30 - 40, allergy history, etc.). For each audience group, the system will also collect information on skin type (such as oily, dry, sensitive, etc.) and whether there is an allergy history, so as to make subsequent predictions more personalized. Through the above characteristics, the system can establish an accurate first audience characteristics record library according to the type of immediate adverse reactions and the specific information of the audience, providing strong data support for subsequent cosmetic formula optimization and adverse reaction prediction.

[0032] Step S430: Collect audience characteristics based on the delayed adverse reaction record library to obtain the second audience characteristics record library. Specifically, extract audience characteristics related to long-term use from the delayed adverse reaction record library and store the characteristic data in the second audience characteristics record library. For example, assume that an anti-aging essence containing retinol causes dryness and peeling after long-term use. The system will record the data of the delayed reaction and associate it with the characteristics of a specific audience, such as age, gender, skin type, etc. Female users with dry skin aged 30 - 40 may experience more significant skin dryness or peeling when using retinol, so the above data will be recorded as a specific audience characteristic. Similarly, for male users with oily skin aged 20 - 30 using the same product, the system may not record obvious delayed adverse reactions because users with oily skin usually have a higher tolerance to retinol. The system will also collect other characteristics of different audience groups, such as whether there is an allergy history, usage frequency, etc., especially for ingredients that can cause pigmentation and allergic reactions after long-term use, such as products containing certain preservatives or fragrances. Through the data, the system can generate a complete second audience characteristics record library, providing a more accurate basis for subsequent risk prediction and ensuring the safety of products during long-term use.

[0033] Step S440: Based on the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, perform optimization learning on the registered cosmetic ingredient library, the first record library of audience characteristics, and the immediate adverse reaction record library to generate the immediate adverse reaction prediction channel. Specifically, through multiple adverse reaction prediction learners (machine learning models) and the adverse reaction prediction loss parser, perform optimization learning on the registered cosmetic ingredient library, the first record library of audience characteristics, and the immediate adverse reaction record library, thereby generating the immediate adverse reaction prediction channel. For example, the system uses decision trees to analyze whether certain ingredients (such as fragrances) will cause allergic reactions in a specific audience (such as women aged 30 - 40). Through the training of machine learning models, the system can identify which ingredients will cause allergies in users with oily skin and fewer reactions in users with dry skin. Support vector machines help the system handle the complex relationships between different ingredients and user characteristics, especially for classifying immediate reactions under the combined action of multiple characteristics (such as skin type, allergy history, age, etc.). During the model training process, the loss parser optimizes the learner by calculating the error between the predicted result and the actual reaction. For example, if a fragrance ingredient causes an unexpected allergic reaction in some user groups, the loss parser will adjust the model, reduce this error, and readjust the usage ratio of the ingredient. Through this process, multiple learners finally generate an immediate adverse reaction prediction channel that can real-time predict immediate adverse reactions after using cosmetics, such as allergies, stings, and redness and swelling, etc.

[0034] Step S450: Based on the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, perform optimization learning on the registered cosmetic ingredient library, the second audience characteristic record library, and the delayed adverse reaction record library to generate the delayed adverse reaction prediction channel. Specifically, use the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser to perform optimization learning on the registered cosmetic ingredient library, the second audience characteristic record library, and the delayed adverse reaction record library to generate the delayed adverse reaction prediction channel. For example, when the system analyzes ingredients such as retinol, multiple machine learning models (such as decision trees, neural networks, random forests) are used to learn the long-term effects of retinol on users of different ages and skin types. Suppose for users with dry skin over 40 years old, the system may find that the retinol ingredient causes delayed reactions such as skin dryness and peeling after long-term use. After the model inputs this data and further analyzes it through a support vector machine, the system can identify which age groups, genders, and skin types of users are most likely to have side effects. During this process, the loss parser optimizes the prediction results of the model by calculating the error of each model. For example, suppose the system's prediction of the possible delayed adverse reactions caused by a certain ingredient does not match the actual data. The loss parser will make adjustments based on the error to reduce this prediction deviation and improve accuracy. Through the above method, the system continuously optimizes the prediction model and generates the final delayed adverse reaction prediction channel, which can real-time warn of possible side effects after long-term use, such as skin dryness, pigmentation, and long-term allergies, thus providing data support for the safety optimization of product formulations.

[0035] Step S460: Connect the immediate adverse reaction prediction channel and the delayed adverse reaction prediction channel as parallel nodes to generate the dual-channel for predicting cosmetic adverse reactions. Specifically, the system first generates an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel, which independently analyze the adverse reactions of cosmetic ingredients during short-term and long-term use. The immediate adverse reaction prediction channel focuses on short-term reactions, such as the immediate adverse reactions like allergies, stings, or redness that fragrance ingredients may cause in users with oily skin. Through multiple machine learning models (such as decision trees and support vector machines), the system learns and predicts these short-term reactions, issuing warnings in real time and providing formulation adjustment suggestions. On the other hand, the delayed adverse reaction prediction channel targets adverse reactions after long-term use, such as the skin dryness, peeling, or pigmentation problems that retinol ingredients may cause during long-term use. Through deep learning and neural network models, the system can predict the occurrence of these delayed side effects after weeks or months of use and provide timely suggestions, such as reducing the usage frequency or replacing ingredients. To improve the efficiency and prediction accuracy of the system, these two channels are connected in a parallel node manner, ensuring that immediate and delayed reactions can be processed simultaneously and predicted independently, ultimately generating a dual-channel for predicting cosmetic adverse reactions. This dual-channel system can not only promptly identify short-term adverse reactions, such as allergic reactions caused by fragrance, but also predict long-term risks, such as the dryness or peeling that retinol may bring to dry skin, thus comprehensively evaluating the safety of cosmetics and providing precise suggestions for product optimization.

[0036] In a possible implementation manner, according to the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, an optimization learning of the immediate adverse reaction prediction loss is performed on the registered cosmetic ingredient library, the first record library of the audience characteristics, and the immediate adverse reaction record library to generate the immediate adverse reaction prediction channel. Step S440 further includes step S441, using the registered cosmetic ingredient library and the first record library of the audience characteristics as input information, and using the immediate adverse reaction record library as output information, respectively supervising and training the multiple adverse reaction prediction learners. Every time a predetermined number of training times is completed, multiple immediate adverse prediction loss coefficients are calculated according to the adverse reaction prediction loss parser. Specifically, the system uses the registered cosmetic ingredient library and the first record library of the audience characteristics as input data, and uses the immediate adverse reaction record library as output information to perform multiple trainings to generate an immediate adverse reaction prediction model. First, multiple adverse reaction prediction learners (such as decision trees, support vector machines, neural networks, etc.) are used to train the model. By learning the relationship between cosmetic ingredients and audience characteristics, it predicts which ingredients may cause allergic reactions, stinging sensations, redness and swelling, etc. in the short term. For example, fragrance ingredients may cause allergies in users with oily skin, but have less impact on users with dry skin. During each training process, the system calculates the immediate adverse prediction loss coefficient according to the adverse reaction prediction loss parser to evaluate the error between the prediction result and the actual reaction. Suppose the model predicts that a certain ingredient (such as a preservative) will cause allergic reactions in 20% of users, but the actual data shows that only 15% of users have reactions. The loss parser will calculate the error and adjust the model parameters. If the loss coefficient is lower than the preset immediate adverse prediction loss threshold, the system will consider that the model has reached the required accuracy and generate multiple immediate adverse reaction prediction models. These models are connected together through an ensemble learning method (such as weighted average or voting mechanism) to form an immediate adverse reaction prediction channel. For example, for a certain cosmetic product, the system can predict the immediate allergic reaction of the fragrance ingredient in users with oily skin, and at the same time predict the stinging sensation that the preservative ingredient may cause in sensitive skin, and finally provide safety evaluation and formulation adjustment suggestions for R & D personnel.

[0037] Step S442, if the multiple immediate adverse prediction loss coefficients are less than the immediate adverse prediction loss threshold, multiple immediate adverse reaction prediction models are generated. Specifically, the system uses the registered cosmetic ingredient library and the first record library of audience characteristics as input data, and the immediate adverse reaction record library as output data, and conducts supervised training through multiple adverse reaction prediction learners (such as decision trees, support vector machines, neural networks, etc.). Each time during training, the system uses the data in the training set (including cosmetic ingredients and user characteristics) to predict the immediate reactions (such as allergies, redness, or stinging) that cosmetic ingredients may cause to different groups (such as users with oily skin, dry skin, or sensitive skin). Through training, the model will gradually optimize and identify which ingredients are more likely to cause these reactions in specific audience groups. For example, fragrance ingredients may cause stinging or allergic reactions in users with dry skin, while such reactions may not occur in users with oily skin. The system calculates the loss coefficient of each training through an adverse reaction prediction loss parser to measure the error between the prediction result and the actual reaction. If the multiple immediate adverse prediction loss coefficients are less than the set immediate adverse prediction loss threshold, the system considers that the model has achieved the expected accuracy and will generate multiple immediate adverse reaction prediction models. For example, the system will generate a decision tree model that can identify the immediate reactions of fragrance ingredients to different age groups and skin types and adjust its predictions based on past training data. Next, the system connects these models through an ensemble learning method (such as weighted average or voting mechanism) to form a final immediate adverse reaction prediction channel, which can predict in real time the immediate adverse reactions that cosmetic ingredients may cause during use and provide accurate early warnings for product formula optimization and safety assessment.

[0038] Step S443: Connect the multiple immediate adverse reaction prediction models to obtain the immediate adverse reaction prediction channel. Specifically, in the cosmetic adverse reaction prediction system, when multiple immediate adverse reaction prediction models are generated after training, the system will connect the models through an ensemble learning method to form an immediate adverse reaction prediction channel. For example, suppose there are three models: decision tree, support vector machine (SVM), and neural network. Each model learns the relationship between cosmetic ingredients and the audience based on different features. The decision tree model may find that fragrance ingredients have a stronger allergic reaction in oily skin users, while the SVM model may predict that fragrance is more irritating to sensitive skin. The neural network provides a comprehensive prediction based on more dimensional data (such as age, allergy history, etc.). Through methods such as weighted average or voting mechanism, the system will integrate the prediction results of the models. For instance, if both the decision tree and neural network models predict that a certain fragrance ingredient will cause an allergic reaction in dry skin users, while the SVM model does not predict such a reaction, then the system will make a final judgment based on the prediction results of the majority of models. In addition, the performance of each model during the training process determines its weight in the ensemble, and models with better performance will obtain higher weights. Finally, by integrating the models, the immediate adverse reaction prediction channel can comprehensively evaluate the immediate reactions (such as redness, stinging, allergy, etc.) that cosmetics may cause in user groups with different skin types, ages, and allergy histories. For example, when it is predicted that a certain fragrance-containing cream will cause redness in female users with sensitive skin, the system will provide feedback and suggest reducing the use of fragrance or adjusting the ingredient ratio, thereby optimizing the product formula and improving its safety.

[0039] In a possible implementation manner, multiple adverse reaction prediction learners and an adverse reaction prediction loss resolver are introduced to perform optimization learning on the adverse reaction prediction loss of the registered cosmetic ingredient library, and a dual-channel cosmetic adverse reaction prediction is built. Among them, the dual-channel cosmetic adverse reaction prediction includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel. Step S400 further includes step S470. The adverse reaction prediction loss resolver includes an adverse reaction prediction loss resolution function, and the adverse reaction prediction loss resolution function is: where LOSS represents the adverse prediction loss coefficient, M represents the predetermined number of training times, m represents the m-th training, both M and m are positive integers, 1 ≤ m ≤ M, SUO m represents the number of adverse reaction prediction samples in the m-th training, SUX mCharacterize the number of correctly predicted samples of adverse reactions in the m-th training. Specifically, in the cosmetic adverse reaction prediction system, the adverse reaction prediction loss parser evaluates the accuracy of model prediction by calculating the loss coefficient. During each training process, the system calculates the loss coefficient for each round of training according to the adverse reaction prediction loss parsing function, and the form of this function is: where, SUO m represents the number of adverse reaction prediction samples in the m-th round of training, SUX m represents the number of correctly predicted samples, and M is the number of training rounds. For example, in a certain round of training, if the system processes 100 samples and successfully predicts 80 correct reactions, the loss coefficient is 0.2, which indicates that the prediction accuracy of the model needs to be improved. After each round of training, the system evaluates the performance of the model according to the calculated loss coefficient. If the loss coefficient is less than the preset loss threshold, it means that the model prediction is accurate enough. At this time, multiple models will generate a final immediate adverse reaction prediction model. For example, assume that the system discovers through training that spice ingredients can cause allergic reactions in users of certain age groups. After calculation by the loss parser, if the loss coefficient is still higher than the threshold, the system will adjust the model parameters, such as reducing the use of spices or optimizing different formulas for sensitive populations, until the loss coefficient meets the requirements. The final generated prediction model will be able to accurately reflect the immediate reactions of spices to user groups with different skin types, ages, and allergy histories. This process ensures that each prediction can identify potential adverse reactions in cosmetics as accurately as possible by continuously optimizing the loss coefficient, thereby providing a scientific basis for the safety of products.

[0040] Step S500, input the cosmetic ingredient information and the multi-cluster audience characteristic stream into the dual-channel for predicting cosmetic adverse reactions to generate a prediction report on cosmetic adverse reactions. Specifically, in the system for predicting cosmetic adverse reactions, the system takes the cosmetic ingredient information and the multi-cluster audience characteristic stream as input data and transmits them to the dual-channel for predicting cosmetic adverse reactions for processing. First, the system makes short-term (immediate adverse reactions) and long-term (delayed adverse reactions) predictions based on the cosmetic ingredient information, such as hyaluronic acid, retinol, preservatives, and fragrances, in combination with the multi-cluster audience characteristic stream (including the user's skin type, age, gender, allergy history, etc.). For example, the system will analyze the immediate reaction of fragrance ingredients to users with oily skin and predict the possible allergic reactions; at the same time, it will also evaluate the long-term effects of retinol ingredients after long-term use, especially the risk of skin dryness or peeling in users with dry skin. The immediate adverse reaction prediction channel will handle reactions such as redness, allergy, or stinging that may be caused by fragrance ingredients and make different predictions for different skin types and age groups. The delayed adverse reaction prediction channel analyzes the long-term use effects of cosmetic ingredients to predict problems such as skin dryness and pigmentation that may be caused by long-term use of retinol. Finally, the system combines the prediction results of these two channels to generate a detailed prediction report on cosmetic adverse reactions, which will list the short-term and long-term reactions of each ingredient in different user groups. For example, fragrance ingredients may cause allergies in female users with sensitive skin, while retinol may cause dryness or peeling in users with dry skin over 40 years old. Through this dual-channel processing, R & D personnel can adjust the product formula according to the report, such as reducing the use concentration of fragrance or adjusting the content of retinol, so as to optimize the safety of the product and ensure its applicability in the market.

[0041] In a possible implementation, the cosmetic ingredient information and the multi-cluster audience characteristic stream are input into the dual-channel for predicting cosmetic adverse reactions to generate a report on predicting cosmetic adverse reactions. Step S500 further includes step S510 of inputting the cosmetic ingredient information and the multi-cluster audience characteristic stream into the immediate adverse reaction prediction channel to obtain multi-cluster audience-immediate adverse prediction results. Specifically, in the cosmetic adverse reaction prediction system, the cosmetic ingredient information and the multi-cluster audience characteristic stream are used as inputs and enter the immediate adverse reaction prediction channel for processing. First, the system analyzes the cosmetic ingredient information, such as ingredients like fragrance, retinol, preservatives, etc., and combines it with the multi-cluster audience characteristic stream, which includes information such as the user's age, gender, skin type (such as oily skin, dry skin, sensitive skin), and allergy history. The system uses multiple machine learning models (such as decision trees, SVMs, neural networks, etc.) to evaluate the immediate reactions of these ingredients in different user groups. For example, the fragrance ingredient may cause an allergic reaction in the oily skin group but have little reaction in dry skin users. The system trains the models to accurately predict the immediate adverse reactions of different ingredients for each group, such as redness, stinging, allergies, etc. If the decision tree model predicts an 80% probability of an allergic reaction to the fragrance ingredient in the oily skin group, and the neural network model predicts a 60% probability of a redness reaction caused by this ingredient in sensitive skin users, the system will generate immediate adverse reaction prediction results for each audience group based on the prediction results of these different models. These results will help the R & D team identify the immediate allergic reactions that cosmetic ingredients may cause in different user groups and provide a scientific basis for product formulation adjustment. Finally, the multi-cluster audience-immediate adverse prediction results output by the system will detail the risk of immediate reactions that specific cosmetic ingredients may cause in each audience group, thus providing data support for product optimization and safety assessment.

[0042] Step S520: Input the cosmetic ingredient information and the multi-cluster audience characteristic stream into the delayed adverse reaction prediction channel to obtain the multi-cluster audience-delayed adverse prediction results. Specifically, the cosmetic ingredient information and the multi-cluster audience characteristic stream are input into the delayed adverse reaction prediction channel. The main task of this channel is to evaluate the potential impact of cosmetic ingredients on different groups after long-term use. The system first analyzes the information of cosmetic ingredients, such as retinol, hyaluronic acid, preservatives, etc. These ingredients may affect the skin after several weeks or months. For example, retinol may cause peeling or dryness in users with dry skin, but has less impact on the oily skin group. Then, the multi-cluster audience characteristic stream provides detailed information about each user group, including age, gender, skin type (such as oily, dry, sensitive skin), allergy history, etc. Based on this information, the system uses multiple machine learning models (such as decision trees, SVMs, neural networks, etc.) to predict the potential reactions of different ingredients during long-term use. For example, for the dry skin group over 40 years old, the system may predict that the probability of the retinol ingredient causing skin dryness or peeling is 60%, while for the young oily skin group, this probability may be only 20%. The system will calculate the delayed reaction risk of each cosmetic ingredient in this group based on the characteristics of each group, such as pigmentation, skin dryness or allergies. Finally, the system outputs the multi-cluster audience-delayed adverse prediction results, providing personalized delayed adverse reaction predictions for each group, helping the R & D team evaluate the safety of ingredients and optimize the product formula. For example, if the system detects that retinol may cause peeling reactions in the dry skin group, the R & D team may choose to reduce the concentration of this ingredient or develop a retinol-free alternative product for such users.

[0043] Step S530: Perform data fusion based on the multi-cluster audience-immediate adverse prediction results and the multi-cluster audience-delayed adverse prediction results, and output the cosmetic adverse reaction prediction report. Specifically, the cosmetic adverse reaction prediction report generated by data fusion of the immediate adverse reaction prediction result and the delayed adverse reaction prediction result evaluates the possible adverse reactions that each ingredient may cause in the short term (immediate reaction) and long term (delayed reaction) according to the characteristics of different audience groups (such as skin type, age, gender, allergy history, etc.) and cosmetic ingredient information. For example, fragrance ingredients may cause a higher immediate allergic reaction in users with oily skin, while retinol ingredients may cause dryness and peeling after long-term use in the dry skin group. The immediate adverse reaction prediction channel and the delayed adverse reaction prediction channel respectively predict these reactions to generate their respective results. Then, the system fuses these two types of prediction results through a weighted average or voting mechanism. For example, the immediate reaction prediction of fragrance ingredients in users with oily skin is 80%, while the delayed reaction prediction in users with dry skin is 20%. The system will perform weighted calculations, assign different weights according to the group characteristics, and obtain a comprehensive evaluation. If fragrance ingredients have a high risk of immediate reaction in the sensitive skin group, and retinol has a high risk of delayed reaction in the dry skin group, the report will issue warnings for these two types of ingredients respectively, prompting the R & D team to reduce the fragrance ingredient or adjust the concentration of retinol. In addition, the system comprehensively considers the characteristics of each group through a decision tree method. The finally generated report will list in detail the immediate and delayed reaction risks of each group and provide suggestions for optimizing the formula. For example, it is recommended to reduce the use of fragrance for sensitive skin users, or provide an alternative product version without retinol for dry skin users.

[0044] In the embodiment of the present application, the attributes and ingredient information of the target cosmetic are obtained, and then the audience characteristics of the target cosmetic are predicted according to the information to form a multi-cluster audience characteristic stream. The cosmetic ingredient database is registered and optimized using the cosmetic attributes and ingredient information to determine the registered ingredient library. Multiple adverse reaction prediction learners and loss analyzers are used for learning to build a dual-channel for immediate and delayed adverse reaction prediction. The cosmetic ingredients and audience characteristics are input into the dual-channel to generate a cosmetic adverse reaction prediction report, achieving the technical effect of improving the prediction accuracy by realizing personalized and accurate prediction of cosmetic adverse reactions.

[0045] In the above text, reference is made to Figure 1 The method for predicting cosmetic adverse reactions driven by a cosmetic ingredient database according to an embodiment of the present invention is described in detail. Next, reference will be made to Figure 2 Describe a cosmetic adverse reaction prediction system driven by a cosmetic ingredient database according to an embodiment of the present invention.

[0046] The cosmetic adverse reaction prediction system driven by the cosmetic ingredient database according to the embodiments of the present invention is used to solve the technical problem in the existing prediction of cosmetic adverse reactions that personalized and accurate prediction cannot be provided. By realizing the personalized and accurate prediction of cosmetic adverse reactions, the technical effect of improving the prediction accuracy is achieved. The cosmetic adverse reaction prediction system driven by the cosmetic ingredient database includes: a prediction instruction acquisition module 10, an audience characteristic prediction module 20, a database registration optimization module 30, a prediction dual-channel construction module 40, and an adverse reaction prediction report generation module 50.

[0047] The prediction instruction acquisition module 10 is used to obtain a cosmetic adverse reaction prediction instruction, wherein the cosmetic adverse reaction prediction instruction includes cosmetic attribute information and cosmetic ingredient information corresponding to the target cosmetic.

[0048] The audience characteristic prediction module 20 is used to perform audience characteristic prediction on the target cosmetic according to the cosmetic attribute information to obtain multiple clusters of audience characteristic streams.

[0049] The database registration optimization module 30 is used to perform registration optimization on the cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information to determine a registered cosmetic ingredient library.

[0050] The prediction dual-channel construction module 40 is used to introduce multiple adverse reaction prediction learners and adverse reaction prediction loss analyzers to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, and construct a cosmetic adverse reaction prediction dual-channel, wherein the cosmetic adverse reaction prediction dual-channel includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel.

[0051] The adverse reaction prediction report generation module 50 is used to input the cosmetic ingredient information and the multiple clusters of audience characteristic streams into the cosmetic adverse reaction prediction dual-channel to generate a cosmetic adverse reaction prediction report.

[0052] Next, the specific configuration of the audience characteristic prediction module 20 will be described in detail. As described above, based on the cosmetic attribute information, the target cosmetics are subjected to audience characteristic prediction to obtain multiple clusters of audience characteristic streams. The audience characteristic prediction module 20 further includes: a target audience population determination unit, which is used to perform audience prediction on the target cosmetics according to the cosmetic attribute information to determine the target audience population; a target audience characteristic library acquisition unit, which is used to collect the age information, gender information, and skin type information of the target audience population to obtain a target audience characteristic library; an initial classification result acquisition unit, which is used to perform age and gender classification according to the target audience characteristic library to obtain multiple clusters of initial audience classification results; a skin type classification unit, which is used to perform skin type classification according to the multiple clusters of initial audience classification results to obtain the multiple clusters of audience characteristic streams.

[0053] Next, the specific configuration of the database registration optimization module 30 will be described in detail. As described above, based on the cosmetic attribute information and the cosmetic ingredient information, the cosmetic ingredient database is registered and optimized to determine a registered cosmetic ingredient library. The database registration optimization module 30 further includes: a cosmetic ingredient library acquisition unit, which is used to perform associated screening on the cosmetic ingredient database according to the cosmetic attribute information to obtain a same-attribute cosmetic ingredient library; a twin degree evaluation unit, which is used to evaluate the twin degree of each same-attribute cosmetic ingredient data in the same-attribute cosmetic ingredient library according to the cosmetic ingredient information to obtain multiple ingredient registration coefficients; an optimization selection unit, which is used to perform optimization selection on the multiple ingredient registration coefficients according to an ingredient registration threshold to obtain an ingredient registration optimization distribution greater than or equal to the ingredient registration threshold; a cosmetic ingredient library screening unit, which is used to screen the same-attribute cosmetic ingredient library according to the ingredient registration optimization distribution to generate the registered cosmetic ingredient library.

[0054] Next, the specific configuration of the prediction dual-channel building module 40 will be described in detail. As described above, multiple adverse reaction prediction learners and adverse reaction prediction loss analyzers are introduced to perform optimization learning on the adverse reaction prediction loss of the registered cosmetic ingredient library, and a dual-channel for cosmetic adverse reaction prediction is built. Among them, the dual-channel for cosmetic adverse reaction prediction includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel. The prediction dual-channel building module 40 further includes: an adverse reaction record library acquisition unit, which is used to retrieve adverse reaction records according to the registered cosmetic ingredient library to obtain an immediate adverse reaction record library and a delayed adverse reaction record library; a first audience characteristic record library acquisition unit, which is used to collect audience characteristics according to the immediate adverse reaction record library to obtain a first audience characteristic record library; a second audience characteristic record library acquisition unit, which is used to collect audience characteristics according to the delayed adverse reaction record library to obtain a second audience characteristic record library; an immediate adverse reaction prediction channel generation unit, which is used to perform optimization learning on the immediate adverse reaction prediction loss of the registered cosmetic ingredient library, the first audience characteristic record library, and the immediate adverse reaction record library according to the multiple adverse reaction prediction learners and the adverse reaction prediction loss analyzer to generate the immediate adverse reaction prediction channel; a delayed adverse reaction prediction channel generation unit, which is used to perform optimization learning on the delayed adverse reaction prediction loss of the registered cosmetic ingredient library, the second audience characteristic record library, and the delayed adverse reaction record library according to the multiple adverse reaction prediction learners and the adverse reaction prediction loss analyzer to generate the delayed adverse reaction prediction channel; a parallel node connection unit, which is used to connect the immediate adverse reaction prediction channel and the delayed adverse reaction prediction channel as parallel nodes to generate the dual-channel for cosmetic adverse reaction prediction.

[0055] Among them, according to the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, the registered cosmetic ingredient library, the first record library of audience characteristics, and the immediate adverse reaction record library are subjected to immediate adverse reaction prediction loss optimization learning to generate the immediate adverse reaction prediction channel. The immediate adverse reaction prediction channel generation unit further includes: a loss coefficient calculation sub-unit, which is used to use the registered cosmetic ingredient library and the first record library of audience characteristics as input information, and the immediate adverse reaction record library as output information, and respectively perform supervised training on the multiple adverse reaction prediction learners. Each time a predetermined number of training times is completed, multiple immediate adverse prediction loss coefficients are calculated according to the adverse reaction prediction loss parser; an adverse reaction prediction model generation sub-unit, which is used to generate multiple immediate adverse reaction prediction models if the multiple immediate adverse prediction loss coefficients are less than the immediate adverse prediction loss threshold; an immediate adverse reaction prediction channel acquisition sub-unit, which is used to connect the multiple immediate adverse reaction prediction models to obtain the immediate adverse reaction prediction channel.

[0056] Among them, the prediction dual-channel construction module 40 further includes: an adverse reaction prediction loss parser composition unit, which is used for the adverse reaction prediction loss parser to include an adverse reaction prediction loss parsing function, and the adverse reaction prediction loss parsing function is: Among them, LOSS represents the adverse prediction loss coefficient, M represents the predetermined number of training times, m represents the m-th training, both M and m are positive integers, 1 ≤ m ≤ M, SUO m represents the number of adverse reaction prediction samples in the m-th training, SUX m represents the number of correctly predicted adverse reaction samples in the m-th training.

[0057] Next, the specific configuration of the adverse reaction prediction report generation module 50 will be described in detail. As described above, the cosmetic ingredient information and the multi-cluster audience characteristic stream are input into the dual-channel cosmetic adverse reaction prediction system to generate a cosmetic adverse reaction prediction report. The adverse reaction prediction report generation module 50 further includes: a multi-cluster audience - immediate adverse prediction result acquisition unit, which is used to input the cosmetic ingredient information and the multi-cluster audience characteristic stream into the immediate adverse reaction prediction channel to obtain the multi-cluster audience - immediate adverse prediction result; a multi-cluster audience - delayed adverse prediction result acquisition unit, which is used to input the cosmetic ingredient information and the multi-cluster audience characteristic stream into the delayed adverse reaction prediction channel to obtain the multi-cluster audience - delayed adverse prediction result; a data fusion unit, which is used to perform data fusion based on the multi-cluster audience - immediate adverse prediction result and the multi-cluster audience - delayed adverse prediction result, and output the cosmetic adverse reaction prediction report.

[0058] The cosmetic adverse reaction prediction system driven by the cosmetic ingredient database provided by the embodiments of the present invention can execute the cosmetic adverse reaction prediction method driven by the cosmetic ingredient database provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0059] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server. The included units and modules are only divided according to the functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of the functional units are only for the convenience of mutual distinction and do not limit the protection scope of the present invention.

[0060] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A cosmetics adverse reaction prediction method driven by a cosmetics ingredient database, characterized in that: The method comprises: Obtaining a cosmetic adverse reaction prediction instruction, wherein the cosmetic adverse reaction prediction instruction includes cosmetic attribute information and cosmetic ingredient information corresponding to the target cosmetic; Predicting audience characteristics of the target cosmetics according to the cosmetics attribute information to obtain multiple clusters of audience characteristic streams; Performing registration and optimization on the cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information to determine a registered cosmetic ingredient database; Introducing multiple adverse reaction prediction learners and adverse reaction prediction loss parsers to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, and building a dual channel for cosmetic adverse reaction prediction, wherein the dual channel for cosmetic adverse reaction prediction includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel; The cosmetic ingredient information and the multi-cluster audience feature streams are input into the cosmetic adverse reaction prediction dual channel to generate a cosmetic adverse reaction prediction report.

2. The method according to claim 1, characterized in that Predicting the audience characteristics of the target cosmetics according to the cosmetics attribute information to obtain multiple clusters of audience characteristic flows, including: Performing audience prediction for the target cosmetics according to the cosmetics attribute information to determine the target audience; Collecting the age information, gender information and skin quality information of the target audience to obtain a target audience feature library; Perform age and gender classification according to the target audience feature library to obtain multiple audience cluster initial classification results; Skin type classification is performed according to the initial classification results of the multiple clusters of audiences to obtain the multiple clusters of audience feature streams.

3. The method according to claim 1, characterized in that Performing registration and optimization on the cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information to determine the registered cosmetic ingredient database includes: Performing association screening on the cosmetic ingredient database according to the cosmetic attribute information to obtain a cosmetic ingredient database with the same attribute; According to the cosmetic ingredient information, twin degree evaluation is performed on each cosmetic ingredient data of the same attribute in the cosmetic ingredient library of the same attribute to obtain multiple ingredient calibration coefficients; Optimizing and selecting the plurality of component calibration coefficients according to a component calibration threshold value to obtain a component calibration optimal distribution that is greater than or equal to the component calibration threshold value; The cosmetic ingredient library with the same attribute is screened according to the component registration optimal distribution to generate the registered cosmetic ingredient library.

4. The method according to claim 1, characterized in that Introduce multiple adverse reaction prediction learners and adverse reaction prediction loss parsers to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, and build a dual channel for cosmetic adverse reaction prediction, including: Perform adverse reaction record retrieval according to the registered cosmetic ingredient library to obtain an immediate adverse reaction record library and a delayed adverse reaction record library; Collecting audience characteristics according to the immediate adverse reaction record library to obtain a first audience characteristics record library; Collecting audience characteristics according to the delayed adverse reaction record library to obtain a second audience characteristics record library; According to the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, the registered cosmetic ingredient library, the first audience feature record library and the immediate adverse reaction record library are subjected to immediate adverse reaction prediction loss optimization learning to generate the immediate adverse reaction prediction channel; According to the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, performing delayed adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, the second audience feature record library and the delayed adverse reaction record library to generate the delayed adverse reaction prediction channel; The immediate adverse reaction prediction channel and the delayed adverse reaction prediction channel are connected as parallel nodes to generate the cosmetic adverse reaction prediction dual channel.

5. The method according to claim 4, characterized in that According to the multiple adverse reaction prediction learners and the adverse reaction prediction loss parser, the registered cosmetic ingredient library, the first audience feature record library and the immediate adverse reaction record library are subjected to immediate adverse reaction prediction loss optimization learning to generate the immediate adverse reaction prediction channel, including: Using the registered cosmetic ingredient library and the first record library of audience characteristics as input information and the immediate adverse reaction record library as output information, respectively, the plurality of adverse reaction prediction learners are supervised trained, and each time a predetermined number of training times are trained, a plurality of immediate adverse reaction prediction loss coefficients are calculated according to the adverse reaction prediction loss parser; If the multiple immediate adverse reaction prediction loss coefficients are less than the immediate adverse reaction prediction loss threshold, generating multiple immediate adverse reaction prediction models; The multiple immediate adverse reaction prediction models are connected to obtain the immediate adverse reaction prediction channel.

6. The method according to claim 1, characterized in that The adverse reaction prediction loss parser includes an adverse reaction prediction loss parsing function, and the adverse reaction prediction loss parsing function is: Among them, LOSS represents the adverse reaction prediction loss coefficient, M represents the scheduled number of training times, m represents the mth training, M and m are both positive integers, 1≤m≤M, SUOm represents the number of adverse reaction prediction samples in the mth training, and SUXm represents the number of correct adverse reaction prediction samples in the mth training.

7. The method according to claim 1, characterized in that Inputting the cosmetic ingredient information and the multi-cluster audience feature streams into the cosmetic adverse reaction prediction dual channel to generate a cosmetic adverse reaction prediction report, including: Inputting the cosmetic ingredient information and the multi-cluster audience feature streams into the immediate adverse reaction prediction channel to obtain a multi-cluster audience-immediate adverse reaction prediction result; Inputting the cosmetic ingredient information and the multi-cluster audience feature streams into the delayed adverse reaction prediction channel to obtain a multi-cluster audience-delayed adverse reaction prediction result; Data fusion is performed based on the multiple-cluster audience-immediate adverse reaction prediction results and the multiple-cluster audience-delayed adverse reaction prediction results to output the cosmetic adverse reaction prediction report.

8. A cosmetic adverse reaction prediction system driven by a cosmetic ingredient database, characterized in that: The system is used to implement the cosmetic ingredient database-driven cosmetic adverse reaction prediction method according to any one of claims 1 to 7, and the system comprises: A prediction instruction acquisition module, the prediction instruction acquisition module is used to obtain a cosmetic adverse reaction prediction instruction, wherein the cosmetic adverse reaction prediction instruction includes cosmetic attribute information and cosmetic ingredient information corresponding to the target cosmetic; An audience characteristic prediction module, the audience characteristic prediction module is used to predict the audience characteristics of the target cosmetics according to the cosmetics attribute information to obtain multiple clusters of audience characteristic streams; A database registration and optimization module, the database registration and optimization module is used to perform registration and optimization on the cosmetic ingredient database according to the cosmetic attribute information and the cosmetic ingredient information, and determine the registered cosmetic ingredient database; A prediction dual-channel building module, wherein the prediction dual-channel building module is used to introduce multiple adverse reaction prediction learners and adverse reaction prediction loss analyzers to perform adverse reaction prediction loss optimization learning on the registered cosmetic ingredient library, and build a cosmetic adverse reaction prediction dual channel, wherein the cosmetic adverse reaction prediction dual channel includes an immediate adverse reaction prediction channel and a delayed adverse reaction prediction channel; The adverse reaction prediction report generation module is used to input the cosmetic ingredient information and the multi-cluster audience feature streams into the cosmetic adverse reaction prediction dual channel to generate a cosmetic adverse reaction prediction report.

Images