Quality performance evaluation method and device, electronic equipment, medium and program product

Abstract

The invention provides a quality performance evaluation method and device, electronic equipment, a medium and a program product, and relates to the technical field of computers. The quality performance evaluation method comprises the steps that fusion processing is carried out on different types of quality related data collected from a multi-source heterogeneous data system to obtain multi-dimensional quality fusion data, and the quality related data represents data related to quality management; quantitative characteristics of the quality performance are obtained; an evaluation model which is adaptive to the quantitative characteristics and has multi-dimensional quality performance indexes is constructed based on the domain requirements and the business scenes, and the evaluation model determines the weights of the performance indexes based on an analytic hierarchy process; and inputting the quantitative characteristics of the quality performance into the evaluation model to obtain a quality performance evaluation result. Through the technical scheme disclosed by the invention, dominant indexes and potential association rules of all links of quality management are comprehensively captured, so that the objectivity, accuracy and systematicness of evaluation results are improved.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to a quality performance evaluation method, a quality performance evaluation device, an electronic device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] In modern enterprise management, quality performance evaluation refers to a management activity that comprehensively evaluates the quality performance of enterprise management processes and other aspects based on established standards and indicators. It is an important means for enterprises to improve product quality, optimize management processes, and enhance market competitiveness. Existing quality performance evaluation methods are mostly based on manual operation, requiring staff to manually collect quality-related data from various business scenarios, and then integrate the data through manual sorting and summarization before completing the performance evaluation analysis based on established standards.

[0003] However, the cumbersome process of manually collecting and organizing data not only leads to low efficiency in the evaluation process, but also makes it easy for data errors to occur due to human error, affecting the accuracy of the evaluation. In addition, the independent nature of the multi-source heterogeneous systems within the enterprise forms data silos, making it difficult to achieve cross-system data sharing and collaborative analysis, so that the quality performance evaluation results cannot comprehensively and accurately reflect the enterprise's quality management level.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this disclosure is to provide a quality performance evaluation method, a quality performance evaluation device, and a computer program product, which at least to some extent overcome the problem that the quality performance evaluation results in related technologies cannot fully and accurately reflect the quality management level of enterprises.

[0006] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0007] According to one aspect of this disclosure, a quality performance evaluation method is provided, comprising: fusing different types of quality-related data collected from a multi-source heterogeneous data system to obtain multi-dimensional quality fusion data, wherein the quality-related data represents data related to quality management; performing association rule mining and cluster analysis on the multi-dimensional quality fusion data to obtain quantitative characteristics of quality performance; constructing an evaluation model with multi-dimensional quality performance indicators adapted to the quantitative characteristics based on domain requirements and business scenarios, wherein the evaluation model determines the weights of the performance indicators based on the analytic hierarchy process; and inputting the quantitative characteristics of quality performance into the evaluation model to obtain a quality performance evaluation result.

[0008] In one embodiment of this disclosure, multi-dimensional quality fusion data is obtained by fusing different types of quality-related data collected from a multi-source heterogeneous data system. This includes: cleaning the collected different types of quality-related data to obtain cleaned data; the multi-source heterogeneous data system includes a quality management system, a production management system, a supply chain management system, and an after-sales management system; performing a unified format conversion operation on the cleaned data to obtain standardized structured fusion data; and performing content association fusion on the standardized structured fusion data to obtain the multi-dimensional quality fusion data. The dimensions of the multi-dimensional quality fusion data include process execution dimension, product quality dimension, supply chain collaboration dimension, and customer feedback dimension.

[0009] In one embodiment of this disclosure, the standardized structure fusion data is subjected to content association fusion to obtain the multidimensional quality fusion data, including: extracting product batch number, production order number, and material code from the standardized structure fusion data as cross-data source association identifiers; using the product batch number as a first association key to associate the standardized structure fusion data from the production management system and the quality management system, so that the process parameters and equipment operation data of the same product batch correspond to the inspection pass rate and defect type data, to obtain a first set of associated data; using the production order number as a second association key to associate the first set of associated data with the standardized structure fusion data from the supply chain management system, so that the raw material pass rate and supplier information corresponding to the same production order are associated with the first set of associated data, to obtain a second set of associated data; using the material code as a third association key to associate the second set of associated data with the standardized structure fusion data from the after-sales management system, so that the customer feedback data corresponding to the same material is integrated with the second set of associated data, to obtain the multidimensional quality fusion data.

[0010] In one embodiment of this disclosure, association rule mining and cluster analysis are performed on the multidimensional quality fusion data to obtain quantitative features of quality performance. This includes: discretizing the continuous indicator data in the multidimensional quality fusion data to map it to corresponding numerical intervals, obtaining discretized data; performing feature encoding on the categorical data in the multidimensional quality fusion data to obtain categorical feature data suitable for mining operations; constructing a mining analysis dataset based on the discretized data, the categorical feature data, and the core fields in the multidimensional quality fusion data; performing association rule mining and cluster analysis operations on the mining analysis dataset to obtain operation results including a set of association rules and quality group clustering results; and quantifying the operation results to obtain the quantitative features of quality performance, wherein the quantitative features include process execution performance features, product quality performance features, supply chain collaboration performance features, and customer feedback performance features.

[0011] In one embodiment of this disclosure, association rule mining and clustering analysis are performed on the mining and analysis dataset to obtain the operation result, including: mining the association relationships between data of different dimensions in the mining and analysis dataset based on the association rule mining algorithm to obtain the corresponding association rule set; selecting corresponding quality-related indicators based on different dimensions; using the quality-related indicators as cluster centers, performing clustering processing on the mining and analysis dataset based on the clustering algorithm to obtain quality groups with different risks; using the mining and analysis dataset as input features, and taking the future batch quality performance level and the existence of specific defects as prediction targets, outputting the corresponding prediction result based on the decision tree algorithm; and obtaining the operation result based on the association rule set, the quality groups with different risks, and the prediction result.

[0012] In one embodiment of this disclosure, the operation result is quantified to obtain the quantitative characteristics of the quality performance, including: filtering the set of association rules to obtain strong association rules with confidence greater than a first threshold and lift greater than a second threshold; extracting causal relationships between different dimensional indicators based on the strong association rules and converting them into first numerical features; extracting features from the quality groups with different risks and calculating corresponding second numerical features based on the dimensional distribution characteristics of samples within the cluster; mapping the future batch quality performance level to a numerical level quantification value and converting the existence of a specific defect into a defect risk probability to obtain corresponding third numerical features; and mapping the first numerical features, the second numerical features, and the third numerical features to the corresponding quantitative characteristics of the quality performance based on the process execution dimension, the product quality dimension, the supply chain collaboration dimension, and the customer feedback dimension.

[0013] In one embodiment of this disclosure, an evaluation model with multidimensional quality performance indicators adapted to the quantitative characteristics is constructed based on domain requirements and business scenarios. The evaluation model determines the weights of the performance indicators based on the analytic hierarchy process (AHP), including: determining the multidimensional quality performance indicators based on the domain requirements and business scenarios; the multidimensional quality performance indicators include product quality-related indicators under the unit performance dimension, functional performance-related indicators under the grassroots performance dimension, and PDCA management-related indicators under the process performance dimension; and establishing a hierarchical structure model based on the AHP, with quality performance evaluation as the target layer, and the unit performance dimension, the grassroots performance dimension, and the process performance dimension as a... The system employs a first-level indicator layer, where product quality-related indicators, functional performance-related indicators, and PDCA management-related indicators are designated as second-level indicator layers. A relative importance judgment is performed on the first-level and second-level indicator layers to obtain the judgment results. A corresponding judgment matrix is ​​constructed based on these results, where the judgment matrix is ​​an N-order positive reciprocal matrix, where N is the number of indicators at the corresponding level, and each element in the matrix represents the relative importance between two corresponding indicators. Based on the consistency check and eigenvalue calculation of the judgment matrix, the weights of the performance indicators are determined. Finally, the evaluation model is constructed by fusing the multidimensional quality performance indicators with their corresponding weights using a fuzzy comprehensive evaluation framework.

[0014] In one embodiment of this disclosure, the weights of the performance indicators are determined based on the consistency check and eigenvalue calculation operations of the judgment matrix, including: calculating the largest eigenvalue and the corresponding eigenvector of the judgment matrix; calculating a consistency index based on the largest eigenvalue and the order of the judgment matrix; selecting a corresponding random consistency index based on the order of the judgment matrix; calculating the ratio between the consistency index and the random consistency index as the consistency ratio; if the consistency ratio is less than a preset standard, determining that the judgment matrix passes the consistency check; if the consistency check passes, normalizing the eigenvector corresponding to the largest eigenvalue to obtain the weights of the performance indicators at the corresponding level.

[0015] In one embodiment of this disclosure, the method further includes: configuring a corresponding visualization format for the quality performance evaluation results based on at least one of data dimension attributes, numerical relationship types, and analysis target scenarios; and / or determining the display content and display method corresponding to the quality performance evaluation results based on user roles and / or permissions.

[0016] In one embodiment of this disclosure, a corresponding visualization format is configured for the quality performance evaluation results based on data dimension attributes, numerical relationship types, and analysis target scenarios. This includes: if the data dimension attribute is a key quality performance indicator, the corresponding display format includes a dashboard; if the data dimension attribute is multi-category / multi-time period data, the corresponding display format includes a bar chart or line chart; if the numerical relationship type is percentage-based data, the corresponding display format includes a pie chart or donut chart; if the analysis target scenario includes geographically related data, the corresponding display format includes a map; if the data dimension attribute is detailed data, the corresponding display format includes a table or report; and if the data dimension attribute is multi-level data, the corresponding display format includes a tree diagram or sunburst chart.

[0017] According to another aspect of this disclosure, a quality performance evaluation device is provided, comprising: a fusion processing module for fusing different types of quality-related data collected from a multi-source heterogeneous data system to obtain multi-dimensional quality fusion data, wherein the quality-related data represents data related to quality management; a mining and analysis module for performing association rule mining and cluster analysis on the multi-dimensional quality fusion data to obtain quantitative characteristics of quality performance; a construction module for constructing an evaluation model with multi-dimensional quality performance indicators adapted to the quantitative characteristics based on domain requirements and business scenarios, wherein the evaluation model determines the weights of the performance indicators based on the analytic hierarchy process; and an evaluation module for inputting the quantitative characteristics of quality performance into the evaluation model to obtain quality performance evaluation results.

[0018] According to another aspect of this disclosure, an electronic device is provided, which is a computer device or a VR device, comprising: a processor; and a memory for storing executable instructions of the processor; the processor is configured to perform the quality performance evaluation method described in the first aspect by executing the executable instructions.

[0019] According to another aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the computer program product described above.

[0020] According to another aspect of this disclosure, a computer program product is provided, on which a computer program is stored, which, when executed by a processor, implements the computer program product described above.

[0021] The quality performance evaluation scheme provided in the embodiments of this disclosure integrates quality-related data from multiple heterogeneous systems throughout the entire process. Through fusion processing, it forms a multi-dimensional data system covering process execution, product quality, supply chain collaboration, and customer feedback. It extracts quantitative characteristics of quality performance using association rule mining and cluster analysis, constructs a multi-dimensional performance indicator system based on domain needs, and determines the indicator weights using the analytic hierarchy process (AHP) to obtain the evaluation model. This achieves an organic combination of qualitative and quantitative evaluation, preventing the subjective limitations of traditional evaluations. It comprehensively captures the explicit indicators and potential correlation patterns of each link in quality management, while improving the objectivity, accuracy, and systematic nature of the evaluation results. This provides reliable decision support for enterprises to identify quality management shortcomings, optimize production processes, improve supply chain collaboration efficiency, and enhance customer satisfaction.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0024] Figure 1 A schematic flowchart of a quality performance evaluation scheme according to an embodiment of this disclosure is shown; Figure 2 A schematic flowchart of another quality performance evaluation method in an embodiment of this disclosure is shown; Figure 3 A schematic flowchart of another quality performance evaluation method in an embodiment of this disclosure is shown; Figure 4 This diagram shows a structural diagram of a quality performance evaluation device according to an embodiment of the present disclosure; Figure 5 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0025] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0026] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0027] Traditional enterprise quality performance evaluation faces numerous challenges: First, data collection is difficult, with data from different systems isolated, hindering effective sharing and collaborative analysis; second, the evaluation process is inefficient, relying on manual data collection and processing, which is error-prone and time-consuming; third, evaluation results are incomplete and fail to accurately reflect the level of quality management; and fourth, there is a lack of effective visualization tools, preventing decision-makers from intuitively understanding performance and hindering timely decision-making and improvement. To address any of these problems, this paper presents a data processing and visualization method for quality performance evaluation in a QMS system based on big data fusion, aiming to achieve efficient, accurate, and intuitive quality performance evaluation and help enterprises improve their quality management effectiveness.

[0028] The steps of the computer program product in this exemplary embodiment will now be described in more detail with reference to the accompanying drawings and embodiments.

[0029] like Figure 1 As shown, a quality performance evaluation method according to an embodiment of this disclosure, applied to a computer device, includes: Step S102: The different types of quality-related data collected from the multi-source heterogeneous data system are fused to obtain multi-dimensional quality fusion data. The quality-related data represents data related to quality management.

[0030] The data related to quality management includes, but is not limited to, the execution status of the entire quality management process, product quality level, customer feedback, and management effectiveness.

[0031] In some embodiments, a stable data connection channel can be established with multi-source heterogeneous data systems such as QMS systems, production management systems, and supply chain management systems to ensure that data can be transmitted in real time and accurately. For example, standard interface protocols and data exchange formats, such as WebService, XML, and JSON, can be used to achieve seamless integration with various systems to collect quality-related data such as process data, product data, inspection data, and customer feedback data.

[0032] The dimensions of multidimensional quality fusion data include process execution, product quality, supply chain collaboration, and customer feedback. This allows for the comprehensive collection of process data related to quality performance, including quality indicators, nonconformity information, process quality issues, quality management behaviors, quality improvement, stakeholder feedback, and quality assessments.

[0033] In the multidimensional process execution dimension, the execution status of the production process can be characterized by data such as process parameter deviation values, process execution accuracy, and equipment operating parameters.

[0034] In terms of product quality, data such as inspection pass rate, defect type and quantity are used to characterize the physical quality level of the final product.

[0035] In terms of supply chain collaboration, data such as raw material qualification rate, supplier on-time delivery rate, and supplier name are used to characterize the effectiveness of quality collaboration in the procurement process.

[0036] In terms of customer feedback, data such as the number of customer complaints, types of malfunctions, and repair records are used to characterize the market's feedback on product quality.

[0037] Step S104: Perform association rule mining and cluster analysis on the multidimensional quality fusion data to obtain the quantitative characteristics of quality performance.

[0038] Quality performance refers to the comprehensive performance of an enterprise in the entire quality management process, focusing on product quality, process execution, customer feedback, and management effectiveness. It is reflected through quantitative indicators (such as pass rate and defect rate) and potential patterns (such as problem patterns and correlations) and serves as the basis for measuring the level of quality management.

[0039] The quantitative characteristics of quality performance refer to the core quality-related attributes extracted from the fused data through association rule mining, cluster analysis, and classification algorithms, which can be represented by specific values, ratios, or quantitative labels.

[0040] Step S106: Based on domain requirements and business scenarios, construct an evaluation model with multi-dimensional quality performance indicators that is adapted to quantitative characteristics. The evaluation model determines the weights of the performance indicators based on the analytic hierarchy process.

[0041] In some embodiments, the domain requirements can be the ISO22163 standard, and the business scenario can be the enterprise standard. A multi-dimensional quality performance indicator evaluation model is constructed, which includes unit performance, grassroots performance, and process performance dimensions. The evaluation model determines the weight of each indicator through the analytic hierarchy process and uses the fuzzy comprehensive evaluation algorithm to achieve comprehensive calculation.

[0042] Step S108: Input the quantitative characteristics of quality performance into the evaluation model to obtain the quality performance evaluation results.

[0043] The evaluation results can be presented in the form of a comprehensive score, performance levels of each dimension, and membership degree of the corresponding comment set, which can intuitively quantify the overall performance of the entire quality management process and the degree of excellence or inferiority of each dimension.

[0044] In this embodiment, by integrating quality-related data from multiple heterogeneous systems throughout the entire process, a multi-dimensional data system covering process execution, product quality, supply chain collaboration, and customer feedback is formed through fusion processing. Quantitative features of quality performance are extracted using association rule mining and cluster analysis. A multi-dimensional performance indicator system is constructed based on domain requirements, and the evaluation model is obtained by determining the indicator weights through the analytic hierarchy process. This achieves an organic combination of qualitative and quantitative evaluation, preventing the subjective limitations of traditional evaluations. It comprehensively captures the explicit indicators and potential correlation patterns of each link in quality management, while improving the objectivity, accuracy, and systematic nature of the evaluation results. This provides reliable decision support for enterprises to identify quality management shortcomings, optimize production processes, improve supply chain collaboration efficiency, and enhance customer satisfaction.

[0045] In one embodiment of this disclosure, multidimensional quality fusion data is obtained by fusing different types of quality-related data collected from a multi-source heterogeneous data system, including: The collected quality-related data of different types are cleaned to obtain cleaned data. The multi-source heterogeneous data system includes the quality management system, production management system, supply chain management system, and after-sales management system. The cleaned data is then converted into a unified format to obtain standardized structured fused data. The standardized structured fused data is then fused with content association to obtain multi-dimensional quality fused data.

[0046] In some embodiments, the cleaning operation specifically includes removing duplicate records, correcting data with format errors, filling in missing values ​​in key fields (e.g., using the average of similar data or business rules to estimate), and filtering out abnormal data that exceeds a reasonable threshold (e.g., records where process parameters deviate from the standard range by more than 3 times). During the unified format conversion, heterogeneous data from various systems (e.g., tabular data from the quality management system, log data from the production management system, and text feedback data from the after-sales management system) are uniformly converted into a structured data table format according to preset data specifications, achieving global consistency in field names, data types, units of measurement, and encoding formats. In the content association and integration stage, key fields such as product unique identifiers, order numbers, and production batches are used as association links to establish semantic mapping relationships between cross-system data, integrating them into a complete data set including four dimensions: process execution, product quality, supply chain collaboration, and customer feedback. At the same time, semantic conflicts between multi-source data are handled through conflict resolution rules.

[0047] In this embodiment, by cleaning, standardizing, and associating and fusing the original quality-related data, noise and errors in the original data can be eliminated, significantly improving data quality. Unified format conversion breaks down data barriers between systems, while content association and fusion based on core key fields enables collaboration of quality data across systems. It integrates scattered single-point data into multi-dimensional fused data covering the entire quality management process, which not only greatly improves the integrity and relevance of the data, but also reveals the quality correlation patterns hidden in isolated data.

[0048] In one embodiment of this disclosure, content association fusion is performed on standardized structure fusion data to obtain multidimensional quality fusion data, including: Product batch number, production order number, and material code are extracted from standardized structured fusion data as cross-data source association identifiers.

[0049] Using the product batch number as the first association key, standardized structured data from the production management system and quality management system are integrated to match the process parameters and equipment operation data of the same product batch with the inspection pass rate and defect type data, thus obtaining the first set of associated data.

[0050] In some embodiments, the product batch number is used as the primary association key to match the production process data in the production management system with the product inspection data in the quality management system, so that the process parameters, equipment operation data and corresponding inspection pass rate and defect type data of the same product batch correspond one-to-one.

[0051] Using the production order number as the second association key, the first set of associated data is linked to the standardized structured fusion data from the supply chain management system. This allows the raw material qualification rate and supplier information corresponding to the same production order to be associated with the first set of associated data, thus obtaining the second set of associated data. In some embodiments, the production order number is used as the association key to link the matched production-inspection integrated data with the procurement data in the supply chain management system, so that the raw material qualification rate, supplier information and production-inspection data corresponding to the same production order are associated.

[0052] By using the material code as the third association key, the second set of associated data is linked with the standardized structured fusion data from the after-sales management system, so that the customer feedback data corresponding to the same material is integrated with the second set of associated data to obtain multi-dimensional quality fusion data.

[0053] In some embodiments, the material code is used as the association key to link the integrated production, inspection and procurement data with the customer feedback data in the after-sales management system, so that the number of customer complaints, fault types, and repair records corresponding to the same material are connected with the preceding data.

[0054] Perform record-by-record verification on the full data after association, delete isolated data without matching product batch number, production order number or material code, and complete the missing core field information in the associated data.

[0055] The verified associated data is integrated according to the three-dimensional identification of product batch, production order and material code to form a unified data record that includes multi-dimensional information such as production process, product quality, procurement and supply and customer feedback, namely the multi-dimensional quality fusion data.

[0056] In this embodiment, a progressive association strategy using three-dimensional association identifiers of product batch number, production order number, and material code is adopted to aggregate data flows between production management, quality management, supply chain management, and after-sales management systems. This enables the connection and integration of previously scattered production process data, product inspection data, procurement and supply data, and customer feedback data. The resulting multi-dimensional quality fusion data achieves a closed-loop data system covering the entire process from raw material procurement, production execution, quality inspection to after-sales feedback. This completely breaks down data silos in traditional quality management, improves data integrity, consistency, and relevance, and provides coherent and reliable data support for subsequent investigation of the root causes of quality problems and tracing the quality influencing factors throughout the entire chain.

[0057] In one embodiment of this disclosure, association rule mining and cluster analysis are performed on multidimensional quality fusion data to obtain quantitative characteristics of quality performance, including: The continuous index data in the multidimensional quality fusion data are discretized to map to the corresponding numerical range, resulting in discretized data.

[0058] In some embodiments, continuous indicators are characterized by their ability to take any value within a range, allowing for infinite subdivision. For example, temperature can be accurate to 0.1℃, and the pass rate to 0.1%. In practice, the actual value range of each continuous indicator is first determined. For instance, the measured operating temperature of equipment may range from a minimum of 25.3℃ to a maximum of 84.7℃, and the pass rate of raw materials may range from a minimum of 92.5% to a maximum of 99.8%. After sorting all data in ascending order, the average of adjacent data is calculated as a temporary dividing point. Then, the CAIM algorithm is used to select the optimal dividing point, ultimately resulting in 3 to 5 reasonable intervals. This prevents data clutter from too many intervals and loss of key information from too few. For example, the pass rate of raw materials can be divided into 98.0%-100% excellent, 95.0%-98.0% good, and below 95.0% unacceptable. This allows continuous data to be adapted to subsequent algorithms such as association rule mining and cluster analysis, which are easier to process discrete interval data.

[0059] Feature encoding is performed on the categorical data in the multidimensional quality fusion data to obtain categorical feature data suitable for mining operations.

[0060] In some embodiments, for non-quantifiable, categorical data such as defect type (appearance defect, performance defect, assembly defect, etc.), supplier name (manufacturer A, manufacturer B, manufacturer C), and fault type (hardware fault and software fault), one-hot encoding can be used. A binary field is created for each category. If the data belongs to the category, it is marked as 1; otherwise, it is marked as 0. For example, the defect type "appearance defect" is coded as 1 0 0, and the supplier "manufacturer A" is coded as 1 00. For binary classification data such as "qualified" or "not qualified", 0-1 encoding is used directly, where 1 indicates qualification and 0 indicates non-qualification, so as to identify the classification information normally.

[0061] A mining and analysis dataset is constructed based on the core fields of discretized data, categorical feature data, and multidimensional quality fusion data.

[0062] In some embodiments, core fields such as product batch number, production order number, and material code can be extracted from multidimensional quality fusion data first to trace data relationships and ensure that each record can be located to a specific product and batch. Then, the discretized continuous data, such as the numerical range label corresponding to the good grade after interval coding, and the classification feature data after one-hot coding, such as the defect type code, are integrated according to the principle of one-to-one correspondence of the same data record. For example, the structure of a certain data is: product batch number is B20240510, raw material qualification rate is good, corresponding to 95.0%-98.0%, defect type is code 1 0 0, supplier is code 1 0 0, and customer complaint count is 2. Finally, invalid records without core fields or missing key features after integration can be deleted, and missing information of core fields can be filled in, such as using the average of the same batch of data to fill in missing process parameters, forming a unified and complete data mining and analysis dataset.

[0063] Based on the dataset, association rule mining and cluster analysis operations are performed to obtain the results, including the set of association rules and the clustering results of quality groups.

[0064] The operational results are quantified to obtain the quantitative characteristics of quality performance, which include process execution performance characteristics, product quality performance characteristics, supply chain collaboration performance characteristics, and customer feedback performance characteristics.

[0065] In some embodiments, association rule quantification refers to calculating the association strength value for rules such as those that cause appearance defects due to excessive process parameters, based on support (frequency of rule occurrence) × confidence (rule accuracy), with a range of 0-1. For example, a rule with 10% support and 90% confidence has an association strength of 0.09, and only rules with a strength ≥ 0.05 are retained as valid features.

[0066] For clustering results such as high pass rate group or high complaint group, calculate the mean value of the index for each group (e.g., the mean deviation of process parameters for high pass rate group is 0.3℃, and the mean pass rate is 98.5%), and then label the probability (range 0-1) of each data point belonging to a certain cluster group. For example, if 92% of a batch of data belongs to the high pass rate group, the probability value is 0.92.

[0067] The aforementioned association strength values ​​and clustering probability, along with original quantitative indicators such as pass rate and number of complaints, are categorized into four dimensions: process execution, product quality, supply chain collaboration, and customer feedback. These are then standardized into performance characteristic values ​​ranging from 0 to 100. For example, an association strength of 0.09 corresponds to 9 points, and a clustering probability of 0.92 corresponds to 92 points, forming structured quantitative characteristics of quality performance that can be directly input into the evaluation model.

[0068] In this embodiment, continuous data is discretized (clarifying the essence of continuous indicators and converting them into interval data) and categorical data is encoded (converting non-numerical categories into numerical values ​​that can be recognized by the algorithm). This enables the obtained mining and analysis dataset to be compatible with algorithms such as association rule mining and cluster analysis, ensuring that the algorithms can effectively identify hidden patterns in the data and quantify the operation results. Abstract mining rules and clustering groups are transformed into intuitive and comparable numerical features, which not only meets the computational needs of subsequent fuzzy comprehensive evaluation models but also makes the performance of each dimension of quality performance quantifiable, thereby improving the objectivity and persuasiveness of the evaluation results.

[0069] In one embodiment of this disclosure, association rule mining and clustering analysis operations are performed based on a dataset to obtain the operation results, including: Association rule mining algorithms are used to mine the relationships between data in different dimensions of a dataset and obtain the corresponding set of association rules.

[0070] In some embodiments, based on the requirements of quality performance analysis, a minimum support threshold and a minimum confidence threshold are preset. The minimum support threshold is used to filter frequently occurring data combinations, and the minimum confidence threshold is used to filter rules with strong correlations. The dataset is traversed and analyzed to count the frequency of each data item in the dataset. Individual data items with a frequency not lower than the minimum support threshold are selected to form a basic frequent set. The basic frequent set is then combined pairwise to generate candidate combination sets. The common frequency of each candidate combination set in the dataset is counted, and candidate combination sets with a frequency not lower than the minimum support threshold are retained to obtain a secondary frequent set. Higher-level candidate sets are generated iteratively based on the secondary frequent set, and the frequency counting and filtering steps are repeated until no new frequent sets satisfying the minimum support threshold can be generated. For each generated frequent set, the confidence of the association rule it contains is calculated. The confidence is calculated as the ratio of the frequency of the frequent set to the frequency of the preceding data item. Association rules with a confidence not lower than the minimum confidence threshold are selected to form a set of strong association rules containing the preceding data item, the succeeding data item, the support value, and the confidence value.

[0071] Based on different dimensions, corresponding quality-related indicators are selected; using these indicators as cluster centers, the dataset is clustered using a clustering algorithm to obtain quality groups with different risks.

[0072] In some embodiments, the process parameter deviation interval quantification value, raw material qualification rate interval quantification value, inspection qualification rate interval quantification value, and customer complaint number in the data set are selected as clustering feature variables. The clustering feature variables are normalized to map each variable value to the [0,1] interval. The K-Means algorithm is used, and the candidate range for the number of clusters is set to 2-5. The optimal number of clusters is determined by calculating the elbow rule curve under different numbers of clusters. Multiple data samples corresponding to the optimal number of clusters are randomly selected as initial cluster centers, and the Euclidean distance between each data sample and each initial cluster center is calculated. Each data sample is assigned to the cluster where the cluster center with the smallest Euclidean distance is located to form an initial cluster. The mean of the feature variables of all samples in each initial cluster is calculated, and this mean is used as the new cluster center. The steps of distance calculation, sample allocation, and cluster center update are repeated until the clustering process converges. The specific information of each cluster is output, including the mean of the feature variables of each cluster, the number of data samples included, and the clustering label of each sample.

[0073] Using the data mining and analysis dataset as input features, and the future batch quality performance level and the presence of specific defects as prediction targets, the corresponding prediction results are output based on the decision tree algorithm.

[0074] In some embodiments, a decision tree algorithm is employed, using production process-related data, procurement-related data, inspection-related data, and historical quality data from the data mining and analysis dataset as input features, and the quality performance level of future production batches and the existence of specific quality defects as prediction objectives. A classification prediction model is constructed and trained based on the data mining and analysis dataset. The trained classification prediction model predicts the quality performance results corresponding to new combinations of input features, and outputs prediction results including the probability of quality performance level and the probability of defect risk. The prediction results are used to identify quality risks in advance and support the formulation and implementation of quality prevention measures.

[0075] The operational results are obtained based on the set of association rules, quality groups with different risks, and prediction results.

[0076] In this embodiment, association rule mining accurately captures strong correlations between data from different dimensions through standardized threshold screening and iterative calculation. The K-Means clustering algorithm, through feature variable selection, normalization processing, and dynamic iterative optimization, divides massive amounts of data into quality groups with different risk levels, achieving accurate clustering and visualization of quality risks. The decision tree algorithm constructs a prediction model based on multi-dimensional input features, predicting the quality performance and defect risks of future batches in advance, providing forward-looking support for the formulation of preventive measures. The integrated application of the results of these three methods realizes full-chain quality control from historical pattern mining and current risk clustering to future trend prediction, which is conducive to improving the accuracy of subsequent quality analysis, the timeliness of risk identification, and the scientific nature of quality decision-making.

[0077] In one embodiment of this disclosure, the operation results are quantified to obtain quantitative characteristics of quality performance, including: The association rule set is filtered to obtain strong association rules with confidence greater than the first threshold and lift greater than the second threshold. Based on the strong association rules, causal relationships between indicators of different dimensions are extracted and transformed into first numerical features. Features are extracted for quality groups with different risks, and the corresponding second numerical features are calculated based on the dimensional distribution characteristics of samples within clusters. The quality performance level of future batches is mapped to a numerical level quantification value, and the existence of specific defects is transformed into defect risk probabilities to obtain the corresponding third numerical features. Based on the process execution dimension, product quality dimension, supply chain collaboration dimension, and customer feedback dimension, the first, second, and third numerical features are mapped to the corresponding quantitative features of quality performance.

[0078] In some embodiments, one implementation of quantifying association rules includes: setting a first threshold of 0.8 and a second threshold of 3; firstly filtering strong association rules according to confidence and lift; calculating the association strength between antecedents and consequents in the rules using the lift formula; extracting causal relationships such as excessive deviation of process parameters and appearance defects, and substandard raw material qualification rate and customer complaints; and standardizing the association strength value into a first numerical feature within the range of [0,1].

[0079] One implementation of cluster quantification includes: for quality groups with different risks, calculating the mean and variance of features such as process parameter deviation and raw material qualification rate within each cluster, mapping the mean to a quantified score of 0-10, and converting the variance into a risk dispersion coefficient to form a second numerical feature.

[0080] One implementation method for quantifying the quality performance of future batches includes: using a grade mapping method to map the quality performance grades of future batches—excellent, good, average, and poor—to numerical grade quantification values ​​of 90, 75, 60, and 40, respectively; using a decision tree model to output the probability value of the existence of a specific defect as a third numerical feature; and finally, combining the analytic hierarchy process (AHP) to determine the weights of each dimension, and classifying and integrating the three types of numerical features according to the dimensions of process execution, product quality, supply chain collaboration, and customer feedback to generate quality performance quantification features of 0-100 points for each dimension.

[0081] In this embodiment, through multi-dimensional feature quantification and systematic integration, abstract association rules, cluster groups, and prediction results are transformed into standardized and comparable numerical performance characteristics. This not only ensures the reliability of the association by using confidence and lift degree screening, but also realizes the quantification of risks and prediction results through cluster feature extraction and level mapping. This allows the core influencing factors, risk distribution, and future trends of quality performance to be intuitively reflected in quantitative form, thereby improving the scientific nature and operability of quality control.

[0082] like Figure 2 As shown, in one embodiment of this disclosure, an evaluation model with multi-dimensional quality performance indicators adapted to quantitative characteristics is constructed based on domain requirements and business scenarios. The evaluation model determines the weights of the performance indicators based on the analytic hierarchy process (AHP), including: Step S202: Determine multi-dimensional quality performance indicators based on domain requirements and business scenarios. These multi-dimensional quality performance indicators include product quality-related indicators under the unit performance dimension, functional performance-related indicators under the grassroots performance dimension, and PDCA management-related indicators under the process performance dimension.

[0083] Step S204: Establish a hierarchical structure model based on the analytic hierarchy process (AHP). Take quality performance evaluation as the target layer, unit performance dimension, grassroots performance dimension and process performance dimension as the first-level indicator layer, and product quality related indicators, functional performance related indicators and PDCA management related indicators as the second-level indicator layer.

[0084] In some embodiments, based on ISO22163 (the quality management standard for the rail transit industry) and the company's internal quality management specifications, the core evaluation dimensions (such as process control, product conformity, customer satisfaction, etc.) that are required by the standard can be extracted, and candidate quality performance indicators can be determined by combining the company's business processes (procurement-production-inspection-after-sales) and organizational structure.

[0085] Step S206: Perform relative importance judgment on the first-level indicator layer and the second-level indicator layer to obtain the judgment result. Construct the corresponding judgment matrix based on the judgment result. The judgment matrix is ​​an N-order positive reciprocal matrix, where N is the number of indicators at the corresponding level. Each element in the matrix is ​​used to represent the relative importance between the two corresponding indicators.

[0086] Step S208: Based on the consistency check of the judgment matrix and the eigenvalue calculation operation, determine the weight of the performance indicators.

[0087] Step S210: Based on the fuzzy comprehensive evaluation framework, integrate multi-dimensional quality performance indicators and their corresponding weights to construct an evaluation model.

[0088] In some embodiments, taking rail transit component manufacturers as an application scenario, multi-dimensional quality performance indicators can be determined based on the ISO22163 standard and the company's internal quality management specifications, combined with the business processes of procurement, production, inspection, and after-sales service. These indicators include: under the unit performance dimension, product quality-related indicators such as product qualification rate, key characteristic compliance rate, and defect rework rate; under the grassroots performance dimension, functional performance-related indicators such as process execution rate, quality responsibility implementation rate, and problem response timeliness rate; and under the process performance dimension, PDCA management-related indicators such as plan rationality, execution effectiveness, inspection comprehensiveness, and improvement closure rate.

[0089] After establishing a hierarchical structure model based on the Analytic Hierarchy Process (AHP), the relative importance of indicators was evaluated using a 1-9 scale method in conjunction with quality experts. For example, in the first-level indicator layer, the unit performance dimension was significantly more important than the grassroots performance dimension (scale 5) and slightly more important than the process performance dimension (scale 3). The grassroots performance dimension was slightly more important than the process performance dimension (scale 3). Based on this, a first-level indicator judgment matrix was constructed. The second-level indicator layer used the upper-level dimensions as benchmarks, and corresponding judgment matrices were constructed through expert evaluation. The eigenvalues ​​and weight vectors of each judgment matrix were calculated using the square root method. The rationality of the judgment logic was verified through a consistency test (consistency ratio CR < 0.1). Finally, the weights of the first-level indicators were determined to be 0.56 for the unit performance dimension, 0.28 for the grassroots performance dimension, and 0.16 for the process performance dimension. The second-level indicators obtained corresponding local weights according to their relative importance. The weights of each level were combined with the quantitative features mentioned above and embedded into the fuzzy comprehensive evaluation model to complete the full-process quality performance evaluation model.

[0090] In this embodiment, the use of industry standards and actual business operations as dual bases ensures the relevance and comprehensiveness of quality performance indicators. The analytic hierarchy process transforms expert qualitative judgments into quantitative weights, and the consistency test ensures the scientific and reasonable allocation of weights. Furthermore, the integration of a fuzzy comprehensive evaluation model achieves the adaptation of quantitative features and multi-dimensional indicators, breaking the limitations of fragmented indicators and subjective weights in traditional evaluations. This allows quality performance evaluation to cover dimensions such as products, functions, and processes, thereby enabling the output evaluation results to more accurately reflect the actual level of enterprise quality management.

[0091] In one embodiment of this disclosure, the weights of performance indicators are determined based on consistency checks and eigenvalue calculations of the judgment matrix, including: Calculate the largest eigenvalue and its corresponding eigenvector of the judgment matrix; calculate the consistency index based on the largest eigenvalue and the order of the judgment matrix; select the corresponding random consistency index based on the order of the judgment matrix; calculate the ratio between the consistency index and the random consistency index as the consistency ratio; if the consistency ratio is less than the preset standard, the judgment matrix is ​​deemed to have passed the consistency test; if it passes the consistency test, normalize the eigenvector corresponding to the largest eigenvalue to obtain the weight of the performance index for the corresponding level.

[0092] In some embodiments, taking the first-level indicator layer (unit performance dimension, grassroots performance dimension, and process performance dimension) of the quality performance evaluation of rail transit component enterprises as an example, its judgment matrix is ​​a third-order positive reciprocal matrix. The matrix elements are determined based on the expert 1-9 scaling method. The specific judgment matrix is ​​as follows:

[0093] The largest eigenvalue of the matrix was calculated using the eigenvalue method. Through matrix operations, the largest eigenvalue was found to be λmax ≈ 3.0092, and the corresponding eigenvector was [0.577, 0.289, 0.134]. .

[0094] The consistency index is calculated using the formula CI = (λmax - n) / (n - 1) (where n is the matrix order). Substituting the values, we get CI = (3.0092 - 3) / (3 - 1) = 0.0046.

[0095] Consulting the random consistency index table, the random consistency index RI for a 3-order matrix is ​​0.58.

[0096] The consistency ratio CR = CI / RI = 0.0046 / 0.58 ≈ 0.0079 was calculated. The preset standard is CR < 0.1. The judgment matrix passed the consistency test.

[0097] For the eigenvector [0.577, 0.289, 0.134] After normalization, the normalized vector [0.56, 0.28, 0.16] is obtained by summing the values ​​according to the proportion of each element. , which are the weights of the first-level indicator layer, corresponding to 0.56 for unit performance dimension, 0.28 for grassroots performance dimension, and 0.16 for process performance dimension.

[0098] In this embodiment, the qualitative judgment of experts is transformed into quantitative weights through a standardized feature value calculation and consistency verification process. This not only preserves the rationality of domain experience by using the 1-9 scale method, but also prevents subjective bias in the judgment logic through consistency ratio verification, thus ensuring the reliability of weight allocation. The normalized weight vector intuitively reflects the relative importance of each indicator, providing a precise weight basis for the subsequent quantitative calculation of the evaluation model, thereby helping to improve the objectivity of the quality performance evaluation results.

[0099] In one embodiment of this disclosure, the method further includes: configuring a corresponding visualization format for the quality performance evaluation results based on at least one of data dimension attributes, numerical relationship types, and analysis target scenarios; and / or determining the corresponding display content and display method in the quality performance evaluation results based on user roles and / or permissions.

[0100] In one embodiment of this disclosure, a corresponding visualization format is configured for the quality performance evaluation results based on data dimension attributes, numerical relationship types, and analysis target scenarios, including: If the data dimension attribute is a key quality performance indicator, the corresponding display format includes dashboard format.

[0101] If the data dimension attributes are multi-category / multi-time period data, the corresponding display formats include bar charts or line charts.

[0102] If the numerical relationship type is percentage-based data, the corresponding display formats include pie charts or donut charts.

[0103] If the target scenario being analyzed includes geographically related data, the corresponding display format may include a map.

[0104] If the data dimension attribute is detailed data, the corresponding display format includes tables and reports.

[0105] If the data dimension attributes are multi-level data, the corresponding display formats include tree diagrams or sunburst charts.

[0106] In some embodiments, key quality performance indicators such as product qualification rate, key characteristic compliance rate, and customer complaint rate are displayed in the form of a dashboard, with red, yellow and green color intervals indicating the compliance range, and the current value of the core indicator and the target value are presented in real time, so that management can quickly grasp the core quality status. Multi-category data, such as the pass rate of raw materials from different suppliers and the defect rate of each production workshop, are displayed using bar charts to show the differences between categories; multi-time period data, such as the monthly inspection pass rate and quarterly customer satisfaction changes, are displayed using line charts to show trend fluctuations and support historical data comparison and analysis. Percentage-based data, such as the percentage of various defects (appearance defects, performance defects, assembly defects) and the percentage of quality problems in each link of the supply chain, are displayed using a pie chart to clearly show the proportion of data composition.

[0107] Geographically correlated data, such as the distribution of customer complaints in different regions and the quality compliance status of suppliers in different regions, are visualized using maps. The differences in quality performance across geographical dimensions are marked in the form of heat maps or marker points to support regional quality control.

[0108] Detailed data, such as product batch quality inspection results and defect handling details, are displayed in tabular form, including fields such as batch number, indicator value, whether it meets the standard, and the person responsible for handling the issue, to meet the traceability and query needs of business personnel.

[0109] Multi-level data, such as quality performance indicator systems and performance rankings of work teams under each department, can be displayed using sunburst charts to show hierarchical relationships and data proportions, or presented using tree diagrams to show hierarchical structures and performance values ​​of each node, adapting to multi-level management needs.

[0110] In this embodiment, by accurately matching the visualization format according to the data dimension attributes, numerical relationship type and analysis target scenario, the limitation of the single-dimensional display of quality data is broken, so that different types of quality performance information can be presented in an intuitive and easy-to-understand way. Dashboards, line charts and other forms meet the needs of rapid monitoring, while tables, reports and maps support in-depth analysis and traceability. It not only adapts to the viewing needs of different roles such as management and business execution, but also improves the efficiency of quality performance information transmission, and provides efficient visualization support for quality decision-making and problem rectification.

[0111] like Figure 3 As shown, a quality performance evaluation method according to another embodiment of this disclosure, taking a rail transit component manufacturing enterprise as an application scenario, includes: Step S302: Collect quality-related data from multiple sources, including the enterprise's QMS system, production management system, supply chain management system, and after-sales management system. After cleaning, transformation, and integration, construct a standardized quality performance data warehouse.

[0112] In some embodiments, standard interface protocols such as WebService and RESTful API, as well as XML and JSON data exchange formats, are used to establish stable data connection channels with enterprise QMS systems, production management systems, supply chain management systems, and after-sales management systems, breaking down data barriers between systems and ensuring real-time and accurate data transmission.

[0113] In some embodiments, multiple types of quality-related data are collected in real time or periodically according to a preset collection strategy: process data (process parameters, equipment operating status, personnel operation records), product data (design specifications, raw material information, product batches), inspection data (raw material / process / finished product inspection records), and customer feedback data (complaint information, maintenance records, satisfaction survey results), covering the entire quality management process.

[0114] In some embodiments, data cleaning techniques are used to remove noise, outliers, and duplicate values, and missing values ​​in key fields are filled by the average of data from the same batch; heterogeneous data are format-converted and encoded uniformly, and log data, text data, etc. are converted into structured formats to ensure that the data meets the fusion requirements.

[0115] In some embodiments, product batch number, production order number, and material code are used as core association identifiers to establish a semantic mapping relationship across data sources. Production process data and inspection data are associated through batch number, and customer feedback data and product data are matched through unique identifier codes. After handling semantic conflicts of multi-source data, they are integrated to form a unified quality performance data warehouse, realizing standardized and normalized data storage.

[0116] Step S304: Based on industry standards and enterprise quality control needs, establish a multi-dimensional evaluation index system, determine the weights through the analytic hierarchy process, and construct a quantitative evaluation model.

[0117] In some embodiments, based on the ISO22163 standard, enterprise standards, and actual business operations, a three-level evaluation system is constructed: the target layer, the dimension layer, and the indicator layer. The target layer is a comprehensive evaluation of quality performance. The dimension layer includes three primary indicators: unit performance, grassroots performance, and process performance. The indicator layer is further refined into secondary indicators such as product qualification rate, key characteristic compliance rate, defect rework rate (unit performance), process execution rate, quality responsibility implementation rate, and problem response timeliness rate (grassroots performance), and planning rationality, execution effectiveness, inspection comprehensiveness, and improvement closure rate (process performance). Each indicator has a clear definition, calculation method, and data source.

[0118] In some embodiments, a hierarchical model is established to clarify the hierarchical relationship between the target layer, the first-level indicator layer, and the second-level indicator layer; five quality domain experts are organized to use the 1-9 scale method to score the relative importance of each level of indicator and construct a positive and negative judgment matrix; the maximum eigenvalue and corresponding eigenvector of the judgment matrix are calculated by the square root method, the random consistency index table is consulted, the consistency ratio CR is calculated (CR < 0.1 indicates passing the test), and the rationality of the judgment logic is verified; the eigenvectors that pass the consistency test are normalized, and the weights of the first-level indicators are finally determined, and the second-level indicators obtain corresponding local weights according to their relative importance.

[0119] In some embodiments, a fuzzy comprehensive evaluation model is selected, which combines qualitative evaluation with quantitative analysis. The standardized quality data is substituted into the model with the weights of each indicator as input, and the comprehensive quantitative evaluation of quality performance is achieved through fuzzy relation matrix operations, outputting comparable quantitative evaluation results.

[0120] Step S306: Using standardized production processes, product inspections, customer feedback, and other data from the quality performance data warehouse as the processing objects, algorithms such as Apriori association rule mining, K-Means clustering, and decision tree classification are used to extract strong association rules between quality influencing factors, classify quality groups with different risk levels, and predict quality performance trends, thereby forming applicable quality knowledge and quantitative characteristics.

[0121] In some embodiments, the Apriori algorithm is used, with a preset minimum support of 10% and a minimum confidence of 80%. It traverses the quality performance data warehouse, filters frequent data combinations to generate frequent sets, calculates the confidence and lift of association rules, and extracts strong association rules such as process parameter exceedance, appearance defects, unqualified raw materials, and customer complaints, providing a basis for tracing the source of quality problems.

[0122] In some embodiments, the quantified value of the deviation range of process parameters, the pass rate of raw materials, the pass rate of inspection, and the number of customer complaints are selected as clustering feature variables. After normalization, the K-Means algorithm is used to determine the optimal number of clusters (3 classes) through the elbow rule. Through iterative calculation, the data is divided into high-risk, medium-risk, and low-risk quality groups, and the mean value of each cluster feature, the number of samples, and the attribution label are output to identify different quality problem patterns.

[0123] In some embodiments, a decision tree classification model is constructed, with production process, procurement, inspection and historical quality data as input features, and the future batch quality performance level and the existence of specific defects as prediction targets. The model is trained based on a quality performance data warehouse and outputs prediction results including performance level probability and defect risk probability to support the formulation of quality prevention measures.

[0124] Step S308: Using the quantitative features obtained from data mining and the standardized raw data in the quality performance data warehouse as the processing objects, substitute them into the evaluation model to calculate the quantitative scores of each indicator, each dimension and the overall quality performance.

[0125] In some embodiments, standardized data from the data warehouse and quantitative features (association strength value, clustering probability, and predicted risk value) obtained through data mining are substituted into the fuzzy comprehensive evaluation model to calculate the scores of each secondary indicator, the primary dimension, and the overall quality performance comprehensive score, thereby quantitatively presenting the enterprise's quality performance level.

[0126] In some embodiments, the evaluation results of different time periods (month / quarter / year), departments, and product lines are compared to assess the effectiveness of quality improvement measures and performance trends. Further drill-down analysis is conducted on low-scoring dimensions. For example, when the performance of a production unit is low, the core reasons such as poor control of process parameters and inadequate implementation of quality responsibilities are identified by breaking down the performance of its subordinate indicators, and targeted improvement suggestions are formed.

[0127] Step S310: Using the quantitative scores, trend changes, problem distribution, and other results data obtained from performance evaluation, as well as the permission requirements of different user roles, as the processing objects, a responsive interface is designed using a B / S architecture, adapting to diverse display formats such as dashboards, bar charts, and maps, and differentiated display content is configured according to roles.

[0128] In some embodiments, a responsive interface is developed using a B / S architecture, adhering to the principles of simplicity and ease of use. Through clear layout, reasonable color matching, and intuitive icon design, it adapts to multiple terminal devices such as desktop computers and smartphones, ensuring convenient operation for users with different roles.

[0129] In some embodiments, key performance indicators (product qualification rate, customer complaint rate) are displayed on a dashboard, with the acceptable range marked in red, yellow and green, and the difference between the current value and the target value is presented in real time.

[0130] In some embodiments, multi-category data (defect incidence rates in each department) are compared using bar charts, and multi-time period data (monthly pass rate changes) are presented with line charts to show trends.

[0131] In some embodiments, percentage data such as defect type proportion and performance weight of each dimension are displayed using a pie chart.

[0132] In some embodiments, geographic correlation data such as the distribution of customer complaints in different regions and the supplier's quality compliance status are presented through heat maps or marker maps.

[0133] In some embodiments, detailed data such as inspection results and defect handling records are displayed in tabular reports, supporting sorting and filtering operations.

[0134] In some embodiments, data such as multi-level indicator systems and departmental performance rankings are presented using sunburst charts or tree diagrams to show hierarchical relationships.

[0135] In some embodiments, personalized access control provides customized display content based on user roles and permissions. Senior management: View the overall performance overview of the enterprise, department rankings, key indicator trends, and support multi-dimensional data comparison and analysis; Quality management personnel: Access detailed quality data for each stage, the distribution of problems, and the implementation status of improvement measures; detailed data can be drilled down. Frontline employees: Review quality standards, operational performance scores, and team rankings related to their work to enhance their quality awareness.

[0136] Results Export and Sharing: Supports exporting visualization results to Excel, PDF, and Word formats for use in quality improvement meeting reports, audit archives, or external collaborations, providing data support for decision-making.

[0137] It should be noted that the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may, for example, be executed synchronously or asynchronously in multiple modules.

[0138] The following reference Figure 4 To describe the quality performance evaluation device 400 according to this embodiment of the present invention. Figure 4 The quality performance evaluation device 400 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0139] The quality performance evaluation device 400 is represented in the form of hardware modules. Components of the quality performance evaluation device may include, but are not limited to: a fusion processing module 402, used to fuse different types of quality-related data collected from multi-source heterogeneous data systems to obtain multi-dimensional quality fusion data, where the quality-related data represents data related to quality management; a mining and analysis module 404, used to perform association rule mining and cluster analysis on the multi-dimensional quality fusion data to obtain quantitative characteristics of quality performance; a construction module 406, used to construct an evaluation model with multi-dimensional quality performance indicators adapted to the quantitative characteristics based on domain requirements and business scenarios, where the evaluation model determines the weights of performance indicators based on the analytic hierarchy process; and an evaluation module 408, used to input the quantitative characteristics of quality performance into the evaluation model to obtain the quality performance evaluation results.

[0140] Those skilled in the art will understand that various aspects of the present invention can be implemented as systems, methods, or program products. Therefore, various aspects of the present invention can be specifically implemented in the following forms: entirely in hardware, entirely in software (including firmware, microcode, etc.), or in a combination of hardware and software, collectively referred to herein as “circuit,” “module,” or “system.”

[0141] The following reference Figure 5 To describe an electronic device 500 according to this embodiment of the present invention. Figure 5 The electronic device 500 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0142] like Figure 5 As shown, the electronic device 500 is manifested in the form of a general-purpose computing device. The components of the electronic device 500 may include, but are not limited to: at least one processing unit 510, at least one storage unit 520, and a bus 530 connecting different system components (including storage unit 520 and processing unit 510).

[0143] The storage unit stores program code that can be executed by the processing unit 510, causing the processing unit 510 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 510 can perform, as follows: Figure 1 The scheme described in steps S102 to S108 shown.

[0144] Storage unit 520 may include readable media in the form of volatile storage units, such as random access memory (RAM) 5201 and / or cache memory 5202, and may further include read-only memory (ROM) 5203.

[0145] Storage unit 520 may also include a program / utility 5204 having a set (at least one) program module 5205, such program module 5205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0146] Bus 530 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0147] Electronic device 500 can also communicate with one or more external devices 570 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 500, and / or with any device that enables electronic device 500 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 550. Furthermore, electronic device 500 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 560. As shown, network adapter 560 communicates with other modules of electronic device 500 via bus 530. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0148] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or electronic device, etc.) to execute the methods according to the embodiments of this disclosure.

[0149] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible embodiments, various aspects of the invention may also be implemented as a program product comprising program code that, when the program product is run on an electronic device, causes the electronic device to perform the steps of the various exemplary embodiments of the invention described in the "Exemplary Methods" section above.

[0150] According to embodiments of the present invention, a program product for implementing the above-described method may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on an electronic device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0151] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0152] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0153] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0154] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0155] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0156] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0157] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or electronic device, etc.) to execute the methods according to the embodiments of this disclosure.

[0158] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the appended claims.

Claims

1. A method of mass performance evaluation, characterized by, The method comprises the following steps: Fusion processing of different types of quality-related data collected from a multi-source heterogeneous data system to obtain multi-dimensional quality fusion data, wherein the quality-related data represents data related to quality management; Correlation rule mining and clustering analysis of the multi-dimensional quality fusion data to obtain quantitative characteristics of quality performance; Based on the domain requirements and business scenarios, an evaluation model with multi-dimensional quality performance indicators is constructed to adapt to the quantitative characteristics, wherein the evaluation model determines the weight of the performance indicators based on the analytic hierarchy process; The quantitative characteristics of the quality performance are input into the evaluation model to obtain the quality performance evaluation results.

2. The method of claim 1, wherein, Fusion processing of different types of quality-related data collected from a multi-source heterogeneous data system to obtain multi-dimensional quality fusion data, comprising: Cleaning the collected different types of quality-related data to obtain cleaned data, wherein the multi-source heterogeneous data system includes a quality management system, a production management system, a supply chain management system, and an after-sales management system; Performing a unified format conversion operation on the cleaned data to obtain standardized structure fusion data; Content association fusion of the standardized structure fusion data to obtain the multi-dimensional quality fusion data, wherein the dimensions of the multi-dimensional quality fusion data include process execution dimension, product quality dimension, supply chain collaboration dimension, and customer feedback dimension.

3. The method of claim 2, wherein, Content association fusion of the standardized structure fusion data to obtain the multi-dimensional quality fusion data, comprising: Extracting product batch number, production order number, and material code from the standardized structure fusion data as cross-data-source association identifiers; Using the product batch number as a first association key to associate the standardized structure fusion data from the production management system and the quality management system, so that the process parameters and equipment operation data of the same product batch correspond to the inspection pass rate and defect type data to obtain a first set of associated data; Using the production order number as a second association key to associate the first set of associated data with the standardized structure fusion data from the supply chain management system, so that the same production order corresponds to the material pass rate and supplier information associated with the first set of associated data to obtain a second set of associated data; Using the material code as a third association key to associate the second set of associated data with the standardized structure fusion data from the after-sales management system, so that the same material corresponds to the customer feedback data integrated with the second set of associated data to obtain the multi-dimensional quality fusion data.

4. The method of claim 1, wherein, Correlation rule mining and clustering analysis of the multi-dimensional quality fusion data to obtain quantitative characteristics of quality performance, comprising: Discretization processing of continuous index data in the multi-dimensional quality fusion data to map to corresponding numerical intervals to obtain discretized data; Feature encoding processing of classification data in the multi-dimensional quality fusion data to obtain classification feature data adapted to mining operations; Based on the discretized data, the classification feature data, and the core fields in the multi-dimensional quality fusion data, a mining analysis data set is constructed; Performing association rule mining and clustering analysis operations based on the mining analysis dataset, obtaining operation results including a set of association rules and quality group clustering results; Quantitative processing of the operation results to obtain quantitative characteristics of the quality performance, wherein the quantitative characteristics include process execution performance characteristics, product quality performance characteristics, supply chain collaboration performance characteristics, and customer feedback performance characteristics.

5. The method of claim 4, wherein, Performing association rule mining and clustering analysis operations based on the mining analysis dataset, obtaining operation results including: Mining the association relationships between different dimensions of data in the mining analysis dataset based on an association rule mining algorithm to obtain a corresponding set of association rules; Selecting corresponding quality-related indicators based on different dimensions; Performing clustering processing on the mining analysis dataset based on a clustering algorithm with the quality-related indicators as clustering centers to obtain quality groups with different risks; Taking the mining analysis dataset as input features and the future batch quality performance level and the existence of specific defects as prediction targets, outputting corresponding prediction results based on a decision tree algorithm; Obtaining the operation results based on the set of association rules, the quality groups with different risks, and the prediction results.

6. The method of claim 5, wherein, Quantitative processing of the operation results to obtain quantitative characteristics of the quality performance, including: Filtering the set of association rules to obtain strong association rules with a confidence greater than a first threshold and a lift greater than a second threshold, extracting causal association relationships between different dimension indicators based on the strong association rules, and converting them into first numerical features; Extracting features from the quality groups with different risks, calculating corresponding second numerical features based on the dimension distribution characteristics of samples within the clustering clusters; Mapping the future batch quality performance level to a numerical level quantization value and converting the existence of specific defects to a defect risk probability to obtain corresponding third numerical features; Mapping the first numerical features, the second numerical features, and the third numerical features to corresponding quantitative characteristics of the quality performance based on the process execution dimension, the product quality dimension, the supply chain collaboration dimension, and the customer feedback dimension.

7. The method of claim 1, wherein, Based on the domain requirements and business scenarios, an evaluation model with multi-dimensional quality performance indicators is constructed to adapt to the quantitative characteristics, and the evaluation model determines the weights of the performance indicators based on the analytic hierarchy process, including: Determining the multi-dimensional quality performance indicators based on the domain requirements and the business scenarios, the multi-dimensional quality performance indicators including product quality-related indicators under the unit performance dimension, function performance-related indicators under the grassroots performance dimension, and PDCA management-related indicators under the process performance dimension; Establishing a hierarchical structure model based on the analytic hierarchy process to take quality performance evaluation as the target layer, the unit performance dimension, the grassroots performance dimension, and the process performance dimension as the first indicator layer, and the product quality-related indicators, the function performance-related indicators, and the PDCA management-related indicators as the second indicator layer; The relative importance of the first indicator layer and the second indicator layer is judged to obtain a judgment result, and a corresponding judgment matrix is constructed based on the judgment result, wherein the judgment matrix is an N-order positive reciprocal matrix, N is the number of indicators in the corresponding level, and each element in the matrix is used to represent the relative importance between the corresponding two indicators; The weight of the performance indicator is determined based on the consistency test and eigenvalue calculation operation of the judgment matrix; The multi-dimensional quality performance indicator and the corresponding weight are fused based on the fuzzy comprehensive evaluation framework to construct the evaluation model.

8. The method of claim 7, wherein, The weight of the performance indicator is determined based on the consistency test and eigenvalue calculation operation of the judgment matrix, including: The maximum eigenvalue of the judgment matrix and the corresponding eigenvector are calculated; According to the maximum eigenvalue and the order of the judgment matrix, a consistency index is calculated; Based on the order of the judgment matrix, a corresponding random consistency index is selected; The ratio between the consistency index and the random consistency index is calculated as a consistency ratio; If the consistency ratio is less than a preset standard, it is determined that the judgment matrix passes the consistency test; If the consistency test is passed, the eigenvector corresponding to the maximum eigenvalue is normalized to obtain the weight of the performance indicator in the corresponding level.

9. The method of claim 1, wherein, Further comprising: Based on at least one of data dimension attributes, numerical relationship types, and analysis target scenarios, a corresponding visual display form is configured for the quality performance evaluation result; And / or Based on user roles and / or permissions, the corresponding display content and display method in the quality performance evaluation result are determined.

10. The method of claim 9, wherein, Based on data dimension attributes, numerical relationship types, and analysis target scenarios, a corresponding visual display form is configured for the quality performance evaluation result, including: If the data dimension attribute is a key quality performance indicator, the corresponding display form includes a dashboard form; If the data dimension attribute is multi-category / multi-time period data, the corresponding display form includes a column chart or a line chart; If the numerical relationship type is a proportion type data, the corresponding display form includes a pie chart or a ring chart; If the analysis target scenario includes geographic correlation data, the corresponding display form includes a map; If the data dimension attribute is detailed data, the corresponding display form includes a table report; If the data dimension attribute is multi-level data, the corresponding display form includes a tree chart or a sun chart.

Images