A record product risk early warning method and system based on a component library

Abstract

The present application relates to the field of information technology services, and discloses a record product risk early warning method and system based on a component library, comprising: extracting a full component list of a record product from product record data to construct a component database corresponding to the full component list by using a preset component matching rule; establishing a component quick search index and a risk feature set of the component database to match potential risk components corresponding to the full component list in the component database; constructing a multi-dimensional risk portrait corresponding to the record product by calculating the compliance correlation degree of each component in the full component list and the potential risk components; and outputting a risk early warning result of the record product according to a risk assessment conclusion corresponding to the record product and the multi-dimensional risk portrait. The present application can greatly improve the initiative foresight and prevention accuracy of record product risk early warning.

Description

Technical Field

[0001] This invention relates to a risk warning method and system for registered products based on an ingredient library, belonging to the field of information technology services. Background Technology

[0002] Registered products refer to consumer goods that have completed the registration process with regulatory agencies and obtained a unique registration mark, such as cosmetics and disinfection products. The authenticity and completeness of their registration information are the core basis for ensuring product compliance traceability and risk control, and they have a fundamental position in the intelligent management system for consumer product quality and safety.

[0003] Currently, the most common method for risk warning of registered products is manual comparison. This method involves manually entering the registration number on the official website of the regulatory agency, querying and downloading the product's registration information one by one, and then relying on the experience of professionals to read the ingredient list and assess its risks. Therefore, this method only intervenes after the risk has become apparent, rather than proactively intercepting it before the product is registered or distributed, making it difficult to proactively identify and mitigate potential systemic risks. Summary of the Invention

[0004] This invention provides a method and system for risk warning of registered products based on a component library. Its main purpose is to significantly improve the proactive predictability and accuracy of risk warning for registered products.

[0005] To achieve the above objectives, this invention provides a risk warning method for registered products based on an ingredient library, comprising:

[0006] Extract the full ingredient list of the registered products from the product registration data, and construct an ingredient database corresponding to the full ingredient list using preset ingredient matching rules;

[0007] Establish a fast component retrieval index and risk feature set for the component database, and match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set;

[0008] Calculate the compliance correlation between each component in the full ingredient list and the potential risk component, and construct a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation.

[0009] Based on the multidimensional risk profile, the risk tracing characteristics of the ingredient database are determined, and based on the risk tracing characteristics, a risk assessment conclusion corresponding to the registered product is generated;

[0010] Based on the multi-dimensional risk profile and the risk assessment conclusion, the risk warning result of the registered product is output.

[0011] Optionally, a component database corresponding to the full component list is constructed using preset component matching rules, including:

[0012] The target risk parameters corresponding to each risk component category are parsed from the component matching rules to calculate the matching metric value of each component in the full component list to each risk component category.

[0013] Obtain the incompatibilities between the components in the full ingredient list from the preset compatibility rule base;

[0014] Based on the incompatibilities and the matching metric, generate the component risk feature vector corresponding to the full component list;

[0015] Based on the component risk feature vector, a component database corresponding to the full component list is constructed.

[0016] Optionally, based on the incompatibilities and the matching metric, a component risk feature vector corresponding to the full component list is generated, including:

[0017] Based on the matching quantification and the incompatibilities, calculate the risk synergy coefficients among the components in the full component list;

[0018] Based on the risk synergy coefficient and the preset system security target threshold, the system risk feature vector corresponding to the full component list is determined.

[0019] Optionally, based on the matching metric and the incompatibility relationship, the risk synergy coefficients among the components in the full component list are calculated, including:

[0020] Based on the incompatibilities, analyze the interaction strength of each component in the full ingredient list;

[0021] The compatibility influencing factors of each component in the full component list are identified by the interaction strength.

[0022] Based on the compatibility influence factor, the risk synergy coefficient among the components in the full component list is calculated.

[0023] Optionally, establishing a fast component retrieval index for the component database includes:

[0024] The standard naming identifiers and attribute features of each component in the component database are parsed out to generate a set of component feature vectors for the component database;

[0025] Identify the root words in the standard naming identifier to generate a weighted inverted index table of the component database;

[0026] Based on the set of component feature vectors and the weighted inverted index table, a fast component retrieval index is established for the component database.

[0027] Optionally, the component feature vector set includes chemical structure descriptor vectors, toxicological parameter vectors, and functional classification encoding vectors for each component.

[0028] Optionally, based on the compliance correlation, a multi-dimensional risk profile is constructed corresponding to the registered product, including:

[0029] Based on the aforementioned compliance correlation, the coordinates of each risk component of the registered product in the risk space are determined.

[0030] Based on the coordinate positioning of each risk component in the risk space, the risk topology of the registered product is generated.

[0031] The interaction relationships between the risk components are analyzed, and a multi-dimensional risk profile corresponding to the registered product is constructed based on the interaction relationships and the risk topology.

[0032] Optionally, the risk topology is used to quantify the spatial distribution relationship and transmission path of the risk components.

[0033] Optionally, the transmission path is determined by analyzing the compatibility correlation strength between the risk components and then based on the direction and normalized strength value of the compatibility correlation strength.

[0034] To address the aforementioned problems, the present invention also provides a risk warning system for registered products based on an ingredient library, the system comprising:

[0035] The ingredient library construction module is used to extract the full ingredient list of the registered products from the product registration data and construct the ingredient database corresponding to the full ingredient list using preset ingredient matching rules;

[0036] The risk matching module is used to establish a fast component retrieval index and risk feature set for the component database, and to match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set.

[0037] The risk assessment module is used to calculate the compliance correlation between each component in the full ingredient list and the potential risk components, and to construct a multi-dimensional risk profile for the registered product based on the compliance correlation.

[0038] The conclusion generation module is used to determine the risk tracing characteristics of the component database based on the multi-dimensional risk profile, and generate a risk assessment conclusion corresponding to the registered product based on the risk tracing characteristics.

[0039] The early warning output module is used to output the risk warning result of the registered product based on the multi-dimensional risk profile and the risk assessment conclusion.

[0040] Compared to the problems described in the background art, this embodiment of the invention, by utilizing preset ingredient matching rules to construct an ingredient database corresponding to the full ingredient list, can intelligently link scattered registration information with a standardized ingredient knowledge base, forming a systematic and standardized risk identification mechanism. Furthermore, by establishing a rapid ingredient retrieval index and risk feature set for the ingredient database, this embodiment of the invention can break the inefficient traditional manual ingredient data retrieval mode, forming a data retrieval system that can quickly locate and intelligently link components. This provides an efficient analytical foundation for real-time capture of the matching relationship between the full ingredient list and risk rules, ensuring rapid response to the complexity of ingredients and dynamic updates in regulations. Based on the compliance correlation, this embodiment of the invention constructs a multi-dimensional risk profile corresponding to the registered product, transforming abstract risk assessment data into a visualized and quantifiable data. The analytical composite risk characterization accurately correlates the compliance status of components with the product safety situation. Furthermore, this embodiment of the invention generates risk assessment conclusions corresponding to the registered products based on the risk tracing characteristics, accurately linking the potential hazards of components, actual compliance status, and the overall safety level of the product, providing authoritative and transparent decision-making basis for product access, risk management, and subsequent improvements. Finally, this embodiment of the invention outputs risk warning results for the registered products based on the multi-dimensional risk profile and the risk assessment conclusions, allowing the product risk analysis process to deeply align with the dynamic update requirements of component safety standards and the integration and adaptation needs of historical risk evidence. Simultaneously, it avoids risk underreporting caused by reliance on manual experience judgment or risk management delays caused by response lags, significantly improving the proactive predictability and accuracy of risk warnings. Therefore, the component library-based risk warning method and system for registered products provided by this embodiment of the invention can significantly improve the proactive predictability and accuracy of risk warnings for registered products. Attached Figure Description

[0041] Figure 1 A flowchart illustrating a risk warning method for registered products based on an ingredient library, provided in an embodiment of the present invention;

[0042] Figure 2 This is a status flow diagram of a product registration method based on a component library, provided in an embodiment of the present invention.

[0043] Figure 3 This is a schematic diagram of a module for implementing a risk warning system for registered products based on a component library, provided as an embodiment of the present invention.

[0044] Figure 4This is an internal structural diagram of a computer device for implementing a risk warning system for registered products based on a component library, as provided in an embodiment of the present invention.

[0045] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0046] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0047] This application provides a method for risk warning of registered products based on an ingredient library. The executing entity of this method includes, but is not limited to, at least one electronic device that can be configured to execute the method provided in this application, such as a server or a terminal. In other words, the method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0048] Reference Figure 1 The diagram shown is a flowchart illustrating a risk warning method for registered products based on a component library, according to an embodiment of the present invention. In this embodiment, the risk warning method for registered products based on a component library includes:

[0049] S1. Extract the full ingredient list of the registered products from the product registration data, and construct the ingredient database corresponding to the full ingredient list using preset ingredient matching rules.

[0050] This invention provides traceable raw data input for subsequent automated risk warning based on an ingredient library by extracting a complete ingredient list of registered products from product registration data. Registered products refer to consumer goods that have completed the legal registration process with the competent authority and obtained a unique identification code, thus being approved for distribution. These include, but are not limited to, cosmetics, health foods, disinfection products, infant formula milk powder, special medical purpose formula foods, and medical devices. Product registration data refers to a collection of legal documents submitted by the registrant to the competent authority and approved, recording the core technical indicators of the product. Its legal form includes, but is not limited to, standardized documents such as registration certificates, product formula tables, ingredient lists, and technical specifications. The complete ingredient list refers to a collection of all ingredient information extracted from the product registration data that fully reflects the composition of the registered product's formula. It includes the standardized names and corresponding content information of all raw material components in the product formula, and the information of each component is consistent with the formula details declared during product registration.

[0051] Specifically, the process of obtaining product registration data involves judging and processing various scenarios related to the product registration status. For the specific logical flow, please refer to [link / reference needed]. Figure 2 The diagram illustrates the status flow of a product registration process according to an embodiment of the present invention, which provides a risk warning method for registered products based on a component library. This diagram abstracts core steps such as product selection approval, registration information acquisition, registration status judgment, and data processing into coherent logical branches. The status flow and data update operations constitute the pre-processing workflow, intuitively demonstrating the complete path for acquiring registration data and the multi-scenario processing mechanism. Specifically: The system automatically triggers a registration information query task based on the list of products that have passed product selection approval; it obtains the product's registration / filing information by connecting to the regulatory agency's registration database and performs branch judgments on various results such as "not found," "registered (current / historical approval)," and "registered"; furthermore, the system performs operations such as updating data, re-registering, or status labeling based on attributes such as the validity period and cancellation status of the registration status. This provides an accurate and timely registration data foundation for subsequent risk warning analysis and ensures that the registered product data input into the component library intelligent comparison stage has good quality and clear timeliness attributes, providing a key guarantee for the accuracy and reliability of risk warnings from the source.

[0052] Furthermore, this embodiment of the invention utilizes preset ingredient matching rules to construct an ingredient database corresponding to the full ingredient list. This enables the intelligent association between scattered registration information and a standardized ingredient knowledge base, forming a systematic and standardized risk identification mechanism. The ingredient matching rules refer to a pre-defined set of logical conditions used to associate product ingredients with risk characteristics. This set of logical conditions includes ingredient name standardization rules, risk ingredient identification rules, and correlation calculation rules. For example, when "phenoxyethanol" appears simultaneously in the ingredient list and its concentration is higher than one percent, the preservative compounding rule is triggered, and formaldehyde-releasing components are required to be detected simultaneously. The ingredient database refers to a structured data set dynamically constructed based on the full ingredient list and its corresponding ingredient risk feature vectors, used to store and manage product ingredient safety information. By integrating relevant data from cosmetic ingredient libraries, food ingredient libraries, and domestic and international additive standard libraries, this database can form a complete data architecture encompassing three levels: basic attributes, risk characteristics, and compatibility relationships.

[0053] Specifically, the cosmetic ingredient library is used to store standardized data of cosmetic raw materials, including fields such as the Chinese and English names of the ingredients, technical classifications, ingredient types, and intended uses, and is classified and managed according to prohibited ingredients, restricted ingredients, risky ingredients, and permitted ingredients; the food ingredient library is used to collect basic information on food ingredients and related additives, including data such as the legal name of the ingredients, functional categories, scope of use, and limit requirements; the domestic and international additive standards library is used to integrate domestic and international food additive standards and regulations, including key technical indicators such as the permitted scope of use, maximum usage, and residue standards for additives.

[0054] As an embodiment of the present invention, a component database corresponding to the full component list is constructed using preset component matching rules, including:

[0055] The target risk parameters corresponding to each risk component category are parsed from the component matching rules to calculate the matching metric value of each component in the full component list to each risk component category.

[0056] Obtain the incompatibilities between the components in the full ingredient list from the preset compatibility rule base;

[0057] Based on the incompatibilities and the matching metric, generate the component risk feature vector corresponding to the full component list;

[0058] Based on the component risk feature vector, a component database corresponding to the full component list is constructed.

[0059] The risk component category refers to a classification system based on the toxicological characteristics and regulatory restrictions of the components, including technical categories with different regulatory requirements such as prohibited substances, restricted substances, and permitted substances. The target risk parameter refers to a technical indicator used to quantitatively assess the risk level of a component. For example, a maximum permissible concentration of one percent of a preservative in leave-on cosmetics is a specific target risk parameter. The component refers to each independent chemical substance or natural extract that constitutes the product formulation. For example, glycerin, niacinamide, and phenoxyethanol listed in a skin cream formulation are all components of that product. The matching quantification value refers to a numerical result used to characterize the degree of fit between each component in the full ingredient list and its corresponding risk component category, typically ranging from 0-1 or 0-100. The compatibility rule base refers to a knowledge set recording the interaction relationships between components. For example, it may contain rules recording the chemical reaction of benzene produced when vitamin C is mixed with sodium benzoate. The incompatible relationship refers to the type of interaction that may result in safety risks or reduced efficacy when different components are mixed. For example, the interaction between acidic and alkaline components can lead to sudden pH changes and skin irritation; the component risk feature vector refers to a mathematical vector that comprehensively characterizes the risk features of a single component and its combinations. For example, a five-dimensional vector can be used to describe the sensitization risk value, cumulative toxicity risk value, compatibility risk value, concentration risk value, and overall risk level.

[0060] Optionally, the matching metric value of each component in the full component list to each of the risk component categories can be calculated using a cosine similarity algorithm; the component database corresponding to the full component list can be constructed using a MySQL relational database, such as: establishing a three-level association structure based on the MySQL database management system, including a basic component information table, a risk characteristic table, and a compatibility table, realizing the association mapping between component identifiers and risk parameters through primary and foreign key constraints, and establishing a fast retrieval index based on the component name hash value, thereby completing the construction of the component database.

[0061] As another embodiment of the present invention, a component risk feature vector corresponding to the full component list is generated based on the incompatibility relationship and the matching metric value, including:

[0062] Based on the matching quantification and the incompatibilities, calculate the risk synergy coefficients among the components in the full component list;

[0063] Based on the risk synergy coefficient and the preset system security target threshold, the system risk feature vector corresponding to the full component list is determined.

[0064] The risk synergy coefficient refers to a comprehensive risk synergy effect index obtained by quantitatively analyzing the interaction relationships between the components in the full ingredient list; the system safety target threshold refers to a systemic risk tolerance limit pre-set according to product safety standards and regulatory requirements. For example, the limit standard for the total amount of eight heavy metals in infant skin care products shall not exceed 10 parts per million. Optionally, the system safety target threshold can be determined comprehensively based on safety limit data, toxicological reference values, and product usage scenario characteristics in the ingredient database.

[0065] As another embodiment of the present invention, based on the matching quantification value and the incompatibility relationship, the risk synergy coefficients among the components in the full component list are calculated, including:

[0066] Based on the incompatibilities, analyze the interaction strength of each component in the full ingredient list;

[0067] The compatibility influencing factors of each component in the full component list are identified by the interaction strength.

[0068] Based on the compatibility influence factor, the risk synergy coefficient among the components in the full component list is calculated.

[0069] The interaction strength refers to a quantitative indicator of the degree to which different components produce chemical or biological effects after mixing. For example, the quantitative data showing a 30% increase in transdermal absorption rate after mixing ethanol and phenoxyethanol. The compatibility influence factor refers to a characteristic parameter reflecting the direction and degree of interaction between components. For example, the rate constant of the redox reaction between vitamin C and metal ions.

[0070] Optionally, the interaction strength of each component in the full component list can be analyzed by molecular docking simulation; the compatibility influence factor of each component in the full component list can be identified by principal component analysis.

[0071] For example, the risk synergy coefficients among the components in the full ingredient list are calculated using the following formula. It should be noted that this calculation method is only one possible approach and does not affect the implementation of the basic scheme described above:

[0072]

[0073] in, This represents the risk synergy coefficient among the components in the full ingredient list. This represents the compatibility factor between component i and component j. This represents the matching metric value of component i. This represents the matching metric of the j-th component that interacts with the i-th component. Parameters representing differences in chemical properties between components The reference chemical property parameter corresponding to the chemical property difference parameter, n represents the total number of risk components in the full ingredient list, m represents the total number of component pairs with compatibility relationship, i represents the index of the risk component, and j represents the index of the paired component with compatibility relationship with the i-th component.

[0074] Specifically, the compatibility influence factor is used to distinguish between two interaction modes: synergistic enhancement and antagonistic weakening. Its value range is [-1, 1]. For example, if a preservative is combined with a surfactant, the toxicity increases. =0.6; while chelating agents reduce toxicity after binding with metal ions, their =-0.4. The matching quantification value is used to quantify the contribution of a single component to a specific risk category (such as sensitization or carcinogenicity), and its value range is [0,1]. The chemical property difference parameter represents the Euclidean distance between the physicochemical properties of the components, such as logP, molecular weight, and polarity, and is used to correct for ineffective synergistic effects caused by excessive differences in properties. It should be noted that the above formula uses an exponential decay term. Quantifying the inhibitory effect of differences in chemical properties on synergistic effects can avoid ineffective correlation assessments of components with vastly different properties. Furthermore, by integrating compatibility influence factors, quantified matching values, and differences in chemical properties into a single synergistic coefficient, a comprehensive mapping from microscopic interactions to macroscopic risks of each component can be achieved. Finally, the normalization term in the denominator is used... Automatically adjusting the contribution weight of highly matched components in the synergistic effect can eliminate the dominant influence of a single component with excessively high matching on the overall synergistic effect, ensure that the assessment results more evenly reflect the comprehensive effect of all related components, and avoid the risk assessment model being overly sensitive to specific highly matched components.

[0075] S2. Establish a fast component retrieval index and risk feature set for the component database, and match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set.

[0076] This invention, through the establishment of a rapid component retrieval index and risk feature set in the component database, can break the inefficient mode of traditional manual component data retrieval and form a data retrieval system that can quickly locate and intelligently associate components. This provides an efficient analytical foundation for real-time capture of the matching relationship between the full component list and risk rules, ensuring rapid response of risk warnings to component complexity and regulatory updates.

[0077] The fast component retrieval index refers to a data structure built based on key component attributes to enable rapid queries. It maps the component's Chinese and English names, CAS numbers, molecular formulas, and other identifiers to physical storage locations in the database, establishing a hash table or B-tree index mechanism to support millisecond-level retrieval of massive component data. For example, for the component "phenoxyethanol," its CAS number "122-99-6" can immediately locate all relevant data records for that component in the database. The risk feature set refers to a collection of component risk attributes organized according to a specific classification system. It integrates toxicological data, regulatory limits, and clinical evidence to form a structured data set containing standardized risk labels such as prohibition markers, restricted concentrations, and sensitization levels. Optionally, the risk feature set of the component database can be established using a decision tree classification model. For example, using a decision tree classification model, with the component's chemical structure, toxicological data, and regulatory status as feature nodes, an "IF-THEN" rule set is constructed to achieve automatic risk classification and label generation for components.

[0078] It is important to emphasize that the ingredient database and risk feature set described in this solution integrate multi-source data from the authoritative EU risk substance limit database and the common alternative risk database to ensure that risk identification complies with both international regulatory standards and industry realities. Specifically, the data involved in the authoritative EU risk substance limit database includes, but is not limited to: the list of prohibited and restricted substances in the EU Cosmetics Regulation ((EC) No 1223 / 2009), the list of restricted substances in Annex XVII of the REACH Regulation (EC No 1907 / 2006), specific migration limits in the EU Food Contact Materials Regulation ((EC) No 1935 / 2004), and the SVHC candidate substance list published by ECHA (European Chemicals Agency); the common alternative risk database is based on long-term industry practice and includes risk data of common alternative ingredients in actual production, registration, and distribution, including: potential risk cases of commonly used alternative raw materials (such as the risk of functional analogs that replace plasticizers and preservatives), undisclosed risky combinations reported internally by the industry, and records of safety incidents involving non-standard ingredients in actual use. Therefore, this solution integrates authoritative regulatory data with industry practice data to systematically construct a risk identification mechanism driven by both "regulatory compliance" and "industry realities," which significantly improves the coverage and actual response capabilities of risk warnings.

[0079] As an embodiment of the present invention, establishing a fast component retrieval index for the component database includes:

[0080] The standard naming identifiers and attribute features of each component in the component database are parsed out to generate a set of component feature vectors for the component database;

[0081] Identify the root words in the standard naming identifier to generate a weighted inverted index table of the component database;

[0082] Based on the set of component feature vectors and the weighted inverted index table, a fast component retrieval index is established for the component database.

[0083] The standard naming identifier refers to the unique identifier defined by internationally recognized ingredient naming standards. For example, "L-ascorbic acid" is the standard naming identifier for vitamin C, and "propylparaben" is the standard naming identifier for propylparaben. The attribute characteristics refer to the physicochemical properties and functional classification parameters of the ingredients. Examples include the HLB value of surfactants and the effective pH range of preservatives. The ingredient feature vector set refers to a structured data set that converts the standardized naming identifiers and attribute characteristics of each ingredient into numerical feature vectors through vectorization methods. For example, it includes a composite vector group containing the ingredient's molecular descriptor, toxicological parameters, and functional classification code. The root word refers to the smallest word-forming unit with a specific chemical meaning that constitutes the standard naming identifier of an ingredient. For example, in the ingredient name "phenoxyethanol," "benzene," "oxygen," "ethanol," and "alcohol" are all root words with clear chemical structure indications. The weighted inverted index table is a retrieval data structure that records the correspondence between root words and ingredient identifiers and represents the strength of the association through statistical weights. This table uses the root word as the index key, with the ingredient identifier containing that root word and its weight value as the corresponding entry. For example, a hash mapping table that records all component IDs and corresponding weights corresponding to the root word "propyl".

[0084] Optionally, the component feature vector set of the component database can be generated by the BERT embedding model; the word roots in the standard naming identifier can be identified using an n-gram statistical model; and the fast component retrieval index of the component database can be established using the Faiss vector search engine.

[0085] Furthermore, this embodiment of the invention, by using the fast retrieval index of the ingredients and the risk feature set, matches the potential risk ingredients in the ingredient database that correspond to the full ingredient list. This enables accurate identification and efficient screening of ingredient risks in registered products, ensuring that the matched potentially risky ingredients, such as prohibited substances, restricted ingredients, and risky combinations, are highly correlated with the product formulation data. The potential risk ingredients refer to chemical substances or raw materials that may pose a threat to human health or product safety under specific usage conditions. It should be noted that when matching the potential risk ingredients, standard compliance screening based on the authoritative EU database is required, and enhanced identification of actual industry risks based on a database of common alternatives is necessary. In particular, it is necessary to target ingredients or combinations that have not yet been included in regulations but have already shown risk signals in the industry.

[0086] As an embodiment of the present invention, based on the fast component retrieval index and the risk feature set, potential risk components corresponding to the full component list in the component database are matched, including:

[0087] The component name sequence, component concentration information, component function type, and product usage conditions of the full ingredient list are parsed out.

[0088] Based on the component name sequence and the component function type, a batch query instruction sequence for the component database is generated through the component fast retrieval index;

[0089] The product usage conditions are mapped to the activation conditions of the screening rules of the risk feature set, wherein the product usage conditions include the application site, the dwell time, and the applicable population;

[0090] Based on the component concentration information, quantify the limit compliance parameters of each component in the risk feature set;

[0091] By integrating the batch query instruction sequence, the filtering rule activation conditions, and the limit compliance parameters, potential risk components in the component database that correspond to the full component list are matched.

[0092] The component name sequence refers to an ordered set of standardized ingredient names arranged in the product formula sequence; the component concentration information refers to the content range or precise ratio data of each ingredient in the formula, including quantitative expressions such as mass fraction, volume fraction, and mass concentration; the component function type refers to the technical category classified according to the role of the ingredient in the product. For example, ingredients are classified into functional categories such as moisturizers, preservatives, whitening agents, and emulsifiers; the product usage conditions refer to product application scenario parameters that affect the ingredient risk assessment. For example, the product usage conditions of a children's face cream are: it can be applied to the face and body as needed and is suitable for children aged 3-12 years; the batch query command sequence refers to a group of batch query commands for ingredients automatically generated through the retrieval index; the screening rule activation condition refers to the application scenario threshold that triggers a specific risk screening rule. For example, when the product usage conditions include "suitable for children", the screening rule for prohibited ingredients in children's products is activated, or when the application site is "lips", the specific limit standard for lip products is activated; the limit compliance parameter refers to the quantitative value of the degree of compliance between the actual concentration of the ingredient and the regulatory limit standard.

[0093] It should be noted that the aforementioned limits include, but are not limited to: maximum permissible concentrations, such as the requirement that "phenoxyethanol" in cosmetics must not exceed 1.0% in leave-on products and 1.5% in rinse-off products; lists of prohibited substances, such as prohibited ingredients explicitly listed in China's "Cosmetic Safety Technical Specifications," such as "mercury compounds" and "chloramphenicol"; conditions for restricted use of substances, such as the requirement that the concentration of "salicylic acid" in skincare products must not exceed 2.0%, and it can only be used on non-mucous membrane areas; limits on incompatibilities, such as the requirement that the molar ratio of "vitamin C" and "metal ions" must not exceed 1:1 to avoid oxidation reactions; and cumulative limits, such as the requirement that the total content of the eight heavy metals (lead, arsenic, mercury, cadmium, etc.) in infant skincare products must not exceed 10 mg / kg. These limits are derived from regulatory documents from authoritative agencies such as the National Medical Products Administration, the EU Cosmetics Regulation (EC) No 1223 / 2009, and the US Food and Drug Administration (FDA), and are dynamically updated in the ingredient database to ensure the timeliness and compliance of risk warnings.

[0094] Optionally, a rule-based reasoning system can be used to map the product usage conditions to the screening rule activation conditions of the risk feature set; the limit compliance parameters of each component in the risk feature set can be quantified by the hazard quotient method.

[0095] S3. Calculate the compliance correlation between each component in the full ingredient list and the potential risk component, and construct a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation.

[0096] This invention, by calculating the compliance correlation between each component in the full ingredient list and the potential risk component, can accurately map the dynamic adaptation relationship between the actual use of the components and regulatory safety standards, significantly enhancing the risk warning system's ability to quantify the compliance status of components and predict safety boundaries. The compliance correlation refers to the correlation strength index between each component in the product corresponding to the full ingredient list and the identified potential risk component at the regulatory compliance level.

[0097] Optionally, the compliance correlation between each component in the full ingredient list and the potential risk component can be calculated using a logistic regression model.

[0098] Furthermore, in this embodiment of the invention, by constructing a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation, abstract risk assessment data can be transformed into a visualized and parsable composite risk representation, accurately linking the compliance status of components with the product safety situation. The multi-dimensional risk profile refers to a structured data model used to comprehensively represent the overall risk situation of the registered product.

[0099] As an embodiment of the present invention, a multi-dimensional risk profile corresponding to the registered product is constructed based on the compliance correlation, including:

[0100] Based on the aforementioned compliance correlation, the coordinates of each risk component of the registered product in the risk space are determined.

[0101] Based on the coordinate positioning of each risk component in the risk space, the risk topology of the registered product is generated.

[0102] The interaction relationships between the risk components are analyzed, and a multi-dimensional risk profile corresponding to the registered product is constructed based on the interaction relationships and the risk topology.

[0103] The term "risk ingredient" refers to a chemical substance identified through ingredient matching that may affect product safety. For example, phenoxyethanol (a preservative) and benzophenone-3 (a sunscreen) identified in a cosmetic product are both considered risk ingredients. The term "risk space" refers to a multidimensional mathematical space constructed using risk characteristic dimensions as coordinate axes to quantify the risk distribution of ingredients. For example, a three-dimensional risk space constructed using three risk indicators—carcinogenicity, sensitization, and environmental toxicity—as orthogonal coordinate axes. The term "coordinate positioning" refers to the technical process of determining the specific location of each risk ingredient in the risk space using mathematical methods. For example, principal component analysis can be used to map the carcinogenicity data of a preservative to a set of specific coordinate values ​​in the risk space. The term "risk topology" refers to the geometric structure characterizing the distribution and connectivity of risk ingredients in the feature space, used to quantify the spatial distribution and transmission paths of the risk ingredients. For example, three high-risk ingredients in a product may form a triangular configuration in the risk space with two strongly correlated edges. The term "interaction relationship" refers to the chemical or biological effects resulting from the mixing of different risk ingredients. For example, the synergistic effect of retinol and fruit acid in a product leading to increased skin irritation.

[0104] Optionally, the interactions between the risk components can be analyzed through molecular docking simulation; the multidimensional risk profile corresponding to the registered product can be constructed using Tableau's multidimensional data visualization tool.

[0105] S4. Based on the multi-dimensional risk profile, determine the risk tracing characteristics of the component database, and generate a risk assessment conclusion corresponding to the registered product based on the risk tracing characteristics.

[0106] This invention, through the multi-dimensional risk profile, determines the risk tracing characteristics of the ingredient database, providing traceable and verifiable data support for risk cause analysis, compliance verification, and subsequent risk management of registered products. The risk tracing characteristics refer to the associated data attributes used to trace and verify the causes of risks, including the source of the ingredients' regulatory basis, historical safety incident records, indexes of relevant toxicological research literature, and risk case identifiers of similar products.

[0107] As an embodiment of the present invention, the risk tracing characteristics of the component database are determined based on the multi-dimensional risk profile, including:

[0108] Identify the risk level distribution characteristics and key risk components in the multidimensional risk profile;

[0109] Establish the correlation between the key risk components and historical security events in the component database;

[0110] Based on the aforementioned relationship, a risk evidence chain corresponding to the key risk component is constructed;

[0111] Based on the risk level distribution characteristics and the risk evidence chain, the risk tracing characteristics of the component database are determined.

[0112] The risk level distribution characteristics refer to the spatial distribution patterns and statistical characteristics of different risk levels across various component dimensions in the multidimensional risk profile; the historical safety events refer to publicly disclosed safety issues related to specific components recorded in the component database. For example, a case of sensitization notification of a certain preservative recorded in the EU cosmetics database; the risk evidence chain refers to a multidimensional set of evidence constructed based on historical safety events that systematically reflects the safety issues of key risk components.

[0113] Optionally, the association between the key risk component and historical safety events in the component database can be established using an entity association model based on knowledge graphs; the risk evidence chain corresponding to the key risk component can be constructed using a Bayesian network, such as constructing a risk evidence chain for the component "phenoxyethanol" using a Bayesian network, with the component as the root node, associating it with sub-nodes such as regulatory limit change events in different countries, clinical reports of skin sensitization, and cases of exceeding the limit and being penalized in different product types, and quantifying the causal strength between each evidence node based on a conditional probability table to form a complete risk evolution path.

[0114] Furthermore, this embodiment of the invention generates risk assessment conclusions corresponding to the registered products based on the risk tracing characteristics. This can accurately link the potential hazards of the components, their actual compliance status, and the overall safety level of the product, providing authoritative and transparent decision-making basis for product access, risk management, and subsequent improvements. The risk assessment conclusions refer to a comprehensive judgment result on the overall safety status of the registered products, generated based on the risk tracing characteristics and through preset judgment rules. Its specific form may include a structured data object containing the product risk level, a list of major risk components, risk cause analysis, and management suggestions.

[0115] As an embodiment of the present invention, based on the risk tracing characteristics, a risk assessment conclusion corresponding to the registered product is generated, including:

[0116] The historical security events corresponding to the risk tracing characteristics were analyzed;

[0117] The frequency of high-risk components in the aforementioned historical security incidents was statistically analyzed.

[0118] Read the number of associated security alerts from the risk tracing features;

[0119] Calculate the risk correlation index between the frequency of occurrence of the high-risk component and the number of associated safety alarms;

[0120] Based on the risk correlation index and the preset industry safety benchmark value, the risk assessment conclusion corresponding to the registered product is output.

[0121] The historical safety incidents refer to cases of safety issues related to a specific ingredient that have been reported by official agencies or recorded in authoritative documents. For example, a case of sensitization notification of a preservative ingredient recorded in the EU cosmetics database. The frequency of occurrence of high-risk ingredients refers to the number of times a chemical ingredient identified as having significant risk characteristics has appeared in historical safety incidents. The number of associated safety alerts refers to the total number of safety warning records directly related to the target ingredient. For example, a sunscreen ingredient has triggered five safety alerts in the China Drug Administration's warning system. The risk correlation index refers to the correlation strength indicator between the frequency of occurrence of high-risk ingredients and the number of associated safety alerts, calculated using a statistical algorithm. The industry safety benchmark value refers to a safety evaluation reference standard set according to industry regulatory requirements.

[0122] Optionally, the number of associated security alerts in the risk tracing features can be obtained by reading security notification records from the regulatory agency's database; the risk assessment conclusion corresponding to the registered product can be output using a rule engine reasoning system.

[0123] It should be noted that when generating the aforementioned risk assessment conclusions, it is necessary not only to rely on the limits and risk classifications in EU regulations, but also to incorporate industry-specific cases and early warning records from the common alternative risk database, forming a dual chain of evidence: "regulation + practice." For example, if a component is a "restricted substance" under EU regulations, but there are multiple "actual use sensitization reports" in the common database, the system will automatically upgrade its risk level and indicate in the conclusion that "risk signals have appeared in industry practice, and it is recommended to strengthen monitoring." If a substitute component is not yet explicitly restricted by EU regulations, but is marked as a "common risk substitute" in the common database, the system can still output an early warning, achieving "pre-regulatory early warning."

[0124] S5. Based on the multi-dimensional risk profile and the risk assessment conclusion, output the risk warning result of the registered product.

[0125] This invention, through its embodiment, outputs risk warning results for the registered products based on the multi-dimensional risk profile and the risk assessment conclusions. This allows the product risk assessment process to deeply align with the dynamic update requirements of component safety standards and the integration and adaptation needs of historical risk evidence. At the same time, it can avoid the problems of risk underreporting caused by relying on human experience judgment or risk handling delays caused by response lags, and significantly improve the proactive predictability and accuracy of risk warnings.

[0126] The risk warning result refers to a set of structured instructions generated based on multi-dimensional risk profiles and risk assessment conclusions to guide product risk management. This set of structured instructions includes standardized outputs of warning levels, risk positioning, handling suggestions, and traceability basis. Specifically, it includes: First, based on the risk assessment conclusions, quantifying and assigning values ​​to core parameters such as the warning level and risk positioning of the registered product to obtain standardized warning instructions; the warning instructions output control parameters to the product management system, such as specific risk level codes, risk component lists, and handling priorities; then, the product management system executes corresponding risk management measures according to the warning instructions, such as automatically intercepting high-risk products, triggering manual review processes, and restricting shelf access, and recording the handling process and results in real time through system logs; finally, the system structurally integrates key node data during the handling process, such as automatically associating risk components with raw material supplier files, matching warning records with historical cases, and combining this with the implemented control measures to form a complete risk warning result file.

[0127] like Figure 3 The diagram shown is a functional block diagram of a risk warning system for registered products based on an ingredient library, according to the present invention.

[0128] The ingredient library-based risk warning system 200 for registered products described in this invention can be installed in an electronic device. Depending on the functions implemented, the ingredient library-based risk warning system may include an ingredient library construction module 201, a risk matching module 202, a risk assessment module 203, a conclusion generation module 204, and a warning output module 205. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0129] In this embodiment of the invention, the functions of each module / unit are as follows:

[0130] The ingredient library construction module 201 is used to extract the full ingredient list of the registered products from the product registration data and construct the ingredient database corresponding to the full ingredient list using preset ingredient matching rules;

[0131] The risk matching module 202 is used to establish a fast component retrieval index and a risk feature set for the component database, and to match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set.

[0132] The risk assessment module 203 is used to calculate the compliance correlation between each component in the full ingredient list and the potential risk component, and to construct a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation.

[0133] The conclusion generation module 204 is used to determine the risk tracing characteristics of the component database based on the multi-dimensional risk profile, and generate a risk assessment conclusion corresponding to the registered product based on the risk tracing characteristics.

[0134] The early warning output module 205 is used to output the risk warning result of the registered product based on the multi-dimensional risk profile and the risk assessment conclusion.

[0135] In detail, the modules in the ingredient library-based risk warning system 200 for registered products described in this embodiment of the invention employ the same methods as described above. Figure 1 The method used here is the same as the one described in the article on risk warning for registered products based on a component library, and can produce the same technical effect, so it will not be elaborated here.

[0136] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a server-side method for risk warning of registered products based on a component library.

[0137] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0138] Obtain the full ingredient list of the registered product, and construct an ingredient database corresponding to the full ingredient list using preset ingredient matching rules;

[0139] Establish a fast component retrieval index and risk feature set for the component database, and match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set;

[0140] Calculate the compliance correlation between each component in the full ingredient list and the potential risk component, and construct a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation.

[0141] Based on the multidimensional risk profile, the risk tracing characteristics of the ingredient database are determined, and based on the risk tracing characteristics, a risk assessment conclusion corresponding to the registered product is generated;

[0142] Based on the multi-dimensional risk profile and the risk assessment conclusion, the risk warning result of the registered product is output.

[0143] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0144] Obtain the full ingredient list of the registered product, and construct an ingredient database corresponding to the full ingredient list using preset ingredient matching rules;

[0145] Establish a fast component retrieval index and risk feature set for the component database, and match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set;

[0146] Calculate the compliance correlation between each component in the full ingredient list and the potential risk component, and construct a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation.

[0147] Based on the multidimensional risk profile, the risk tracing characteristics of the ingredient database are determined, and based on the risk tracing characteristics, a risk assessment conclusion corresponding to the registered product is generated;

[0148] Based on the multi-dimensional risk profile and the risk assessment conclusion, the risk warning result of the registered product is output.

[0149] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0150] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0151] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0152] 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 present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0153] Finally, it should be noted that in the above embodiments, each embodiment can be combined with each other or independent. Deleting any one of them will not affect the technical implementation of other embodiments. The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for risk warning of registered products based on an ingredient library, characterized in that, The method includes: Extract the full ingredient list of the registered products from the product registration data, and construct an ingredient database corresponding to the full ingredient list using preset ingredient matching rules. Constructing the ingredient database corresponding to the full ingredient list using preset ingredient matching rules includes: The target risk parameters corresponding to each risk component category are parsed from the component matching rules to calculate the matching metric value of each component in the full component list to each risk component category. Obtain the incompatibilities between the components in the full ingredient list from the preset compatibility rule base; Based on the incompatibilities and the matching metric, a component risk feature vector corresponding to the full component list is generated. This generation of the component risk feature vector corresponding to the full component list includes: Based on the metric of matching and the incompatibilities, the risk synergy coefficients among the components in the full component list are calculated, wherein calculating the risk synergy coefficients among the components in the full component list based on the metric of matching and the incompatibilities includes: Based on the incompatibilities, analyze the interaction strength of each component in the full ingredient list; The compatibility influencing factors of each component in the full component list are identified by the interaction strength. Based on the compatibility influence factor, the risk synergy coefficient among the components in the full component list is calculated using the following formula: ； in, This represents the risk synergy coefficient among the components in the full ingredient list. This represents the compatibility factor between component i and component j. This represents the matching metric value of component i. This represents the matching metric of the j-th component that interacts with the i-th component. Parameters representing differences in chemical properties between components The reference chemical property parameter corresponding to the chemical property difference parameter, n represents the total number of risk components in the full ingredient list, m represents the total number of component pairs with compatibility relationship, i represents the index of the risk component, and j represents the index of the paired component that has a compatibility relationship with the i-th component. Based on the risk synergy coefficient and the preset system security target threshold, determine the system risk feature vector corresponding to the full component list; Based on the component risk feature vector, construct a component database corresponding to the full component list; Establish a fast component retrieval index and risk feature set for the component database. Based on the fast component retrieval index and the risk feature set, match potential risk components in the component database that correspond to the full component list. The matching of potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set includes: The component name sequence, component concentration information, component function type, and product usage conditions of the full ingredient list are parsed out. Based on the component name sequence and the component function type, a batch query instruction sequence for the component database is generated through the component fast retrieval index; The product usage conditions are mapped to the activation conditions of the screening rules of the risk feature set, wherein the product usage conditions include the application site, the dwell time, and the applicable population; Based on the component concentration information, quantify the limit compliance parameters of each component in the risk feature set; By integrating the batch query instruction sequence, the filtering rule activation conditions, and the limit compliance parameters, potential risk components in the ingredient database that correspond to the full ingredient list are matched. Calculate the compliance correlation between each component in the full ingredient list and the potential risk component, and construct a multi-dimensional risk profile corresponding to the registered product based on the compliance correlation. Based on the multidimensional risk profile, the risk tracing characteristics of the component database are determined, wherein determining the risk tracing characteristics of the component database based on the multidimensional risk profile includes: Identify the risk level distribution characteristics and key risk components in the multidimensional risk profile; Establish the correlation between the key risk components and historical security events in the component database; Based on the aforementioned relationship, a risk evidence chain corresponding to the key risk component is constructed; Based on the risk level distribution characteristics and the risk evidence chain, the risk tracing characteristics of the component database are determined, and based on the risk tracing characteristics, a risk assessment conclusion corresponding to the registered product is generated. Based on the multi-dimensional risk profile and the risk assessment conclusion, the risk warning result of the registered product is output.

2. The method for risk warning of registered products based on a component library as described in claim 1, characterized in that, Establishing a fast component retrieval index for the component database includes: The standard naming identifiers and attribute features of each component in the component database are parsed out to generate a set of component feature vectors for the component database; Identify the root words in the standard naming identifier to generate a weighted inverted index table of the component database; Based on the set of component feature vectors and the weighted inverted index table, a fast component retrieval index is established for the component database.

3. The method for risk warning of registered products based on a component library as described in claim 2, characterized in that, The component feature vector set includes chemical structure descriptor vectors, toxicological parameter vectors, and functional classification coding vectors for each component.

4. The method for risk warning of registered products based on a component library as described in claim 1, characterized in that, Based on the aforementioned compliance correlation, a multi-dimensional risk profile is constructed for the registered product, including: Based on the aforementioned compliance correlation, the coordinates of each risk component of the registered product in the risk space are determined. Based on the coordinate positioning of each risk component in the risk space, the risk topology of the registered product is generated. The interaction relationships between the risk components are analyzed, and a multi-dimensional risk profile corresponding to the registered product is constructed based on the interaction relationships and the risk topology.

5. The method for risk warning of registered products based on a component library as described in claim 4, characterized in that, The risk topology is used to quantify the spatial distribution and transmission path of the risk components.

6. The method for risk warning of registered products based on a component library as described in claim 5, characterized in that, The transmission path is determined by analyzing the compatibility correlation strength between the risk components and then using the direction and normalized strength value of the compatibility correlation strength.

7. A risk warning system for registered products based on an ingredient library, used to execute a risk warning method for registered products based on an ingredient library as described in any one of claims 1-6, characterized in that, The system includes: The ingredient library construction module is used to extract the full ingredient list of registered products from product registration data, and construct an ingredient database corresponding to the full ingredient list using preset ingredient matching rules. The construction of the ingredient database corresponding to the full ingredient list using preset ingredient matching rules includes: The target risk parameters corresponding to each risk component category are parsed from the component matching rules to calculate the matching metric value of each component in the full component list to each risk component category. Obtain the incompatibilities between the components in the full ingredient list from the preset compatibility rule base; Based on the incompatibilities and the matching metric, a component risk feature vector corresponding to the full component list is generated. This generation of the component risk feature vector corresponding to the full component list includes: Based on the metric of matching and the incompatibilities, the risk synergy coefficients among the components in the full component list are calculated, wherein calculating the risk synergy coefficients among the components in the full component list based on the metric of matching and the incompatibilities includes: Based on the incompatibilities, analyze the interaction strength of each component in the full ingredient list; The compatibility influencing factors of each component in the full component list are identified by the interaction strength. Based on the compatibility influence factor, the risk synergy coefficient among the components in the full component list is calculated using the following formula: ； in, This represents the risk synergy coefficient among the components in the full ingredient list. This represents the compatibility factor between component i and component j. This represents the matching metric value of component i. This represents the matching metric of the j-th component that interacts with the i-th component. Parameters representing differences in chemical properties between components The reference chemical property parameter corresponding to the chemical property difference parameter, n represents the total number of risk components in the full ingredient list, m represents the total number of component pairs with compatibility relationship, i represents the index of the risk component, and j represents the index of the paired component that has a compatibility relationship with the i-th component. Based on the risk synergy coefficient and the preset system security target threshold, determine the system risk feature vector corresponding to the full component list; Based on the component risk feature vector, construct a component database corresponding to the full component list; The risk matching module is used to establish a fast component retrieval index and a risk feature set for the component database, and to match potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set. The matching of potential risk components in the component database that correspond to the full component list based on the fast component retrieval index and the risk feature set includes: The component name sequence, component concentration information, component function type, and product usage conditions of the full ingredient list are parsed out. Based on the component name sequence and the component function type, a batch query instruction sequence for the component database is generated through the component fast retrieval index; The product usage conditions are mapped to the activation conditions of the screening rules of the risk feature set, wherein the product usage conditions include the application site, the dwell time, and the applicable population; Based on the component concentration information, quantify the limit compliance parameters of each component in the risk feature set; By integrating the batch query instruction sequence, the filtering rule activation conditions, and the limit compliance parameters, potential risk components in the ingredient database that correspond to the full ingredient list are matched. The risk assessment module is used to calculate the compliance correlation between each component in the full ingredient list and the potential risk components, and to construct a multi-dimensional risk profile for the registered product based on the compliance correlation. The conclusion generation module is used to determine the risk tracing characteristics of the component database based on the multidimensional risk profile, wherein determining the risk tracing characteristics of the component database based on the multidimensional risk profile includes: Identify the risk level distribution characteristics and key risk components in the multidimensional risk profile; Establish the correlation between the key risk components and historical security events in the component database; Based on the aforementioned relationship, a risk evidence chain corresponding to the key risk component is constructed; Based on the risk level distribution characteristics and the risk evidence chain, the risk tracing characteristics of the component database are determined, and based on the risk tracing characteristics, a risk assessment conclusion corresponding to the registered product is generated. The early warning output module is used to output the risk warning result of the registered product based on the multi-dimensional risk profile and the risk assessment conclusion.

Images