Big data enabled consumer goods quality tracing and risk early warning management and control system

Abstract

The invention discloses a big data enabled consumer goods quality tracing and risk early warning management and control system, and belongs to the technical field of consumer goods quality management and control. The big data enabled consumer goods quality tracing and risk early warning management and control system comprises a multi-source heterogeneous data processing module, a full-chain dynamic tracing module, an intelligent dynamic early warning module and a risk pre-judgment module. The problems that in the prior art, quality risks cannot be avoided in advance, problems can only be found during spot check, and management and control timeliness and foresight are seriously insufficient are solved. According to the method, the granularity of the whole life cycle from raw material purchase to terminal consumption is accurately traced, a source node can be quickly locked when a quality problem occurs, the disposal efficiency and accuracy are improved, a differentiated plan is configured according to an early warning level, supervision resources are optimized, a historical trend is analyzed by using a time sequence prediction algorithm, potential risks are pre-judged, and the quality of the product is improved. Quality control is promoted to be converted from passive post-treatment to active pre-prevention, and the initiative and effectiveness of risk control are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of consumer product quality control, in particular to a consumer product quality traceability and risk early warning control system empowered by big data. BACKGROUND

[0002] With the rapid development of the consumer product market, consumers' attention to product quality is increasing, and consumer product quality traceability and risk early warning control have become a core link to protect market order and consumers' rights and interests.

[0003] A Chinese patent with publication number CN106296230A discloses a daily consumer product quality safety monitoring system and method, which includes a business entity identity subsystem, a product subsystem, a sales subsystem, and a quality monitoring subsystem. Each subsystem has its own communication module, realizes decentralized connection and forms a communication network, and communicates in a broadcast manner. Each subsystem has its own synchronization module and database, and the synchronization module sends broadcast update content to other subsystems when the database is updated. In addition, each subsystem also has an authentication module that authenticates the received update content and updates the content to its own database after successful authentication.

[0004] The above-mentioned patent can only provide spot check product information based on existing data in actual use, which belongs to a passive monitoring mode of post-checking and does not have a proactive risk early warning design, so it cannot issue an alarm at the quality problem germination stage, thereby failing to avoid quality risks in advance and only discovering problems during spot checks, which seriously lacks timeliness and foresight in control. Therefore, it does not meet the existing needs, and for this reason, we propose a consumer product quality traceability and risk early warning control system empowered by big data. SUMMARY

[0005] The present application aims to provide a consumer product quality traceability and risk early warning control system empowered by big data, which builds a full-chain dynamic traceability and risk transmission path identification mechanism, and uses industrial internet identification analysis technology to assign a unique dynamic code to each consumer product, realizes granular accurate traceability from raw material procurement to terminal consumption full life cycle, builds a risk transmission model combined with dynamic coding, accurately locates the risk source and clarifies the transmission path, quickly locks the source node when quality problems occur, traces the diffusion path and clarifies the link responsibility, improves disposal efficiency and accuracy, at the same time, constructs an intelligent dynamic early warning and hierarchical control mechanism, uses a random forest algorithm to dynamically calculate risk values to improve evaluation accuracy, configures differentiated plans according to early warning levels to optimize supervision resources, uses time series prediction algorithm to analyze historical trends to predict potential risks, promotes quality control from passive post-disposal to proactive prevention, improves the initiative and effectiveness of risk control, and solves the problems raised in the above background technology.

[0006] In order to achieve the above object, the present application provides the following technical scheme: a big data enabled consumer product quality traceability and risk early warning management and control system, comprising: A multi-source heterogeneous data processing module is used to acquire multi-source data of consumer products to form a global data pool, and after structured processing and standardization, the data is stored in a distributed manner by using blockchain encryption; A full-chain dynamic traceability module is used to dynamically encode consumer products by using industrial internet identification, build a risk transmission model, and trace the consumer products with quality problems by using dynamic coding and the risk transmission model; An intelligent dynamic early warning module is used to fuse three types of risk indicators, i.e., basic, dynamic and management, dynamically calculate risk values by using a random forest algorithm, generate early warning levels and trigger differentiated plans; A risk prediction module is used to analyze historical data trends by using a time series prediction algorithm and predict potential risks in advance.

[0007] Preferably, the multi-source heterogeneous data processing module specifically comprises: Industrial internet identification analysis data, internet of things real-time sensing data, consumer multi-channel feedback data, supervision sampling data and industry standard data are acquired to form a global data pool; The acquired structured data is subjected to format verification and consistency correction, and for unstructured text data, natural language processing algorithm is used for word segmentation, entity recognition, keyword extraction and semantic normalization processing to convert it into structured data; A field mapping relationship library of different data sources is established by using dynamic data mapping technology, and corresponding mapping rules are automatically matched according to the data source type to uniformly standardize different format data; After processing, the standardized data is stored in a distributed manner by using blockchain encryption technology.

[0008] Preferably, the uniform standardization processing of different format data specifically comprises: A cross-domain data element model is constructed to define the core data element, data type, data length and value range of different domain data; A dynamic mapping rule library is established to develop mapping rules between data elements for heterogeneous data formats of different enterprises and different departments; A dynamic mapping engine is used to automatically map and convert heterogeneous data according to the rules in the mapping rule library to convert data of different formats into standardized data conforming to the cross-domain data element model.

[0009] Preferably, the construction process of the risk transmission model specifically comprises: The full-chain data in the global data pool is acquired, each link of the industry chain is taken as a node, and the data correlation between each link is taken as an edge to construct an industry chain node correlation graph; The industry chain node correlation graph is trained based on a graph neural network algorithm, potential correlation relationships between each node are mined, and risk transmission weights of each node are determined; When receiving quality problem feedback, the dynamic coding information of the problem consumer goods is extracted, and the corresponding core node is located; Based on the trained risk transmission model, the upstream associated nodes of the core node are traced back in reverse, and the risk source node is determined; The downstream associated nodes of the core node are traced forward, the full path of risk diffusion is combed, and the risk source node information, risk diffusion path information and corresponding responsibility subject information of each node are output.

[0010] Preferably, the intelligent dynamic early warning module specifically comprises: The basic risk indicators, dynamic risk indicators and management risk indicators are acquired, and the basic risk indicators, dynamic risk indicators and management risk indicators are fused; The risk value of the fused risk indicators is dynamically calculated through a random forest algorithm, an early warning level signal is automatically generated according to the calculated dynamic risk value, and a differentiated disposal plan is triggered for different early warning levels.

[0011] Preferably, the process of risk value dynamic calculation specifically comprises: The definition, data source and quantitative value of the basic risk indicators, dynamic risk indicators and management risk indicators are determined; The subjective weights of the basic risk indicators, dynamic risk indicators and management risk indicators are determined through an analytic hierarchy process; The objective weights of each indicator are calculated based on historical data through an entropy weight method, the subjective weights and the objective weights are weighted and fused to obtain the comprehensive weights of each indicator; A risk value calculation model is constructed based on a random forest algorithm, the quantitative values of each indicator and the corresponding comprehensive weights are input into the model, and the real-time risk value of the consumer goods is calculated.

[0012] Preferably, the process of automatically generating an early warning level signal according to the calculated dynamic risk value and triggering a differentiated disposal plan for different early warning levels specifically comprises A risk value threshold range is preset, the early warning level is determined according to the threshold range where the risk value is located, and the corresponding disposal plan is triggered for different early warning levels; The risk value is recalculated by collecting the sampling data, complaint data and public opinion data after disposal, and it is judged whether the risk is effectively controlled; If the risk value does not decrease below the threshold value of the corresponding early warning level, the disposal plan is adjusted and the disposal measures are strengthened.

[0013] Preferably, the process of early prediction of potential risks specifically includes: Collecting historical data in the global data pool, cleaning, deduplicating, and standardizing the historical data, and screening out feature data related to potential risks; Using a time series prediction algorithm to construct a potential risk prediction logic, dividing the processed historical feature data into training data and validation data according to time series, optimizing the prediction logic parameters with the training data, and verifying the effectiveness of the prediction logic with the validation data; Inputting real-time collected dynamic data into the optimized time series prediction logic to predict the risk value change trend in the future period and identify the potential risk type; When predicting potential risks, push forward-looking early warning information to regulatory departments and enterprises, and provide special prevention and control suggestions for potential risks.

[0014] Preferably, after obtaining the multi-source data of consumer goods, it further includes: Standardizing the multi-source data, and time series slicing the standardized multi-source data according to data sources and business links to form an initial time series set; Applying a causal relationship discovery algorithm to construct an association relationship network between sequences in the initial time series set, wherein nodes represent sequences and edges represent statistically significant guiding or influencing relationships; Based on the association relationship network, performing collaborative quality inspection on sequence groups with associations in the initial time series set, identifying low-quality sequences and outputting a high-quality sequence dataset; Mining key risk indicators and dynamic risk patterns from the high-quality sequence dataset, and retrieving historical risk events according to the key risk indicators; Marking the historical risk events as key points on a time axis, and extracting all sequence data segments within the time window before and after the event as analysis samples; Using a time series pattern mining algorithm and regression analysis to identify a frequently occurring and strongly correlated risk indicator combination change pattern from the analysis samples as a dynamic risk pattern; Calculating the fitting coefficients of each single sequence element in the high-quality sequence dataset and the dynamic risk pattern, and selecting a key sequence element set according to the fitting coefficients; For each element in the key sequence element set, based on its historical performance data, rolling calculating a statistical control limit, and setting a specification limit according to business rules; Real-time detecting the sequence value of the key element, generating a graded early warning signal according to the comparison result of the sequence value and the specification limit of the key element, and generating an early abnormality prompt according to the situation of touching the control limit; Based on early anomaly alerts and historical trends of key elements, a time-series prediction algorithm is used to comprehensively predict potential risks.

[0015] Preferably, before acquiring multi-source consumer product data to form a comprehensive data pool, and after structured processing and standardization, and before employing blockchain-based encrypted distributed storage, the following steps are also included: Determine the historical production batches and outgoing defect parameters of consumer products, and identify multiple potential abnormal quality events of consumer products based on the historical production batches and outgoing defect parameters; Obtain the influence propagation intensity of each potential abnormal quality event at each node in the industry chain, and determine the risk observation value of each potential abnormal quality event based on the influence propagation intensity of each node; The developmental impact factors for each potential abnormal quality event are determined based on risk observations. Obtain abnormal observation data for each potential abnormal quality event from multi-source consumer product data, and obtain user evaluation data for key abnormal data; Determine the user's confidence and support for the abnormal key data based on the evaluation data; The uncertainty factor of the source parameters of multi-source consumer product data is calculated based on users' confidence and support for the key abnormal data and the development impact factors of each potential abnormal event: Where Q represents the uncertainty factor of the source parameter of the multi-source data for consumer goods, N represents the number of potential abnormal quality events, and i represents the i-th potential abnormal quality event. Let be the development impact factor of the i-th potential abnormal quality event. This represents the confidence level related weight, with a value of 0.6. This represents the user's confidence level regarding the critical data related to potential abnormal quality events (i items). This represents the support-related weight, with a value of 0.4. Let represent the user's support for the key data of the i-th potentially abnormal quality event, ln represent the natural logarithm, k represent the baseline fluctuation factor of the consumer goods supply chain, and S represent the steady-state connectivity index of the consumer goods supply chain. This refers to the controllable factors in the development of consumer product quality incidents; The uncertainty factor of the source parameter of the multi-source data of consumer products is compared with a preset threshold. If the uncertainty factor is higher than the preset threshold, an enhanced data collection instruction is generated, and the specific data source that needs to be collected is determined based on the enhanced data collection instruction. Collect relevant parameters from a specific data source and insert them into multi-source data.

[0016] Compared with the prior art, the beneficial effects of the present invention are: The application realizes accurate tracking of the whole life cycle of consumer goods and clear identification of the risk transmission path by constructing a full-chain dynamic tracking and risk transmission path identification mechanism, and based on the industrial internet identification analysis technology, a unique dynamic code is assigned to each consumer good, and accurate tracking from raw material procurement to terminal consumption can be realized; by constructing a risk transmission model combined with dynamic coding, the problems of inaccurate risk source positioning and unclear risk transmission path analysis are effectively solved; when quality problems occur, the risk source node can be quickly locked, the complete path of risk diffusion is traced in a forward direction, the responsibility of each link is clarified, and the efficiency and accuracy of quality problem disposal are significantly improved; by constructing an intelligent dynamic early warning and hierarchical control mechanism, dynamic risk assessment, hierarchical control and forward-looking prediction are realized, the accuracy of risk assessment is significantly improved by using the random forest algorithm to realize dynamic calculation of risk value, different disposal plans are configured for different early warning levels to realize accurate allocation of supervision resources, and by using the time series prediction algorithm to analyze the historical data trend, the potential risks are predicted, the quality risk control is changed from post-disposal to pre-prevention, and the initiative and effectiveness of risk control are effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0017] Fig. 1 Module diagram of the big data enabled consumer goods quality tracking and risk early warning control system of the application; Fig. 2 Flowchart of the big data enabled consumer goods quality tracking and risk early warning control system of the application. DETAILED DESCRIPTION

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

[0019] In order to solve the problem that the prior art cannot avoid quality risks in advance, but can only find problems during spot checks, and the timeliness and forward-looking of control are seriously insufficient, please refer to Figs. 1-2 The embodiment provides the following technical solutions: The big data enabled consumer goods quality tracking and risk early warning control system comprises: A multi-source heterogeneous data processing module is used to acquire multi-source data of consumer goods to form a global data pool, and after structured processing and standardization, the data is stored by using block chain encryption distributed storage; A full-chain dynamic traceability module is configured to dynamically encode consumer goods by using industrial internet identification, build a risk transmission model, trace consumer goods with quality problems by using dynamic coding and the risk transmission model, and assign a unique dynamic code to each consumer good based on industrial internet identification analysis technology, mine the relevance of data of each industry chain node by using a graph neural network algorithm, and build a risk transmission model; when a quality problem occurs, the full life cycle data of the risk consumer goods are located based on the dynamic coding, the risk source node is quickly located by using the risk transmission model, and the full path of risk diffusion is traced in a forward direction to determine the responsibility subject of each link; An intelligent dynamic early warning module is configured to fuse three types of risk indicators, dynamically calculate risk values by using a random forest algorithm, generate early warning levels, and trigger differentiated plans; A risk prediction module is configured to analyze historical data trends by using a time series prediction algorithm and predict potential risks in advance.

[0020] A multi-source heterogeneous data processing module specifically includes: Industrial internet identification analysis data, internet of things real-time sensing data, consumer multi-channel feedback data, regulatory sampling data, and industry standard data are obtained to form a global data pool; the internet of things real-time sensing data includes cold chain temperature and humidity data, warehouse environment data, and transportation trajectory data; the consumer multi-channel feedback data includes e-commerce evaluation data, 12315 complaint data, and public opinion data; The obtained structured data is subjected to format verification and consistency correction, and the unstructured text data is subjected to word segmentation, entity recognition, keyword extraction, and semantic normalization processing by using a natural language processing algorithm to convert it into structured data; A field mapping relationship library of different data sources is established by using a dynamic data mapping technology, and corresponding mapping rules are automatically matched according to the data source type to uniformly standardize different format data; The standardized data is subjected to distributed storage by using a blockchain encryption technology to realize traceability and non-tamperability of the whole process of data collection, transmission, and use; enterprise business secret data and consumer privacy data are processed by using a data desensitization algorithm to ensure data security. Unstructured data cleaning specifically includes: the collected unstructured text data is preprocessed, including removing special characters and stop words, text segmentation, and part-of-speech tagging; A text classification model is built based on a natural language processing technology to classify the preprocessed text data, and the classification categories include consumer quality complaints, consumer product function failures, service quality feedback, positive evaluations, and neutral evaluations; A named entity recognition model is constructed to extract key entity information from the classified text data, including consumer product name, consumer product model, fault part, fault phenomenon, complaint time, complainant region, and involved enterprise name. The extracted key entity information is structured and converted according to a preset data structure to generate standardized structured data and stored in a global data pool.

[0021] Different formats of data are standardized, including: A cross-domain data meta-model is constructed to define the core data elements, data types, data lengths, and value ranges of different domain data, including production, circulation, consumption, and supervision domains. A dynamic mapping rule library is established to develop mapping rules between data elements for different enterprises and departments with heterogeneous data formats; the mapping rules include one-to-one, one-to-many, and many-to-one mapping rules. A dynamic mapping engine is used to automatically map and convert heterogeneous data according to the rules in the mapping rule library, converting different formats of data into standardized data conforming to the cross-domain data meta-model.

[0022] The construction process of the risk transmission model includes: Global data pool data is obtained, including raw material procurement data, production process data, storage data, transportation data, sales data, consumer feedback data, and supervision data, with each link of the industry chain as a node and the data association between each link as an edge, to construct an industry chain node association graph; the nodes include raw material supplier nodes, production workshop nodes, storage nodes, transportation nodes, sales terminal nodes, consumer nodes, and supervision nodes. Based on the graph neural network algorithm, the industry chain node association graph is trained to mine the potential association relationships between nodes and determine the risk transmission weights of each node; the risk transmission weights are determined based on the business association closeness, data interaction frequency, and historical risk transmission records between nodes. When receiving quality problem feedback, the dynamic coding information of the problem consumer product is extracted to locate the corresponding core node. Based on the trained risk transmission model, the upstream associated nodes of the core node are traced back to determine the risk source node. The downstream associated nodes of the core node are traced forward to comb the full path of risk diffusion, output the risk source node information, risk diffusion path information, and corresponding responsibility subject information of each node, providing a basis for subsequent disposal.

[0023] The intelligent dynamic early warning module includes: Acquire basic risk indicators, dynamic risk indicators, and management risk indicators, and integrate them. Basic risk indicators include inherent attribute indicators such as consumer goods business format, raw material characteristics, and production process complexity. Dynamic risk indicators include variable indicators such as real-time monitoring data, complaint density, public opinion heat value, and sampling failure rate. Management risk indicators include management-related indicators such as enterprise self-inspection and rectification rate, historical violation records, and quality system certification status. The random forest algorithm is used to dynamically calculate the risk value of the fused risk indicators. Based on the calculated dynamic risk value, an early warning level signal is automatically generated, and a differentiated response plan is triggered for different early warning levels.

[0024] The process of dynamically calculating the risk value specifically includes: Define the basic risk indicators, dynamic risk indicators, and management risk indicators, including their definitions, data sources, and quantification values. The basic risk indicators are quantified using a combination of qualitative and quantitative methods, the dynamic risk indicators are quantified using real-time data, and the management risk indicators are quantified using a combination of historical and real-time data. The subjective weights of basic risk indicators, dynamic risk indicators, and management risk indicators are determined using the analytic hierarchy process (AHP). The objective weights of each indicator are calculated based on historical data using the entropy weight method. The subjective weights and objective weights are then weighted and fused together to obtain the comprehensive weights of each indicator. A risk value calculation model is constructed based on the random forest algorithm. The quantitative values ​​of each indicator and their corresponding comprehensive weights are input into the model to calculate the real-time risk value of consumer products. The random forest algorithm predicts the risk value through multiple decision trees and takes the average of the prediction results of multiple decision trees as the final risk value.

[0025] The system automatically generates early warning level signals based on the calculated dynamic risk values ​​and triggers differentiated response plans for different early warning levels, including: A preset risk threshold range is established, and the warning level is determined based on the threshold range in which the risk value falls. Specifically, a risk value ≥ 85 is a red warning, 70 ≤ risk value < 85 is a yellow warning, 50 ≤ risk value < 70 is a blue warning, and a risk value < 50 is a green warning. Corresponding contingency plans are triggered for different warning levels: a red warning activates a cross-departmental joint enforcement mechanism, conducts a comprehensive inspection of the involved company, recalls the same batch of consumer products, and suspends the market sales of the involved consumer products; a yellow warning strengthens spot checks at key points, increases the frequency and sample size of spot checks, and requires companies to conduct self-inspection and submit rectification reports; a blue warning sends risk alerts to companies, requiring them to strengthen control over key aspects and report control status regularly; and a green warning implements annual random inspections and reduces the frequency of routine supervision. By collecting the post-treatment sampling data, complaint data, public opinion data, recalculating the risk value, and judging whether the risk is effectively controlled; If the risk value does not decrease below the corresponding early warning level threshold, adjust the disposal plan and strengthen the disposal measures.

[0026] The process of pre-judging potential risks specifically includes: Collect historical data in the global data pool, including historical risk value data, historical quality problem data, historical raw material quality data, historical production process data, and historical consumption feedback data; clean, deduplicate, and standardize the historical data, and select feature data related to potential risks, focusing on risk feature data; Use time series prediction algorithms to build potential risk prediction logic, divide the processed historical feature data into training data and validation data according to time sequence, optimize the prediction logic parameters with training data, and verify the effectiveness of the prediction logic with validation data; Input real-time collected dynamic data into the optimized time series prediction logic to predict the risk value change trend in the future period, identify potential risk types, including raw material quality fluctuation risk, production process deviation risk, and potential risk; When predicting potential risks, push forward-looking early warning information to regulatory departments and enterprises, and provide special prevention and control suggestions for potential risks.

[0027] Working principle: when using the big data empowered consumer product quality traceability and risk early warning management and control system of the application, according to Fig. 1 and Fig. 2 , the following steps are included: S1: Obtain multi-source data of consumer products to form a global data pool, and after structured processing and standardization, use blockchain encryption distributed storage; S2: Based on industrial internet identification analysis technology, give each consumer product a unique dynamic code, and update the code information in real time in the storage, handling and transportation links; S3: When quality problems occur, use the risk transmission model built by the graph neural network algorithm to locate the risk consumer product life cycle data, quickly lock the risk source and diffusion path, and clarify the responsibility subject; S4: Dynamically calculate the risk value by the random forest algorithm, generate an early warning level signal according to the risk value, and trigger differentiated disposal plans for different early warning levels; S5: Analyze the historical data trend by time series prediction algorithm, predict the risk value change trend, identify potential risk types, and push forward-looking early warning and prevention and control suggestions to regulatory departments and enterprises.

[0028] In one embodiment, after obtaining the multi-source data of consumer products, it further includes: The multi-source data is standardized, and the standardized multi-source data is time-sliced according to data sources and business links to form an initial time series set; A causal relationship discovery algorithm is applied to construct a correlation relationship network among the sequences in the initial time series set, wherein a node represents a sequence, and an edge represents a statistically significant guiding or influencing relationship. Based on the correlation relationship network, a collaborative quality inspection is performed on sequence groups having correlations in the initial time series set, low-quality sequences are identified, and a high-quality sequence dataset is output. Key risk indicators and dynamic risk patterns are mined from the high-quality sequence dataset, and historical risk events are retrieved according to the key risk indicators. The historical risk events are marked as key points on a time axis, and all sequence data segments in a time window before and after the event are extracted as analysis samples. A time series pattern mining algorithm and regression analysis are used to identify a frequently occurring and strongly correlated risk indicator combination change pattern from the analysis samples as a dynamic risk pattern. The fitting coefficients of each single sequence element in the high-quality sequence dataset and the dynamic risk pattern are calculated, and a key sequence element set is selected according to the fitting coefficients. For each element in the key sequence element set, statistical control limits are calculated based on historical performance data, and specification limits are set based on business rules. The sequence values of the key elements are detected in real time, and a graded early warning signal is generated according to the comparison result of the sequence values and the specification limits of the key elements, and an early abnormality prompt is generated according to the touch control limit. Based on the early abnormality prompt and the historical trend of the key elements, a time series prediction algorithm is used to make a comprehensive prediction of potential risks.

[0029] In this embodiment, time slicing means that the standardized data is cut into continuous time sequence segments according to fixed time intervals or according to key events of business links.

[0030] In this embodiment, the causal relationship discovery algorithm uses Granger causality test. Under a given significance level (such as p<0.05), if the historical values of time sequence X can significantly predict the current values of time sequence Y, it is considered that there is a guiding relationship from X to Y, and a directed edge is established in the correlation relationship network.

[0031] In this embodiment, the coordinated quality test means checking whether the dynamic changes of the sequence groups with causal relationship are coordinated in time sequence. The specific method is: calculating the dynamic time warping distance or cross-correlation coefficient of the leading sequence and the guided sequence in the sliding window; if the distance exceeds the threshold or the correlation coefficient is lower than the threshold, the guided sequence is determined as a low-quality sequence, because its change mode does not match the causal expectation.

[0032] In this embodiment, the low-quality sequence means the sequence that fails the test in the coordinated quality test, or refers to the sequence with missing rate exceeding 10%, continuous constant value being too long, or variance being zero. The system labels the low-quality sequence and triggers the data review process. After interpolation, smoothing or rejection processing, the processed sequence and the original sequence that passes the test together constitute a high-quality sequence dataset.

[0033] In this embodiment, the dynamic risk pattern is determined by using a time sequence pattern mining algorithm and regression analysis, including: Feature construction: statistical features (mean, variance, slope) and morphological features (peak, valley) are extracted from each sequence segment.

[0034] Pattern recognition: association rule mining or sequence pattern mining is used to find out the frequent co-occurrence of feature combinations before the occurrence of risk events, which is recorded as a candidate risk pattern.

[0035] Pattern screening: the statistical correlation strength of each candidate risk pattern with the occurrence of risk events is analyzed by using logistic regression or Cox proportional hazards regression model, and the strongly correlated pattern is selected as the final dynamic risk pattern.

[0036] In this embodiment, the fitting coefficient is obtained by calculating the reciprocal of the dynamic time warping distance or the absolute value of the Pearson correlation coefficient between each single sequence element and the dynamic risk pattern sequence. The higher the fitting coefficient, the more similar the morphology of the sequence element to the risk pattern.

[0037] In this embodiment, the statistical control limit is represented as the center line and upper and lower control limits (such as ±3 times the standard deviation) calculated by using the exponential weighted moving average control chart method based on the historical performance data of each element, with the data of the last N periods (such as N=30) rolling.

[0038] In this embodiment, the specification limit means the upper and lower limits of allowable fluctuation set for the element according to the national mandatory standards, industry standards or enterprise internal quality control documents, which is used to determine whether the product / process is qualified.

[0039] In this embodiment, the early abnormality prompt is triggered when the real-time sequence value of any key sequence element exceeds its statistical control limit but is still within the specification limit.

[0040] In this embodiment, the early abnormality prompt signal, the key element historical trend feature and other context features are used as input variables of a gradient boosting tree model or a long short-term memory network model to output the probability of occurrence of a potential risk event in a future period. When the probability exceeds a preset threshold, a potential risk warning is generated.

[0041] The above technical solution has the following beneficial effects: by setting control limits for key indicators, statistical abnormal fluctuations in the production or supply chain process can be captured before quality problems cause substantial non-conformities, early abnormality prompts are issued, and the practicality is improved. The correlation network between indicators is constructed through a cause-effect relationship discovery algorithm. When an early abnormality prompt is issued for a certain indicator, the system can immediately locate other indicators that have a causal relationship with it and perform collaborative quality inspection. This not only verifies the authenticity of the abnormality, but also helps to locate the path through which the abnormality may be transmitted, greatly improving the explainability of the alarm and the efficiency of root cause analysis. The risk perception is transformed from macro-lagging to micro-advance, significantly improving the initiative, accuracy and intelligent level of risk control.

[0042] In one embodiment, after obtaining the multi-source data of consumer goods to form a global data pool, and after structured processing and standardization, before adopting blockchain encryption distributed storage, the following steps are further included: determining the historical production batch and factory defect parameters of the consumer goods, and determining a plurality of potential abnormal quality events of the consumer goods based on the historical production batch and factory defect parameters; obtaining the industry chain node influence propagation strength of each potential abnormal quality event, and determining a risk observation value of each potential abnormal quality event based on the node influence propagation strength; determining a development influence factor of each potential abnormal quality event based on the risk observation value; obtaining abnormal observation data of each potential abnormal quality event from the multi-source data of consumer goods, and obtaining evaluation data of users on abnormal key data; determining the confidence and support of users on the abnormal key data according to the evaluation data; calculating a source parameter uncertainty factor of the multi-source data of consumer goods according to the confidence and support of users on the abnormal key data and the development influence factor of each potential abnormal event: wherein Q represents the source parameter uncertainty factor of the multi-source data of consumer goods, N represents the number of potential abnormal quality events, and i represents the i-th potential abnormal quality event, represents the development influence factor of the i-th potential abnormal quality event, represents the confidence-related weight, which is 0.6, denotes the confidence of the user on the abnormal key data of the i-th potential abnormal quality event, denotes the support-related weight, which is set to 0.4, denotes the support of the user on the abnormal key data of the i-th potential abnormal quality event, ln denotes the natural logarithm, k denotes the baseline fluctuation factor of the consumer goods industry chain, and S denotes the steady-state linkage index of the consumer goods industry chain, denotes the development controllable factor of the consumer quality event. The source parameter uncertainty factor of the consumer multi-source data is compared with a preset threshold. If the uncertainty factor is higher than the preset threshold, an enhanced data collection instruction is generated, and specific data sources that need to be supplemented are determined based on the enhanced data collection instruction. The related parameters of the specific data sources are inserted into the multi-source data.

[0043] In this embodiment, the potential abnormal quality event refers to an event that may pose a risk to the quality of consumer goods, which is identified by any one or combination of the following methods: (a) applying statistical process control rules to historical production batch data to identify batches that exceed the control limit; (b) discovering strong association combinations between specific raw material batches and specific factory defects based on association rule mining; (c) using isolated forest or a type of support vector machine algorithm for unsupervised learning of factory detection parameters to identify abnormal samples corresponding to events.

[0044] In this embodiment, the industry chain node influence propagation intensity is represented as the occurrence node of the potential abnormal quality event as the source, based on the industry chain node association graph, using a random walk algorithm or a network infection model to simulate the probability of the event influence reaching other nodes in the graph within a preset number of steps. The probability value is the propagation intensity of the event on the corresponding node.

[0045] In this embodiment, the risk observation value is obtained by multiplying the propagation intensity and the inherent risk vulnerability coefficient of each node (set according to the historical quality problem frequency of each node).

[0046] In this embodiment, the development influence factor is used to represent the relative importance of the i-th potential abnormal quality event. The calculation method is to weight average or take the maximum value of the risk observation value of the event on all nodes, and then normalize the value so that the sum of all events is 1.

[0047] In this embodiment, the abnormal key data refers to the monitoring or detection data directly related to each potential abnormal quality event extracted from the global data pool, such as the content of a specific component of the N-th batch of raw materials, and the value of a specific process parameter in the production process.

[0048] In this embodiment, the evaluation data is represented as the evaluation conclusion issued by the domain expert system or the quality audit rule engine for the abnormal key data.

[0049] In this embodiment, the confidence level is represented as the probability value output by the classification model (such as logistic regression) trained according to historical data for the judgment that the current abnormal key data belongs to a real abnormality.

[0050] In this embodiment, the support degree is represented as the proportion of other collaborative indicator data that is also abnormal in association with each abnormal key data.

[0051] In this embodiment, the baseline fluctuation factor is represented as the coefficient of variation of the monthly average factory pass rate of various consumer goods in the past year, which is used to measure the baseline fluctuation level of the industry quality, and is directly calculated from the historical statistical database.

[0052] In this embodiment, the steady-state connection index is represented as the geometric mean of the data reporting punctuality rate between all key path nodes calculated based on the industry chain node correlation graph, which is used to measure the stability of the data flow in the industry chain.

[0053] In this embodiment, the development controllable factor is represented as a coefficient calculated based on the quality system internal audit score and the rectification closure rate of the enterprise in the last quarter, which is used to measure the ability of the enterprise to control the development of quality events.

[0054] The beneficial effects of the above technical solutions are: by calculating the source parameter uncertainty factor, the uncertainty of different consumer goods and different batches at the data level is quantified. Based on the source parameter uncertainty factor, data collection items can be dynamically planned (for example, more detection items are added for high-risk batches, and the sensor sampling frequency is increased). This realizes the adaptive allocation of data collection resources, meets the needs of risk control with the least data cost, realizes the optimal configuration of data resources under the risk control target, and guarantees the efficiency and effectiveness of the overall quality traceability and early warning system from the beginning of the data supply chain.

[0055] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0056] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and changes can be made by those skilled in the art without departing from the spirit and principles of the present application.

Claims

1. A consumer product quality traceability and risk early warning management system empowered by big data, characterized in that: include: The multi-source heterogeneous data processing module is used to acquire multi-source consumer product data to form a global data pool. After structured processing and standardization, the data is stored in a blockchain-encrypted distributed manner. The full-chain dynamic traceability module is used to dynamically encode consumer products using industrial internet identifiers, build a risk transmission model, and use dynamic coding and risk transmission model to trace the source of consumer products with quality problems; The intelligent dynamic early warning module integrates three types of risk indicators: basic, dynamic, and management. It uses the random forest algorithm to dynamically calculate risk values, generate early warning levels, and trigger differentiated contingency plans. The risk prediction module is used to analyze historical data trends through time-series prediction algorithms to predict potential risks in advance.

2. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 1, characterized in that, The multi-source heterogeneous data processing module specifically includes: Acquire industrial internet identifier resolution data, IoT real-time sensing data, multi-channel feedback data from consumers, regulatory spot check data, and industry standard data to form a comprehensive data pool; The acquired structured data undergoes format validation and consistency correction. For unstructured text data, natural language processing algorithms are used for word segmentation, entity recognition, keyword extraction, and semantic normalization to convert it into structured data. By using dynamic data mapping technology, a field mapping relationship library for different data sources is established, and the corresponding mapping rules are automatically matched according to the data source type to perform unified and standardized processing on data of different formats. After processing, standardized data is distributed and encrypted using blockchain technology.

3. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 2, characterized in that, The standardization process for data in different formats specifically includes: Construct a cross-domain data element model to define the core data elements, data types, data lengths, and value ranges of data from different domains; Establish a dynamic mapping rule base and formulate mapping rules between data elements for heterogeneous data formats of different enterprises and departments; A dynamic mapping engine is used to automatically map and transform heterogeneous data according to the rules in the mapping rule base, converting data of different formats into standardized data that conforms to the cross-domain data meta-model.

4. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 1, characterized in that, The construction process of the risk transmission model specifically includes: Acquire full-chain data from the global data pool, and construct a node relationship graph of the industrial chain with each link of the industrial chain as a node and the data relationship between each link as an edge; The graph neural network algorithm is used to train the node association graph of the industrial chain, explore the potential association between nodes, and determine the risk transmission weight of each node. When a quality issue is reported, the dynamic coding information of the problematic consumer product is extracted to locate the corresponding core node. Based on the trained risk transmission model, the upstream related nodes of the core node are traced back to determine the risk source node; The system traces downstream related nodes of the core node in a forward direction, sorts out the entire path of risk spread, and outputs information on the risk source node, risk spread path information, and the responsible entity information for each node.

5. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 1, characterized in that, The intelligent dynamic early warning module specifically includes: Obtain basic risk indicators, dynamic risk indicators, and management risk indicators, and integrate these indicators. The random forest algorithm is used to dynamically calculate the risk value of the fused risk indicators. Based on the calculated dynamic risk value, an early warning level signal is automatically generated, and a differentiated response plan is triggered for different early warning levels.

6. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 5, characterized in that, The process of dynamically calculating the risk value specifically includes: Define the basic risk indicators, dynamic risk indicators, and management risk indicators, including their definitions, data sources, and quantification values; The subjective weights of basic risk indicators, dynamic risk indicators, and management risk indicators are determined using the analytic hierarchy process (AHP). The objective weights of each indicator are calculated based on historical data using the entropy weight method. The subjective weights and objective weights are then weighted and fused together to obtain the comprehensive weights of each indicator. A risk value calculation model is constructed based on the random forest algorithm. The quantitative values ​​of each indicator and their corresponding comprehensive weights are input into the model to calculate the real-time risk value of consumer products.

7. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 5, characterized in that, The automatic generation of early warning level signals based on the calculated dynamic risk value, and the triggering of differentiated response plans for different early warning levels, specifically includes: Preset risk value threshold range, determine the warning level based on the threshold range where the risk value is located, and trigger the corresponding response plan for different warning levels; By collecting and processing sampling data, complaint data, and public opinion data, the risk value is recalculated to determine whether the risk has been effectively controlled. If the risk value does not drop below the corresponding warning level threshold, the emergency response plan will be adjusted and the response measures will be strengthened.

8. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 1, characterized in that, The process of making advance predictions of potential risks specifically includes: Collect historical data from the global data pool, clean, deduplicate, and standardize the historical data, and filter out feature data related to potential risks; A potential risk prediction logic is constructed using a time series prediction algorithm. The processed historical feature data is divided into training data and validation data according to the time series. The prediction logic parameters are optimized using the training data, and the effectiveness of the prediction logic is verified using the validation data. The real-time collected dynamic data is input into the optimized time series prediction logic to predict the trend of risk value changes over a period of time and identify potential risk types. When potential risks are predicted, forward-looking early warning information is sent to regulatory authorities and enterprises, and specific prevention and control recommendations are provided for potential risks.

9. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 1, characterized in that, After obtaining multi-source data on consumer products, the following is also included: The multi-source data is standardized, and the standardized multi-source data is then sliced ​​into time series according to the data source and business process to form an initial time series set. A causal relationship discovery algorithm is applied to construct a correlation network among the sequences in the initial time series set, where nodes represent sequences and edges represent statistically significant guiding or influencing relationships. Based on the association network, a collaborative quality check is performed on the associated sequence groups in the initial time series set to identify low-quality sequences and output a high-quality sequence dataset. Mining key risk indicators and dynamic risk patterns from high-quality sequence datasets, and retrieving historical risk events based on key risk indicators; Historical risk events are marked as key points on the timeline, and all sequence data fragments within the time window before and after the event are extracted as analysis samples. Using time-series pattern mining algorithms and regression analysis, we identify frequently occurring risk indicator combination change patterns that are strongly correlated with risk events from the analyzed samples, which serve as dynamic risk patterns. Calculate the fitting coefficient between each individual sequence element in the high-quality sequence dataset and the dynamic risk pattern, and select the set of key sequence elements based on the fitting coefficient. For each element in the set of key sequence elements, the statistical control limits are calculated on a rolling basis based on its historical performance data, and the specification limits are set in combination with business rules. Real-time detection of the sequence value of key elements; generation of graded early warning signals based on the comparison results between the sequence value and the specification limit of the key element; and generation of early abnormality prompts based on the situation of reaching the control limit. Based on early anomaly alerts and historical trends of key elements, a time-series prediction algorithm is used to comprehensively predict potential risks.

10. The consumer product quality traceability and risk early warning control system empowered by big data according to claim 1, characterized in that, Before acquiring multi-source consumer product data to form a comprehensive data pool, and after structured processing and standardization, and before employing blockchain-based encrypted distributed storage, the process also includes: Determine the historical production batches and outgoing defect parameters of consumer products, and identify multiple potential abnormal quality events of consumer products based on the historical production batches and outgoing defect parameters; Obtain the influence propagation intensity of each potential abnormal quality event at each node in the industry chain, and determine the risk observation value of each potential abnormal quality event based on the influence propagation intensity of each node; The developmental impact factors for each potential abnormal quality event are determined based on risk observations. Obtain abnormal observation data for each potential abnormal quality event from multi-source consumer product data, and obtain user evaluation data for key abnormal data; Determine the user's confidence and support for the abnormal key data based on the evaluation data; The uncertainty factor of the source parameters of multi-source consumer product data is calculated based on users' confidence and support for the key abnormal data and the development impact factors of each potential abnormal event: Where Q represents the uncertainty factor of the source parameter of the multi-source data for consumer goods, N represents the number of potential abnormal quality events, and i represents the i-th potential abnormal quality event. Let be the development impact factor of the i-th potential abnormal quality event. This represents the confidence level related weight, with a value of 0.

6. This represents the user's confidence level regarding the critical data related to potential abnormal quality events (i items). This represents the support-related weight, with a value of 0.

4. Let represent the user's support for the key data of the i-th potentially abnormal quality event, ln represent the natural logarithm, k represent the baseline fluctuation factor of the consumer goods supply chain, and S represent the steady-state connectivity index of the consumer goods supply chain. This refers to the controllable factors in the development of consumer product quality incidents; The uncertainty factor of the source parameter of the multi-source data of consumer products is compared with a preset threshold. If the uncertainty factor is higher than the preset threshold, an enhanced data collection instruction is generated, and the specific data source that needs to be collected is determined based on the enhanced data collection instruction. Collect relevant parameters from a specific data source and insert them into multi-source data.

Images