Construction site quality management evaluation system based on fuzzy C-means clustering

By introducing nonlinear standardization, adaptive distance and local density into the fuzzy C-mean clustering algorithm, the problem of single sensitivity and distance measurement of initial cluster center selection is solved, and the clustering accuracy of construction site quality management is improved and Reliability, achieving a more comprehensive quality assessment.

CN118761689BActive Publication Date: 2024-11-15SICHUAN ZHIHAO ENG TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411254986.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-09
Publication Date
2024-11-15
Estimated Expiration
2044-09-09

AI Technical Summary

Technical Problem

The existing fuzzy C-mean clustering algorithm has the problem of single initial cluster center selection sensitivity and distance measurement in construction site quality management, which leads to inaccurate clustering results and it is difficult to fully capture the characteristics of complex quality index data.

Method used

The construction site quality management evaluation system based on fuzzy C mean clustering is adopted, and the fuzzy membership matrix update and objective function design are optimized through nonlinear standardized data, adaptive distance calculation and local density considerations, combined with sparse coding and nonlinear feature mapping. Ensure the accuracy and stability of clustering results.

Benefits of technology

It significantly improves the processing ability of complex quality indicator data, improves the accuracy and reliability of clustering results, can better reflect the actual distribution of data, reduces clustering deviations caused by data abnormalities or noise, and enhances quality. Reliability of assessment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118761689B_ABST
    Figure CN118761689B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of data processing technology, and specifically relates to a construction site quality management and evaluation system based on fuzzy C-means clustering, the system comprising: a data acquisition and standardization unit, a fuzzy C-means clustering unit and a quality management and evaluation unit; the data acquisition and standardization unit is used to acquire quality index data of multiple dimensions of the construction site to obtain standardized quality indicators; the fuzzy C-means clustering unit is used to initialize a fuzzy membership matrix and calculate the weighted cluster center of the standardized quality indicator; the quality management and evaluation unit is used to calculate the local density consistency index of each element in the standardized quality indicator in its corresponding cluster based on the clustering result to obtain the quality management and evaluation result. The present invention significantly improves the processing capability of complex quality indicator data, and also ensures the accuracy and reliability of quality evaluation through refined cluster analysis and adaptive adjustment mechanism.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of data processing, and in particular relates to a construction site quality management and evaluation system based on fuzzy C-means clustering. Background Art

[0002] Quality management of construction sites is a key link in engineering construction, which is directly related to the safety, durability and final performance of the project. In the traditional quality management process, it usually relies on methods such as manual inspection, statistical analysis and empirical judgment. Although these methods can guarantee the construction quality to a certain extent, due to the large amount of data and the multidimensional and complex quality indicators, manual methods are often difficult to fully capture and process all factors affecting quality. In addition, manual inspection is subjective and easily affected by the experience and judgment of the inspectors, resulting in inaccurate and incomplete quality assessment results. With the development of information technology, more and more automated and intelligent means are introduced into the quality management of construction sites.

[0003] In the prior art, methods based on data analysis and machine learning are gradually being applied to the field of construction quality management. For example, some studies have proposed the use of traditional machine learning algorithms such as support vector machines and random forests to classify and predict construction quality index data. These methods establish quality assessment models by learning from historical data and apply them to the assessment of new data. However, these traditional machine learning methods have certain limitations when dealing with multi-dimensional, nonlinear and complexly distributed quality index data at construction sites. First, traditional machine learning methods usually rely on a large amount of labeled data for training, but it is not easy to obtain enough high-quality labeled data at actual construction sites. In addition, these methods are prone to the "dimensionality disaster" problem when dealing with high-dimensional data, resulting in a decrease in the generalization ability of the model and difficulty in achieving good results in complex actual scenarios. In response to this problem, the fuzzy C-means clustering (FCM) algorithm has gradually attracted attention in recent years. The fuzzy C-means clustering algorithm is an unsupervised clustering method that simultaneously attributes each data point to multiple clusters through a fuzzy membership matrix and determines the relationship between the data point and each cluster based on the size of the membership. Compared with traditional hard clustering methods, fuzzy C-means clustering can better handle the fuzziness and uncertainty in data, and is especially suitable for environments such as construction sites where data is complex, multidimensional, and full of noise. However, although fuzzy C-means clustering has solved some problems in construction quality management to a certain extent, there are still some challenges in practical applications.

[0004] First, the fuzzy C-means clustering algorithm is sensitive to the selection of the initial cluster center. Different initial cluster centers may lead to different clustering results, which is particularly evident in the multidimensional data environment. The data at the construction site often presents nonlinear and high-dimensional characteristics, which makes the selection of the initial cluster center more complicated and difficult to determine. If the initial cluster center is not selected properly, the clustering results may deviate from the actual situation and reduce the accuracy of the quality assessment. Secondly, the existing fuzzy C-means clustering algorithm usually uses a simple Euclidean distance as a metric when processing multidimensional data. However, the quality indicator data at the construction site often has diversity and complex internal structures, and a single Euclidean distance is difficult to fully reflect the true relationship between data points. For example, for two data points with similar quality indicators but different values, the Euclidean distance may exaggerate the difference between them and ignore their similarity in the feature space. This will cause the clustering results to not accurately reflect the actual distribution of the data, affecting the effect of quality assessment. Summary of the invention

[0005] In view of this, the main purpose of the present invention is to provide a construction site quality management and evaluation system based on fuzzy C-means clustering, which not only significantly improves the processing ability of complex quality index data, but also ensures the accuracy and reliability of quality evaluation through refined clustering analysis and adaptive adjustment mechanism. This has important practical significance for improving the overall quality control level of construction projects, reducing construction risks, and improving engineering safety.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A construction site quality management and evaluation system based on fuzzy C-means clustering, the system comprising: a data acquisition and standardization unit, a fuzzy C-means clustering unit and a quality management and evaluation unit; the data acquisition and standardization unit is used to acquire quality index data of multiple dimensions of the construction site, take the quality index data of each dimension as an element to form a multidimensional quality index vector, perform nonlinear standardization on the multidimensional quality index vector, and obtain a standardized quality index; the fuzzy C-means clustering unit is used to initialize a fuzzy membership matrix and calculate the weighted cluster center of the standardized quality index; calculate the adaptive distance from each element in the standardized quality index to each weighted cluster center; based on the adaptive distance, consider the local density and update the fuzzy membership matrix; set the objective function of the fuzzy C-means clustering, and cluster the standardized quality index based on the objective function; the quality management and evaluation unit is used to calculate the local density consistency index of each element in the standardized quality index in its corresponding cluster based on the clustering result, then add the local density consistency indexes corresponding to all elements to obtain the corresponding quality index, compare the quality index with the preset quality management grade interval, and obtain the quality management evaluation result.

[0008] Furthermore, the quality indicator data of each dimension corresponds to a type of quality indicator data; the quality indicator data of the same dimension includes one or more samples of quality indicator data of this type; the types of quality indicator data include: structural safety indicator data, material quality indicator data, equipment management indicator data, safety management indicator data and progress management indicator data; the structural safety indicator data include: concrete strength and steel bar protection layer thickness; the material quality indicator data include: raw material qualification rate and material pre-use inspection rate; the equipment management indicator data include: equipment integrity rate and equipment regular maintenance rate; the safety management indicator data include: safety accident occurrence rate; the progress management indicator data include: construction period deviation rate, key node completion rate and rework rate.

[0009] Furthermore, the data acquisition and standardization unit uses the following formula to calculate the first The quality index vector of the dimension is nonlinearly normalized to obtain Standardized quality indicators in 4 dimensions:

[0010] ;

[0011] in, For the Standardized quality indicators in 4 dimensions; For the The original quality indicator data of each dimension; For the The average value of the quality indicator data of each dimension; For the The standard deviation of the quality indicator data in each dimension; For the The median of the quality indicator data of each dimension; and Respectively The maximum and minimum values ​​of the quality indicator data in each dimension; For the Adaptive parameters in dimensions; For the The threshold parameters of the dimensions; and All are subscript indexes; It is an absolute value operation; the standardized quality index of each dimension is taken as an element to construct the standardized quality index .

[0012] Further, The adaptive parameters of each dimension are calculated using the following formula:

[0013] ;

[0014] in, For the Dimensional quality indicator data and Covariance of the quality indicator data of the dimensions; is the number of dimensions of the quality indicator data; For the The standard deviation of the quality indicator data of the dimension; Threshold parameters for each dimension Use the following formula to calculate:

[0015] ;

[0016] in, For the The number of samples of quality indicator data in each dimension; For the The quality indicator data of the dimension samples; is the subscript index.

[0017] Furthermore, the fuzzy C-means clustering unit initializes a fuzzy membership matrix ;in, is the fuzzy membership matrix The Row, No. Elements of a column; is the number of clusters; the weighted cluster center is calculated by the following formula:

[0018] ;

[0019] in, is the subscript index; For the Standardized quality indicators in 4 dimensions; is the mean of the elements of the standardized quality index; is the L2 norm; It is the scale parameter in the weighted cluster center calculation, controlling the width of the Gaussian kernel; For the The weighted cluster centers of the clusters.

[0020] Furthermore, the fuzzy C-means clustering unit calculates the adaptive distance from each element in the standardized quality index to the center of each weighted cluster by the following formula;

[0021] ;

[0022] in, For the Normalized quality index of dimensions to weighted cluster centers Adaptive distance; Parameters that control the weight of the reconstruction error in the adaptive distance; is a nonlinear feature mapping function, Mapping to a higher dimensional feature space; is the dictionary matrix of sparse coding; For the The weighted cluster centers in the dictionary Sparse coding under ; is the parameter that controls the weight of the logistic function term in the adaptive distance; is the steepness parameter of the logistic function; is the transpose operation.

[0023] Furthermore, the fuzzy C-means clustering unit updates the fuzzy membership matrix based on the adaptive distance and taking into account the local density through the following formula:

[0024] ;

[0025] in, is the fuzziness parameter, which controls the fuzziness of clustering. The closer it is to 1, the more the clustering tends to be hard partitioned; when As it increases, the clusters become more blurred; For the The local density of weighted cluster centers; For the Normalized quality index of dimensions to weighted cluster centers distance; is the subscript index; is the updated fuzzy membership matrix.

[0026] Furthermore, the local density Use the following formula to calculate:

[0027] ;

[0028] in, For the Standardized quality indicators in 4 dimensions; Parameters that control the weight of cosine similarity in local density calculation; Representation calculation and The cosine similarity of .

[0029] Furthermore, the objective function of fuzzy C-means clustering is Use the following formula to express it:

[0030] ;

[0031] in, Operations to find the trace of a matrix; is a parameter that controls the weight of the local density term in the objective function; is a parameter that controls the strength of the cluster center smoothness constraint; express The gradient of To standardize quality indicators The Laplace matrix of ; by minimizing the objective function, the corresponding clustering result is obtained.

[0032] By adopting the above technical scheme, the present invention produces the following beneficial effects: on the basis of the fuzzy C-means clustering algorithm, the present invention introduces a multi-level data processing and analysis method, including nonlinear feature mapping, sparse coding, local density calculation, etc., which greatly improves the processing ability of complex data on the construction site. At the construction site, quality management data often has characteristics such as multi-dimensionality, nonlinearity and much noise, and traditional clustering algorithms are difficult to effectively cope with these challenges. By mapping the original data to a high-dimensional feature space and combining sparse coding technology, the present invention can better capture the potential structure and pattern in the data, thereby avoiding the problem of reduced accuracy caused by data complexity during the clustering process. In particular, the introduction of nonlinear feature mapping functions enables the system to flexibly adapt to different types of data distributions to ensure the accuracy and stability of clustering results. Secondly, the adaptive distance calculation method proposed in the present invention provides a more refined distance measurement method by combining Euclidean distance, feature mapping error and sparse regularization. This adaptive distance not only takes into account the geometric distance between the data point and the cluster center, but also takes into account the performance of the data point in the high-dimensional feature space, and further reduces the influence of noise and outliers on the clustering results through the regularization of sparse coding. In construction quality management, this multi-dimensional distance calculation method can more accurately reflect the relationship between data points and cluster centers, making the clustering results more in line with the actual situation, avoiding clustering deviations caused by data anomalies or noise, and thus improving the reliability of quality assessment. Third, the present invention comprehensively considers the local density characteristics in the updating process of the fuzzy membership matrix. This innovative design enables the clustering algorithm to pay more attention to the density of the area where the data points are located through the calculation and introduction of local density, thereby giving more weight to high-density areas in the clustering process. This processing method is particularly suitable for the management of construction site quality indicator data, because the data at the construction site is often unevenly distributed, the data density in some areas is high, and other areas are relatively sparse. Through the consideration of local density, the present invention can effectively improve the clustering accuracy of high-density areas, ensure that these key areas are more accurately analyzed and judged in quality assessment, and thus improve the evaluation effect of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A schematic diagram of the system structure of a construction site quality management and evaluation system based on fuzzy C-means clustering provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] All features disclosed in this specification, or steps in all methods or processes disclosed, except mutually exclusive features and / or steps, can be combined in any manner.

[0035] Any feature disclosed in this specification (including any additional claims and abstract), unless otherwise stated, may be replaced by other alternative features that are equivalent or have similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

[0036] Example 1: Reference Figure 1 , a construction site quality management and evaluation system based on fuzzy C-means clustering, the system comprising: a data acquisition and standardization unit, a fuzzy C-means clustering unit and a quality management and evaluation unit; the data acquisition and standardization unit is used to acquire quality index data of multiple dimensions of the construction site, take the quality index data of each dimension as an element to form a multidimensional quality index vector, perform nonlinear standardization on the multidimensional quality index vector, and obtain a standardized quality index; the fuzzy C-means clustering unit is used to initialize a fuzzy membership matrix and calculate the weighted cluster center of the standardized quality index; calculate the adaptive distance from each element in the standardized quality index to each weighted cluster center; based on the adaptive distance, consider the local density and update the fuzzy membership matrix; set the objective function of fuzzy C-means clustering, and cluster the standardized quality index based on the objective function; the quality management and evaluation unit is used to calculate the local density consistency index of each element in the standardized quality index in its corresponding cluster based on the clustering result, then add the local density consistency indexes corresponding to all elements to obtain the corresponding quality index, compare the quality index with the preset quality management level interval, and obtain the quality management evaluation result.

[0037] Specifically, first, the data acquisition process involves collecting quality indicator data of multiple dimensions of the construction site. These data usually include but are not limited to material strength, construction accuracy, environmental conditions, operator operation level, and equipment operation status. These multidimensional data together constitute a comprehensive characterization of the quality of the construction site. However, since these data come from different dimensions and sources, and have different units, dimensions, and distribution characteristics, directly applying them to the clustering algorithm will lead to data imbalance in different dimensions, thereby affecting the accuracy of the clustering results. In order to solve this problem, the data acquisition and standardization unit preprocesses these data through a nonlinear standardization method so that they can be compared and processed on the same scale. A key step in the standardization process is to deal with the dimensional differences between different quality indicators. The quality indicator data of the construction site often come from different measurement tools and sensors. For example, material strength is usually measured in megapascals (MPa), while the ambient temperature may be measured in degrees Celsius (℃). If such data of different dimensions are directly input into the fuzzy C-means clustering algorithm, data of certain dimensions will have too much influence on the clustering results, while data of other dimensions may be ignored. To this end, the present invention adopts a nonlinear standardization method, by standardizing the data of each dimension, the quality index data of different dimensions are converted into dimensionless standardized data, so that it can be processed on the same scale. Nonlinear standardization can not only eliminate the dimensional influence of the data, but also retain the distribution characteristics of the data, so that the relationship between the data of each dimension can be more accurately reflected during clustering. In the standardization process, considering the diversity and complexity of the quality indicators of the construction site, the present invention selects nonlinear standardization rather than the traditional linear standardization method. Linear standardization is usually to compress the data into a fixed interval by linear transformation, for example, compressing all data to between 0 and 1. However, when processing extremely uneven data distribution, linear standardization may lose some important features of the data, especially for the processing of extreme value data, which may cause the clustering algorithm to be unable to effectively identify these key data points. Therefore, the nonlinear standardization method adopted by the present invention is more complicated, and it can be transformed by a specific function so that the standardized data better reflects the distribution characteristics of the original data. This processing method not only retains the extreme value characteristics of the original data, but also enhances the expressiveness of the data in the clustering algorithm, so that the fuzzy C-means clustering algorithm can better identify and distinguish different categories of construction quality index data.

[0038] The standardized data are called standardized quality indicators. These data are used as the input of the fuzzy C-means clustering algorithm and are the basis for subsequent clustering analysis. The accuracy of the standardized quality indicators directly affects the reliability of the clustering results. Therefore, the data acquisition and standardization unit must not only ensure the comprehensiveness of the data, but also ensure the rationality and effectiveness of the standardization process. In this process, the present invention pays special attention to the noise processing and outlier processing of the data. In the actual collection process, the quality indicator data of the construction site will inevitably be affected by the external environment, equipment errors and operator errors, resulting in the possible existence of noise and outliers in the data. If these noises and outliers are not effectively processed, it may have an adverse effect on the standardized results, thereby affecting the accuracy of the clustering analysis. Therefore, before data standardization, the present invention uses data preprocessing technology to filter noise and process outliers on the original data to ensure that the data input to the standardization unit is relatively clean and stable. After the standardized data, the dimensional differences between different dimensions are eliminated, the distribution characteristics of the original data are retained, and after the processing of noise and outliers, the data is more suitable for application in the fuzzy C-means clustering algorithm, thereby providing a stable and reliable data basis for subsequent quality management evaluation.

[0039] In the present invention, the work of the fuzzy C-means clustering unit first starts from initializing the fuzzy membership matrix. Each element of the fuzzy membership matrix represents the probability that a certain data point belongs to a certain cluster, and the range of this probability is between 0 and 1, and for a data point, the sum of its membership in all clusters is equal to 1. Through this probability distribution mode, the fuzzy C-means clustering algorithm can flexibly handle the uncertainty and fuzziness in the data while maintaining global consistency. This processing method is particularly important in the application of construction site quality management, because the data at the construction site are often affected by multiple factors, and there are a large amount of fuzziness and uncertainty, such as environmental changes, material property differences and differences in personnel operations, etc., and the quality index data caused by these factors are difficult to be simply classified into a certain determined category. Therefore, through the introduction of fuzzy membership, the system can establish a fuzzy connection between data points and multiple categories, thereby better reflecting the complexity of actual data. After initializing the fuzzy membership matrix, the fuzzy C-means clustering unit will calculate the weighted cluster center of each cluster according to the standardized quality index data. The weighted cluster center is the center point of each cluster, which is determined by considering the membership of the data point and the weight of each dimension. The weight is introduced to reflect the importance of different quality indicators to the quality management of the construction site, which means that in the calculation process of the weighted cluster center, the system will give higher weights to certain key indicators to ensure that these important indicators can have a greater impact on the clustering results. In this way, the clustering process not only considers the membership of the data point, but also the importance of each quality indicator, so that the clustering results are more in line with the actual situation. Next, the system calculates the adaptive distance from each standardized quality indicator data point to each weighted cluster center. The calculation of the adaptive distance is a key step in the fuzzy C-means clustering algorithm, which determines the relative relationship between the data point and each cluster. The design of the adaptive distance takes into account the distribution characteristics of the data points in the multidimensional space and the attraction of the cluster center to the data point. In this way, the adaptive distance not only reflects the geometric distance between the data point and the cluster center, but also comprehensively considers the local density distribution of the data. Therefore, in the quality management of the construction site, the adaptive distance can help the system accurately distinguish data of different quality categories and can show stronger robustness when dealing with abnormal data. For example, when some data points at a construction site are abnormal due to equipment failure or human error, the calculation of the adaptive distance can identify these abnormal data and reasonably assign them to the corresponding clusters, thereby avoiding the impact of abnormal data on the overall clustering results. After calculating the adaptive distance, the system will update the fuzzy membership matrix based on these distances. The update process is an iterative core link in the fuzzy C-means clustering algorithm. By repeatedly adjusting the membership matrix, the system can gradually optimize the clustering results so that the membership of each data point can better reflect its true cluster affiliation.In this process, the fuzzy C-means clustering algorithm continuously adjusts the membership of each data point in different clusters to optimize the objective function. This process is similar to finding the optimal distribution state of data points in multidimensional space, so that each data point can find a balance point between multiple clusters to achieve the best clustering effect.

[0040] In order to make the clustering process more accurate and effective, the present invention also introduces the concept of local density. Local density refers to the density of other data points near a certain data point, which reflects the distribution characteristics of the data points in its local area. By combining the concept of local density, the system not only considers the global data distribution when updating the fuzzy membership matrix, but also pays special attention to the density of the data points in its local area. This method enables the clustering algorithm to perceive the local structure of the data more keenly, so that it can still maintain a high clustering accuracy when processing uneven data distribution. In the quality management of the construction site, this processing method is particularly important, because the quality indicator data of the construction site often presents a high degree of inhomogeneity, such as certain specific process steps or material properties may cause data to aggregate in certain areas, while in other areas it is relatively sparse. By introducing the adjustment of local density, the system can better adapt to the inhomogeneity of this data, thereby achieving a more accurate quality management evaluation. Finally, the fuzzy C-means clustering unit clusters the standardized quality indicator data based on the objective function of the fuzzy C-means clustering by setting the objective function of the fuzzy C-means clustering. The setting of the objective function takes into account multiple factors such as the degree of membership of the data points, the adaptive distance, and the local density. The purpose is to find an optimal clustering scheme in the multidimensional space so that the data points in the same cluster are as similar as possible, while the data points between different clusters are as different as possible. By optimizing the objective function, the system can gradually converge to an optimal clustering result, thereby realizing the effective classification and evaluation of the quality index data of the construction site. This process not only improves the efficiency of quality management, but also improves the ability to identify potential quality problems, providing strong support for the quality control of the construction site. The quality management evaluation unit further analyzes each standardized quality index through the clustering results. These standardized quality indicators have been classified into different clusters after fuzzy C-means clustering. Each cluster represents a collection of a certain type of quality indicator data. In this process, the system not only simply classifies the quality indicator data, but also calculates the local density consistency index of each data point in its cluster according to the clustering results. The local density consistency index is an indicator that measures the degree of similarity between a data point and other data points in the cluster. The calculation of this index reflects the position of the data point in the cluster to which it belongs and its distance relationship with the cluster center. By calculating this index, the system can determine the stability of each data point in its cluster, that is, whether the data point is representative or whether there are any anomalies.

[0041] In the quality management evaluation unit, the local density consistency index of all standardized quality index data points is accumulated to obtain a comprehensive quality index. This quality index is a comprehensive reflection of the quality status of the entire cluster. It combines the local characteristics of all data points to form an overall evaluation standard. The calculation of this quality index is the core task of the quality management evaluation unit, which directly determines the system's judgment on the quality status of the construction site. In order to ensure the accuracy and fairness of the evaluation, the quality management evaluation unit also compares this quality index with the preset quality management grade interval. The quality management grade interval is set according to the specific requirements and standards of the construction site, and different grades correspond to different quality levels and management requirements. By comparing the calculated quality index with these preset intervals, the system can accurately judge the quality management level of the construction site, identify possible quality problems or risks, and feedback this information to the management personnel. The working principle of the quality management evaluation unit is not only the simple processing of data, but more importantly, through the application of fuzzy C-means clustering results, the quality evaluation process can be more accurate and targeted. In traditional quality management systems, quality assessment often relies on a single indicator or a simple statistical method, which often cannot fully reflect the actual quality status of the site when dealing with complex, multi-dimensional construction site data. However, through the fuzzy C-means clustering algorithm, the quality management assessment unit can identify potential patterns and anomalies in the data, especially when dealing with multi-dimensional data and complex site environments. The introduction of the local density consistency index further enhances the system's assessment capabilities, making the assessment results not only a simple statistics of the overall data, but also an in-depth analysis of the internal structure and characteristics of the data.

[0042] After being processed by the fuzzy C-means clustering unit, the data points of each quality indicator are classified into different clusters. For example, suppose a set of standardized concrete strength data points are classified into the "qualified" cluster. In this cluster, the local density consistency index of each data point is calculated. This index reflects the degree of similarity between the concrete strength data point and other data points in the cluster, that is, whether the data point is representative or an outlier. For example, if a data point is very different from other points in the cluster, its local density consistency index will be low, indicating that the data point may be an outlier. The local density consistency index of all data points belonging to the "qualified" cluster is accumulated to obtain an overall quality index. This quality index reflects the comprehensive quality of the cluster (the "qualified" cluster). Assume that the quality index of the cluster is 85 (within a predetermined scoring range). In practical applications, the quality management grade interval is usually set based on historical data, industry standards, and expert experience. Taking the construction industry as an example, the quality management grade interval may be obtained based on the quality assessment data statistics of a large number of historical projects, or formulated by reference to industry standards. For example, based on a large amount of quality assessment data on concrete strength in previous projects, after statistical analysis, the following quality management grade intervals were obtained: 90-100: Excellent (the construction quality within this interval is high, with basically no quality problems); 70-89: Qualified (the construction quality within this interval basically meets the standards, but there are a few acceptable defects); 0-69: Needs improvement (the construction quality within this interval does not meet the standards and remedial measures are required). The setting of these intervals takes into account the actual situation during the construction process and is adjusted according to industry standards. For example, in a certain project, due to the particularity of the environment (such as high temperature or high humidity environment), the requirements for concrete strength may be appropriately reduced. At this time, the lower limit of the "qualified" interval may be lowered to 65 points.

[0043] Embodiment 2: The quality indicator data of each dimension corresponds to a type of quality indicator data; the quality indicator data of the same dimension includes one or more samples of quality indicator data of this type; the types of quality indicator data include: structural safety indicator data, material quality indicator data, equipment management indicator data, safety management indicator data and progress management indicator data; the structural safety indicator data include: concrete strength and steel bar protection layer thickness; the material quality indicator data include: raw material qualification rate and material pre-use inspection rate; the equipment management indicator data include: equipment integrity rate and equipment regular maintenance rate; the safety management indicator data include: safety accident occurrence rate; the progress management indicator data include: construction period deviation rate, key node completion rate and rework rate.

[0044] Specifically, the structural safety index data includes concrete strength and steel bar protection layer thickness, which are key indicators to ensure the safety and durability of building structures. Concrete strength is the core parameter for measuring the bearing capacity of buildings, while the steel bar protection layer thickness is directly related to the durability and anti-corrosion performance of steel bars. Accurate measurement and evaluation of these indicators can effectively prevent the occurrence of structural safety accidents. Material quality index data focuses on the quality control of raw materials, including the raw material qualification rate and the material pre-use inspection rate. The raw material qualification rate reflects whether the materials entering the construction site meet the design standards and specifications, while the material pre-use inspection rate ensures that the materials have undergone necessary quality inspections before use to prevent unqualified materials from affecting the construction quality. Through strict monitoring and analysis of these data, the system can timely detect material quality problems and reduce construction risks caused by material problems. Equipment management index data covers equipment integrity rate and equipment regular maintenance rate. Equipment integrity rate is a key indicator to ensure the normal operation of construction equipment, while equipment regular maintenance rate reflects the timeliness and effectiveness of equipment maintenance. These indicators are essential to ensure the continuity of the construction process and the safe operation of equipment. By evaluating the equipment management index data, the system can timely identify the weak links in equipment management and prevent the impact of equipment failure on construction progress and quality. Safety management index data mainly includes the incidence rate of safety accidents, which is an important indicator for evaluating the safety management level of construction sites. The incidence rate of safety accidents directly reflects whether the safety management measures on site are in place and whether the workers' safety awareness is high enough. By analyzing this data, the system can comprehensively evaluate the safety status of the construction site, identify potential safety hazards, and then take measures to reduce the occurrence of safety accidents. Progress management index data includes the construction period deviation rate, key node completion rate and rework rate, which are important indicators for evaluating construction progress and plan execution. The construction period deviation rate measures the difference between the actual construction progress and the planned progress, the key node completion rate reflects whether the important nodes in the construction process are completed on time, and the rework rate indicates the stability of construction quality and work efficiency. By analyzing these indicators, the system can identify problems in construction progress management and help managers adjust construction plans in a timely manner to ensure that the project is completed on time and with quality.

[0045] Embodiment 3: The data acquisition and standardization unit uses the following formula to obtain the first The quality index vector of the dimension is nonlinearly normalized to obtain Standardized quality indicators in 4 dimensions:

[0046] ;

[0047] in, For the Standardized quality indicators in 4 dimensions; For the The original quality indicator data of each dimension; For the The average value of the quality indicator data of each dimension; For the The standard deviation of the quality indicator data in each dimension; For the The median of the quality indicator data of each dimension; and Respectively The maximum and minimum values ​​of the quality indicator data in each dimension; For the Adaptive parameters in dimensions; For the The threshold parameters of the dimensions; and All are subscript indexes; It is an absolute value operation; the standardized quality index of each dimension is taken as an element to construct the standardized quality index .

[0048] Specifically, first, the first part of the formula is obtained by With its mean The normalization of the difference between As a nonlinear transformation tool. This transformation is not a simple linear scaling, but a The application of the function allows the distribution characteristics of the data to be displayed more smoothly. The advantage of the hyperbolic sine function is that it can gently stretch and compress the data. Especially when there are extreme values ​​or outliers in the data set, this nonlinear processing can effectively reduce the impact of these outliers on the overall clustering results. At the same time, by introducing a summation term involving other dimensions in the denominator, the formula takes into account the multidimensional characteristics of the data, so that the data of each dimension not only depends on its own statistical characteristics during the standardization process, but is also affected by the data of other dimensions, thereby enhancing the correlation between multidimensional data. This processing method is particularly important in the context of construction site data, because the quality indicators in the construction process are often interrelated, and analyzing the data of a certain dimension alone may ignore its interactive relationship with other dimensions. The next part further smoothes the distribution of the data by logarithmically transforming and normalizing the data. In the formula The influence of outliers on the standardized results is cleverly reduced by performing logarithmic processing on the difference between the original data and its median. The introduction of logarithmic transformation can not only compress a large range of data and make it more concentrated, but also improve the resolution of the data, so that even small changes can be reflected in the cluster analysis. In the context of construction quality management, this treatment is particularly suitable for processing quality fluctuation data that may be caused by factors such as site environment and material batches, so that these data can more accurately reflect the actual situation in subsequent clustering.

[0049] In addition, the formula also compresses the data to a more reasonable range by normalization. This step makes the data of all dimensions normalized to a unified scale regardless of their original magnitude, ensuring the fairness and consistency of subsequent cluster analysis. The normalization process is not only a routine step in data processing, but also specially designed in the present invention to deal with the heterogeneity of multidimensional quality index data. Different quality index data vary on different scales. Through normalization, the system can eliminate these scale differences, so that the fuzzy C-means clustering algorithm can focus on the essential characteristics of the data rather than its magnitude differences. Finally, through the introduction of the Sigmoid function, the formula further nonlinearly maps the data to generate a standardized result between 0 and 1. This step is particularly critical for processing quality index data with threshold characteristics. The S-curve characteristics of the Sigmoid function enable the changes in data to be more significantly expressed near a certain threshold. This processing method is particularly suitable for quality management at construction sites. For example, in concrete strength testing, when the data approaches a critical value, the Sigmoid function can show this approaching risk through more sensitive numerical changes, helping managers to better predict potential problems.

[0050] Embodiment 4: The adaptive parameters of each dimension are calculated using the following formula:

[0051] ;

[0052] in, For the Dimensional quality indicator data and Covariance of the quality indicator data of the dimensions; is the number of dimensions of the quality indicator data; For the The standard deviation of the quality indicator data of the dimension; Threshold parameters for each dimension Use the following formula to calculate:

[0053] ;

[0054] in, For the The number of samples of quality indicator data in each dimension; For the The quality indicator data of the dimension samples; is the subscript index.

[0055] Specifically, adaptive parameters The calculation of is based on the concept of covariance. The purpose of this parameter is to capture the relationship between different quality indicator dimensions. In the multidimensional quality management data of the construction site, the data of each dimension is usually not isolated, and there is a certain degree of correlation between them. In order to accurately reflect these correlations, the present invention quantifies the first The relationship between the first dimension and other dimensions. Covariance is a statistical indicator used to measure the linear relationship between two variables. It can show whether and how the data of another dimension changes when the data of one dimension changes. The data covariance of each dimension is normalized with other dimensions, and combined with the standard deviation of each dimension, an adaptive parameter reflecting the overall correlation of the dimension with other dimensions can be calculated. This parameter plays a regulatory role. In the subsequent standardization and clustering process, it can dynamically adjust the weights of different dimensions, so that the system can more accurately process multidimensional data with strong correlations. In practical applications, for example, when analyzing the relationship between concrete strength and steel bar cover thickness, this adaptive parameter can help the system better understand and analyze the interactive impact between the two, thereby improving the accuracy of clustering. Next, the threshold parameter The calculation of focuses on the distribution characteristics of data, especially high-order statistical characteristics. In the quality management of construction sites, the distribution of data is often not a normal distribution, and there may be skewness or peakedness. In order to capture these more complex distribution characteristics, the calculation of threshold parameters introduces the concept of skewness. Skewness is a statistic that describes the asymmetry of data distribution, which can reveal the degree of skewness of data around the mean. By calculating the first The skewness of each dimension can help us understand the extent to which the data deviates from the normal distribution and adjust the threshold parameters based on this information. Specifically, the calculation of the threshold parameters not only considers the mean and standard deviation of the data, but also takes into account the distribution characteristics of the data through skewness, thereby adjusting the treatment of outliers in the standardization process. For example, when the data on concrete strength is skewed, the threshold parameter It can be adjusted dynamically so that more attention is paid to these deviations in the standardization process, ensuring that these deviations will not be ignored in the subsequent cluster analysis, but will be reasonably weighted. The core idea of ​​this series of parameter calculations is to apply different processing strategies to quality indicator data of different dimensions during the standardization and clustering process, so that the system can flexibly respond to various complex field data. Through the dynamic adjustment of adaptive parameters, the system can effectively identify and process the correlation between different dimensions to ensure the accuracy and stability of clustering results. The introduction of threshold parameters enhances the system's perception of data distribution characteristics, so that the system can not only process evenly distributed data, but also deal with quality indicator data with special distribution characteristics, thereby improving the accuracy and reliability of the overall evaluation. In actual construction quality management applications, this refined parameter adjustment is of great significance. For example, in a complex construction project, data of multiple dimensions such as material quality, equipment management, and progress control are intertwined, and any abnormality in any dimension may affect the quality of the entire project. By introducing adaptive parameters and threshold parameters, the system can dynamically adjust the degree of attention to data in each dimension, thereby more accurately reflecting the quality status in actual construction. Especially when the data has nonlinear relationships or abnormal distributions, these parameters can effectively capture and reflect these characteristics, allowing the system to maintain a high level of analysis and evaluation capabilities when faced with complex data.

[0056] Example 5: Fuzzy C-means clustering unit initializes a fuzzy membership matrix ;in, is the fuzzy membership matrix The Row, No. Elements of a column; is the number of clusters; the weighted cluster center is calculated by the following formula:

[0057] ;

[0058] in, is the subscript index; For the Standardized quality indicators in 4 dimensions; is the mean of the elements of the standardized quality index; is the L2 norm; It is the scale parameter in the weighted cluster center calculation, controlling the width of the Gaussian kernel; For the The weighted cluster centers of the clusters.

[0059] Specifically, the weighted cluster center The calculation of is a key step in the clustering process, which directly determines the center position of each cluster and affects the clustering results and accuracy. In the present invention, the calculation of cluster centers is completed by a complex weighted formula, which combines fuzzy membership and Gaussian kernel function to ensure that the cluster center can more accurately reflect the distribution characteristics of the data. Specifically, the fuzzy membership The introduction of makes it possible to calculate the cluster center by taking into account the degree to which each data point belongs to different clusters, where As a fuzzy control parameter, it determines the weight of membership in cluster center calculation. The value of can make those data points with higher membership have greater influence in the cluster center calculation, so that the cluster center can better represent the core data points of the cluster. This weighted processing is particularly important in construction site quality management, because in practical applications, some data points may be more representative, while other data points may have a weakened influence on clustering due to various reasons (such as abnormal data, measurement errors, etc.). In order to further improve the accuracy of cluster center calculation, the Gaussian kernel function is also introduced in the formula , which measures each data point Standardized quality index mean The influence of data points on the cluster center is adjusted by the distance between them. The role of the Gaussian kernel function is that it can dynamically adjust the weight according to the distance between the data point and the cluster center: data points with closer distances are given higher weights, while data points with farther distances are given lower weights. This processing method can effectively enhance the robustness of the clustering process, especially when the data distribution is relatively dispersed or there are outliers. The introduction of the Gaussian kernel function can reduce the interference of these outliers on the calculation of the cluster center, making the final clustering result more accurate and stable.

[0060] In the context of construction site quality management, this cluster center calculation method combining fuzzy membership and Gaussian kernel function has significant advantages. Construction site data usually comes from a variety of different quality indicators, which not only differ in numerical value but also may show different relationships in physical sense. By introducing fuzzy membership, the system can handle data of different dimensions more flexibly, so that the same data point can belong to multiple clusters at the same time, which is particularly effective when dealing with quality indicators with fuzzy boundaries. For example, when evaluating concrete strength, the strength index may be closely related to the quality of the material and also affected by environmental factors during the construction process. Fuzzy C-means clustering can capture these complex multidimensional relationships and provide more detailed clustering results. The use of Gaussian kernel function further enhances the system's sensitivity to data distribution characteristics. In actual construction, some quality indicator data may be abnormal due to equipment failure or external interference. If these abnormal data are directly used for clustering, they may cause the shift of cluster centers, thereby affecting the overall clustering results. Gaussian kernel function reduces the weight of data points far away from the cluster center, ensuring that the cluster center calculation is more affected by those data points that reflect the actual situation, thereby improving the accuracy of clustering results. This processing method can effectively filter out the negative impact of abnormal data on clustering results, ensuring that the system can still maintain a high level of analytical capabilities when faced with complex construction site data.

[0061] Embodiment 6: The fuzzy C-means clustering unit calculates the adaptive distance from each element in the standardized quality index to the center of each weighted cluster by the following formula;

[0062] ;

[0063] in, For the Normalized quality index of dimensions to weighted cluster centers Adaptive distance; Parameters that control the weight of the reconstruction error in the adaptive distance; is a nonlinear feature mapping function, Mapping to a higher dimensional feature space; is the dictionary matrix of sparse coding; For the The weighted cluster centers in the dictionary Sparse coding under ; is the parameter that controls the weight of the logistic function term in the adaptive distance; is the steepness parameter of the logistic function; is the transpose operation.

[0064] Specifically, the basis of the formula is the square of the Euclidean distance, which directly measures the Standardized quality indicator data points With Weighted cluster centers This part of the distance reflects the relative relationship between the position of the data point in the original feature space and the cluster center, and is the core component of the adaptive distance. However, the quality indicator data of the construction site often have complex nonlinear relationships, and the simple Euclidean distance is difficult to effectively capture this complexity. Therefore, on this basis, the formula further introduces the concepts of nonlinear feature mapping and sparse coding. Nonlinear feature mapping function is a Gaussian kernel mapping that transforms the original data points Mapping to a higher-dimensional feature space, in which the nonlinear relationship between data points is better represented. This mapping makes the differences between data points more significant in the high-dimensional space, which helps to more accurately calculate the distance between them and the cluster center. Sparse coding is a way to express data in a higher-dimensional space, through a dictionary matrix and sparse coding coefficients , each cluster center The representation in feature space is further refined. By calculating the data points The difference from sparse coding is introduced into the adaptive distance, and the formula can better reflect the performance of data points in a higher-dimensional feature space. This processing method is particularly important in construction site quality management, because field data may show highly nonlinear characteristics due to the complexity of the environment, materials or operations. Through nonlinear feature mapping and sparse coding, the system can more accurately capture these characteristics and ensure the accuracy of clustering analysis. In addition, the formula also introduces sparse regularization terms and logical function terms. Sparse regularization term By controlling the complexity of sparse coding, the representation of cluster centers in the feature space is ensured to be as concise and effective as possible to avoid overfitting. This sparse processing helps to eliminate redundant features, making the clustering results more concise and representative. Especially in the case of high data dimensions and more feature redundancy, the sparse regularization term can effectively improve the robustness of clustering. The logistic function term introduces a control parameter , adjusting the intrinsic relationship between data points and cluster centers. The role of the logistic function is to perform nonlinear processing on the product of data points and cluster centers, so that those points close to the cluster center have a higher influence, while those farther away have a weakened influence. This processing method can be regarded as a softening of the relationship between data points and cluster centers, so that the clustering process can better adapt to the actual distribution characteristics of the data. For quality management on construction sites, this means that the system can better identify which data points truly belong to a certain quality category and which may just be noise or outliers.

[0065] Embodiment 7: The fuzzy C-means clustering unit updates the fuzzy membership matrix based on the adaptive distance and taking into account the local density through the following formula:

[0066] ;

[0067] in, is the fuzziness parameter, which controls the fuzziness of clustering. The closer it is to 1, the more the clustering tends to be hard partitioned; when As it increases, the clusters become more blurred; For the The local density of weighted cluster centers; For the Normalized quality index of dimensions to weighted cluster centers distance; is the subscript index; is the updated fuzzy membership matrix.

[0068] Specifically, the core idea of ​​the formula is to update the fuzzy membership matrix so that each data point In different clusters The membership degree in can be adjusted dynamically to reflect its distance from the cluster center and the local density characteristics. First, the formula introduces the fuzziness parameter , this parameter controls the fuzziness of clustering. When it is close to 1, the clustering result tends to be hard partitioning, that is, each data point will belong to a certain cluster more clearly; When it increases, the clustering results will be more ambiguous, and data points can be distributed in multiple clusters with more uniform membership. This design enables the clustering algorithm to adjust under different degrees of ambiguity to adapt to the characteristics of different data sets. In the formula, the adaptive distance It is based on the adaptive distance defined in the above embodiment. This distance not only considers the direct geometric relationship between the data point and the cluster center, but also enhances the accuracy of distance calculation through sparse coding, feature mapping and other means. When updating the fuzzy membership, the formula adopts the form of inversely proportional to the square of the distance, that is, , which means that the data points with smaller distance will get greater weight when their membership is updated. This processing method makes the data points closer to the cluster center occupy a more important position in the updated membership, which conforms to the intuitive logic of the clustering process. In addition, the local density is also introduced in the formula The concept of exponential decay function To adjust the update of membership. Local density reflects the density of other data points around a data point. This feature is particularly important in the clustering process because it can identify the core and edge areas within the cluster. For data points with higher density, their membership in the cluster should be higher, while for points with lower density, the membership should be reduced accordingly. Through this processing, the formula can effectively enhance the ability to recognize the structure within the cluster and ensure that the clustering results not only consider the global distance, but also comprehensively consider the distribution characteristics of local data. Finally, the normalization term in the formula ensures that the sum of all elements of the membership matrix is ​​1, which makes the updated membership still meet the basic requirements of fuzzy clustering. Specifically, the formula is for all clusters The membership of the data points is normalized so that the sum of the membership of each data point is 1. This step ensures the stability and consistency of the membership matrix, and the membership of the data points is reasonably distributed among the clusters during the clustering process. In this way, the update of the fuzzy membership matrix not only depends on the geometric relationship between the data points and the cluster centers, but also comprehensively considers the local density structure and overall distribution characteristics of the data points. This multi-dimensional consideration enables the fuzzy C-means clustering algorithm to more accurately reflect the actual situation when processing complex and multi-dimensional construction site quality indicator data. For the quality management of the construction site, this means that the system can better identify the subtle differences between different quality indicators, discover potential quality problems in a timely manner, and thus improve the quality control level of the entire construction process.

[0069] Example 8: Local Density Use the following formula to calculate:

[0070] ;

[0071] in, For the Standardized quality indicators in 4 dimensions; Parameters that control the weight of cosine similarity in local density calculation; Representation calculation and The cosine similarity of .

[0072] Specifically, the first part of the formula is obtained by using the Gaussian kernel function To measure the data points and The Gaussian kernel function is a commonly used smoothing function that performs exponential decay on the Euclidean distance between data points, so that the closer the distance is, the greater the contribution to the local density. Conversely, the farther the distance is, the faster the contribution to the local density decreases. controls the size of the Gaussian kernel, which determines the decay rate of distance. When is small, only very close data points will have a significant impact on the local density; When is large, data points farther away can also have a certain impact on local density. For the quality index data of the construction site, this processing method can effectively identify those data points that are clustered together and have similar quality, thereby reflecting the density of the local area. However, considering the Euclidean distance alone cannot fully describe the similarity between data points, especially in multidimensional data space. Therefore, the formula also introduces cosine similarity , which further refines the calculation of local density by measuring the directional similarity between data points. Cosine similarity can effectively reflect the angular relationship between two data points in high-dimensional space, that is, whether they point in the same direction in multidimensional space. In quality management data, this similarity often reflects the intrinsic correlation of data points better than simple distance metrics. For example, when evaluating the quality of two different batches of materials, although some of their specific indicators may have numerical differences, if their relative change trends are similar (that is, similar directions), then they may have similar influence in quality assessment. The formula The effect of cosine similarity is further enhanced by introducing the hyperbolic tangent function , the formula can smoothly map the value of cosine similarity to the range of -1 to 1, and the parameter Controls its contribution to local density. This parameter determines the weight of cosine similarity in local density calculation. When is larger, the influence of similarity increases, which means that the system pays more attention to the directional consistency between data points; When is smaller, the importance of directional similarity is weakened and more reliance is placed on the Euclidean distance. In this way, the formula can flexibly adapt to different data characteristics and dynamically adjust the degree of attention paid to similarity when processing different types of quality indicator data. All distances Closer data points This result not only reflects the density of data points in space, but also integrates the information of directional similarity, ensuring that the local density can more accurately reflect the actual distribution of data points in its local area. In construction site quality management, this calculation method can help identify data points that are highly clustered and have similar characteristics in a local area, thereby providing more reliable local information for the fuzzy C-means clustering algorithm.

[0073] Example 9: Objective Function of Fuzzy C-Means Clustering Use the following formula to express it:

[0074] ;

[0075] in, Operations to find the trace of a matrix; is a parameter that controls the weight of the local density term in the objective function; is a parameter that controls the strength of the cluster center smoothness constraint; express The gradient of To standardize quality indicators The Laplace matrix of ; by minimizing the objective function, the corresponding clustering result is obtained.

[0076] Specifically, the first term of the objective function It is the basic component of fuzzy C-means clustering, which measures the With all cluster centers The distance between The sum of weighted memberships. Fuzzy membership is raised to the fuzziness parameter This allows the algorithm to operate with varying degrees of ambiguity. A higher value means more fuzzy clustering, i.e., each data point can be more evenly distributed in multiple clusters, while a lower The value of makes the clustering tend to be hard partitioning. By minimizing this term, the system tries to make the data points as close as possible to the center of the cluster with higher membership, so that the clustering results are more representative. The second term of the objective function It reflects the reconstruction error of data points in high-dimensional feature space. Through nonlinear feature mapping function , the original data point It is mapped into a higher dimensional space to capture the nonlinear characteristics of the data. Through the dictionary These high-dimensional features are sparsely represented. Minimizing this term means that the system attempts to minimize the reconstruction error in the feature space through sparse coding, thereby ensuring that the cluster center can more accurately reflect the high-dimensional structural characteristics of the data. is a sparse regularization term, which minimizes the sparse coding norm, the system can control sparsity, ensure that the feature representation is as concise as possible, and avoid overfitting. This is especially important when processing high-dimensional data, because feature redundancy is prone to occur in high-dimensional space. The sparse regularization term can effectively reduce this redundancy and improve the robustness and accuracy of clustering. The fourth term of the objective function is the local density term, which takes into account the local density of the data points Impact on clustering results. By introducing local density, the system can pay more attention to data points located in high-density areas during the clustering process, ensuring that the clustering results can better reflect the local structural characteristics of the data. Controls the weight of the local density term in the objective function by adjusting , the system can find a balance between global distance and local density, thereby improving the accuracy and robustness of clustering results. is the smoothness constraint of the cluster center. Gradient It reflects the changing trend of cluster centers in the feature space. Minimizing this term can ensure that the cluster centers maintain a certain smoothness in the feature space and avoid drastic changes or unreasonable jumps in the cluster centers, which is very important for ensuring the stability and consistency of clustering results. By introducing smoothness constraints, the system can generate smoother and more reasonable cluster structures, which are suitable for processing complex data distributions in multidimensional space. is the Laplace regularization term, based on the Laplace matrix Constrain the relationship between data points. The Laplacian matrix is ​​usually used to capture the graph structure between data points. By minimizing this, the system can ensure that the relationship between cluster centers and data points conforms to the topological characteristics of the graph structure, thereby improving the rationality of the graph structure of clustering and the global consistency of the results. Control the influence of this term in the objective function by adjusting , the system can find the best balance between the local and global relationships of data points. By combining the above, the objective function It realizes the comprehensive consideration of multi-dimensional factors in the fuzzy C-means clustering process. By minimizing the objective function, the system can find the optimal clustering structure in the complex construction site quality management data, so that the membership degree of each data point, the location of the cluster center and the entire clustering result can accurately reflect the inherent structure and characteristics of the data, thereby providing more scientific and accurate support for construction quality assessment. This multi-dimensional and multi-level objective function design significantly improves the adaptability and accuracy of the fuzzy C-means clustering algorithm, allowing it to maintain efficient clustering effects in complex data environments.

[0077] Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that these specific embodiments are only illustrative, and those skilled in the art may omit, replace, and change the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, merging the above method steps so as to perform substantially the same functions in substantially the same manner to achieve substantially the same results is within the scope of the present invention. Therefore, the scope of the present invention is limited only by the appended claims.

Claims

1. Construction site quality management evaluation system based on fuzzy C-means clustering, It is characterized in that The system includes: a data acquisition and standardization unit, a fuzzy C-means clustering unit and a quality management evaluation unit; the data acquisition and standardization unit is used to acquire quality index data of multiple dimensions of the construction site, take the quality index data of each dimension as an element to form a multidimensional quality index vector, perform nonlinear standardization on the multidimensional quality index vector, and obtain a standardized quality index; the fuzzy C-means clustering unit is used to initialize a fuzzy membership matrix, calculate the weighted cluster center of the standardized quality index; calculate the adaptive distance from each element in the standardized quality index to each weighted cluster center; based on the adaptive distance, consider the local density and update the fuzzy membership matrix; set the objective function of the fuzzy C-means clustering, and cluster the standardized quality index based on the objective function; the quality management evaluation unit is used to calculate the local density consistency index of each element in the standardized quality index in its corresponding cluster based on the clustering result, and then calculate the local density consistency of all elements. The indices are added to obtain the corresponding quality index, and the quality index is compared with the preset quality management level interval to obtain the quality management evaluation result; the quality indicator data of each dimension corresponds to a type of quality indicator data; the quality indicator data of the same dimension includes one or more samples of quality indicator data of this type; the types of quality indicator data include: structural safety indicator data, material quality indicator data, equipment management indicator data, safety management indicator data and progress management indicator data; the structural safety indicator data include: concrete strength and steel bar protection layer thickness; the material quality indicator data include: raw material qualification rate and material pre-use inspection rate; the equipment management indicator data include: equipment integrity rate and equipment regular maintenance rate; the safety management indicator data include: safety accident occurrence rate; the progress management indicator data include: construction period deviation rate, key node completion rate and rework rate; the data acquisition and standardization unit uses the following formula to obtain the first in the multidimensional quality indicator vector The quality index vector of the dimension is nonlinearly normalized to obtain Standardized quality indicators in 4 dimensions: in, For the Standardized quality indicators in 4 dimensions; For the The original quality indicator data of each dimension; For the The average value of the quality indicator data of each dimension; For the The standard deviation of the quality indicator data in each dimension; For the The median of the quality indicator data of each dimension; and Respectively The maximum and minimum values ​​of the quality indicator data in each dimension; For the Adaptive parameters in dimensions; For the The threshold parameters of the dimensions; and All are subscript indexes; It is an absolute value operation; the standardized quality index of each dimension is taken as an element to construct the standardized quality index ; is the total number of dimensions; For the The original quality indicator data of each dimension; For the The average value of the quality indicator data of each dimension; For the The standard deviation of the quality indicator data of the dimension; The adaptive parameters of each dimension are calculated using the following formula: in, For the Dimensional quality indicator data and Covariance of the quality indicator data of the dimensions; is the number of dimensions of the quality indicator data; For the The standard deviation of the quality indicator data of the dimension; Threshold parameters for each dimension Use the following formula to calculate: in, For the The number of samples of quality indicator data in each dimension; For the The quality indicator data of the dimension samples; is the subscript index; the fuzzy C-means clustering unit initializes a fuzzy membership matrix ;in, is the fuzzy membership matrix The Row, No. Elements of a column; is the number of clusters; the weighted cluster center is calculated by the following formula: in, is the subscript index; For the Standardized quality indicators in 4 dimensions; is the mean of the elements of the standardized quality index; is the L2 norm; It is the scale parameter in the weighted cluster center calculation, controlling the width of the Gaussian kernel; For the The weighted cluster centers of the clusters; is the fuzziness parameter, which controls the fuzziness of clustering. The closer it is to 1, the more the clustering tends to be hard partitioned; when When it increases, the clustering becomes more fuzzy; the fuzzy C-means clustering unit calculates the adaptive distance from each element in the standardized quality index to the center of each weighted cluster through the following formula; in, For the Normalized quality index of dimensions to weighted cluster centers Adaptive distance; Parameters that control the weight of the reconstruction error in the adaptive distance; is a nonlinear feature mapping function, Mapping to a higher dimensional feature space; is the dictionary matrix of sparse coding; For the The weighted cluster centers in the dictionary Sparse coding under ; is the parameter that controls the weight of the logistic function term in the adaptive distance; is the steepness parameter of the logistic function; is the transpose operation; is the L1 norm of the matrix; is the preset weight parameter; the fuzzy C-means clustering unit updates the fuzzy membership matrix based on the adaptive distance and considering the local density through the following formula: in, For the The local density of weighted cluster centers; For the Normalized quality index of dimensions to weighted cluster centers distance; is the subscript index; is the updated fuzzy membership matrix; local density Use the following formula to calculate: in, For the Standardized quality indicators in 4 dimensions; Parameters that control the weight of cosine similarity in local density calculation; Representation calculation and The cosine similarity of For the Standardized quality index of dimensions; objective function of fuzzy C-means clustering Use the following formula to express it: in, Operations to find the trace of a matrix; is a parameter that controls the weight of the local density term in the objective function; is a parameter that controls the strength of the cluster center smoothness constraint; express The gradient of To standardize quality indicators The Laplace matrix of ; By minimizing the objective function, the corresponding clustering result is obtained; For the The weighted cluster centers in the dictionary Sparse coding under .