Abnormal data identification method, device and equipment, and storage medium

Abstract

The application discloses an abnormal data identification method and device, equipment and a storage medium. It belongs to the field of data identification. The method comprises the following steps: performing two-class clustering on a plurality of construction data to obtain two data clusters, and obtaining two-class distance sums between each data point in each data cluster and the centroid point of the data cluster; performing one-class clustering to obtain a central centroid, and obtaining one-class distance sums between each construction data and the central centroid; if the two-class distance sum is smaller than the one-class distance sum, taking the cluster with the least data in the two data clusters in the two-class clustering as new construction data; repeatedly performing the two-class clustering and the one-class clustering until the cluster with the least data in the two data clusters has only single construction data; and identifying the single construction data as abnormal data. The application can effectively identify abnormal data.

Description

Technical Field

[0001] This application relates to the field of data identification, and more particularly to an abnormal data identification method, apparatus, device, and storage medium. Background Technology

[0002] In the construction process, many factors affect the identification of abnormal carbon emission data, and common prediction algorithms cannot support the needs of identifying abnormal carbon emission data.

[0003] Application content

[0004] The main objective of this application is to provide an abnormal data identification method, apparatus, device, and storage medium, aiming to solve the technical problem that common prediction algorithms cannot support the needs of identifying abnormal carbon emission data due to the many factors affecting the identification of abnormal carbon emission data.

[0005] Firstly, in order to achieve the above objectives, this application provides an anomaly data identification method, the method comprising:

[0006] Obtain carbon emission data from multiple building construction projects;

[0007] Binary clustering is performed on multiple carbon emission data of building construction to obtain two carbon emission data clusters, and the sum of binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster is obtained.

[0008] A single-class clustering process is performed on multiple building construction carbon emission data to obtain centroids, and the sum of single-class distances between each building construction carbon emission data and the centroids is obtained.

[0009] If the sum of the binary distances is less than the sum of the unary distances, then the cluster with the fewest data points in the two carbon emission data clusters is taken as the new construction carbon emission data. The process of performing binary clustering on the multiple new construction carbon emission data is then repeated to obtain two carbon emission data clusters. The sum of the binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster is obtained. Unary clustering is then performed on the multiple new construction carbon emission data to obtain the centroid. The sum of the unary distances between the multiple new construction carbon emission data and the centroid is obtained. This process continues until the cluster with the fewest data points in the two carbon emission data clusters contains only a single construction carbon emission data point.

[0010] The carbon emission data of the individual building construction is identified as the abnormal carbon emission data.

[0011] Optionally, the binary clustering processing of the multiple building construction carbon emission data includes:

[0012] Determine the average value point of multiple building construction carbon emission data;

[0013] A data point is determined on each side of the average value point to obtain the initial centroid of the cluster;

[0014] The carbon emission data of building construction is clustered based on the initial centroids of the clusters.

[0015] Optionally, the step of clustering the building construction carbon emission data based on the initial centroids includes:

[0016] Clustering the carbon emission data from the building construction yielded two initial carbon emission data clusters.

[0017] Determine whether the centroids of the two initial carbon emission data clusters are further away from the average value point relative to the initial centroid of the cluster;

[0018] If at least one of the two centroids is farther away from the mean point relative to the initial centroid of the cluster, then two new data points are randomly selected as the initial centroids of the cluster, and the clustering process on the carbon emission data of the building construction is returned to obtain two initial carbon emission data clusters; it is determined whether the centroids of the two initial carbon emission data clusters are farther away from the mean point relative to the initial centroid of the cluster, until the clustering iteration is completed.

[0019] Optionally, determining a data point on each side of the average point to obtain the initial centroid of the cluster includes:

[0020] Randomly select a data point within a preset range on both sides of the average value point to obtain the initial centroid of the cluster.

[0021] Optionally, the binary classification distance sum is the sum of the squares of the distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster; and / or

[0022] The first classification distance sum is the sum of the squares of the distances between each data point in the carbon emission data cluster and the centroid of the carbon emission data cluster.

[0023] Optionally, after identifying the individual building construction carbon emission data as the abnormal carbon emission data, the method further includes:

[0024] Based on the characteristics of the various construction carbon emission data, the various construction carbon emission data are classified to obtain construction site-related carbon emission data, vehicle-related carbon emission data, and material-related carbon emission data.

[0025] From the carbon emission data related to the construction site, the carbon emission data related to the vehicle, and the carbon emission data related to the material, delete the abnormal carbon emission data of each type of carbon emission data to obtain the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, and the carbon emission data related to the target material.

[0026] Determine the carbon emission protocols corresponding to the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, and the carbon emission data related to the target material;

[0027] Carbon emission data is calculated based on the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, the carbon emission data related to the target material, and their respective carbon emission agreements.

[0028] Optionally, the plurality of building construction carbon emission data includes at least one of building material carbon emission data, water consumption carbon emission data, and electricity consumption carbon emission data.

[0029] Secondly, this application also provides an abnormal data identification device, the device comprising:

[0030] The acquisition module is used to acquire carbon emission data from multiple building construction projects.

[0031] The first processing module is used to perform binary clustering on the multiple carbon emission data of the building construction to obtain two carbon emission data clusters, and to obtain the sum of binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster.

[0032] The second processing module is used to perform single-class clustering on multiple building construction carbon emission data to obtain centroids, and to obtain the sum of single-class distances between each building construction carbon emission data and the centroids.

[0033] The loop module is used to, if the sum of the binary distances is less than the sum of the unary distances, select the cluster with the fewest data points among the two carbon emission data clusters as the new construction carbon emission data, and return to perform binary clustering on the multiple new construction carbon emission data to obtain two carbon emission data clusters, and obtain the sum of the binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster; perform unary clustering on the multiple new construction carbon emission data to obtain the centroid, and obtain the sum of the unary distances between the multiple new construction carbon emission data and the centroid, until the cluster with the fewest data points among the two carbon emission data clusters contains only a single construction carbon emission data point;

[0034] The identification module is used to identify the carbon emission data of a single building construction as the abnormal carbon emission data.

[0035] Thirdly, this application also provides an abnormal data identification device, including: a processor, a memory, and an abnormal data identification program stored in the memory, wherein the abnormal data identification program is executed by the processor to implement the steps of the abnormal data identification method as described in the first aspect.

[0036] Fourthly, this application also provides a computer-readable storage medium storing an abnormal data identification program, which, when executed by a processor, implements the abnormal data identification method as described in the first aspect.

[0037] Because many factors influence the identification of abnormal carbon emission data during construction in the building industry, common prediction algorithms cannot support the requirements for identifying abnormal carbon emission data. However, the abnormal data identification method proposed in this application performs single-class clustering and binary clustering on multiple building construction carbon emission data. Since abnormal data will affect the clustering results, the clusters containing abnormal data can be identified by comparing the single-class clustering and binary clustering. Then, the above processing is continued on the clusters containing abnormal data until the abnormal carbon emission data is classified from the multiple building construction carbon emission data, thereby effectively identifying abnormal carbon emission data. Attached Figure Description

[0038] Figure 1 This is a structural diagram of an abnormal data identification device for the hardware operating environment involved in the embodiments of this application;

[0039] Figure 2 This is a flowchart illustrating the first embodiment of the abnormal data identification method of this application;

[0040] Figure 3 This is a flowchart illustrating the second embodiment of the abnormal data identification method of this application;

[0041] Figure 4 This is a schematic diagram of the structure of an abnormal data identification device provided in this application.

[0042] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0043] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0044] Because many factors influence the identification of abnormal carbon emission data during the construction process in the construction industry, common prediction algorithms cannot support the needs of identifying abnormal carbon emission data.

[0045] This application provides a solution because many factors affect the identification of abnormal carbon emission data during the construction process in the building industry. Common prediction algorithms cannot support the needs of identifying abnormal carbon emission data. However, the abnormal data identification method proposed in this application involves performing single-class clustering and binary clustering on multiple building construction carbon emission data. Since abnormal data will affect the clustering results, clusters with abnormal data can be identified by comparing the single-class clustering and binary clustering. Then, the above processing is continued on the clusters with abnormal data until the abnormal carbon emission data is classified from the multiple building construction carbon emission data, thereby effectively identifying abnormal carbon emission data.

[0046] refer to Figure 1 , Figure 1 This is a structural diagram of an abnormal data identification device for a hardware operating environment according to an embodiment of this application. The abnormal data identification device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk storage device. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0047] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the abnormal data identification device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0048] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and an abnormal data identification program.

[0049] exist Figure 1In the abnormal data identification device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the abnormal data identification device of this application can be set in the abnormal data identification device. The abnormal data identification device calls the abnormal data identification program stored in the memory 1005 through the processor 1001 and executes the abnormal data identification method provided in the embodiment of this application.

[0050] Based on, but not limited to, the hardware structure of the above-described abnormal data identification device, this application provides a first embodiment of an abnormal data identification method. (Refer to...) Figure 2 , Figure 2 A flowchart illustrating a first embodiment of the abnormal data identification method of this application is shown.

[0051] It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0052] In this embodiment, the method includes:

[0053] Step 10: Obtain carbon emission data from multiple building construction projects.

[0054] It should be understood that the execution subject in this embodiment is the abnormal data identification device.

[0055] First, during the construction process, basic information about the construction site project is configured for two scenarios: office and site construction. This basic information includes personnel information, vehicle type information, number of vehicles, site construction area information, building type information, estimated carbon emission agreement name information, and estimated carbon emission agreement information.

[0056] Secondly, after the basic information of the construction site project is configured, the carbon emission data required for carbon emission protocol calculation is retrieved from the project management system based on the estimated carbon emission protocol. This carbon emission data includes carbon emission data for building materials, fossil fuels, water consumption, and electricity consumption. Specifically, the carbon emission data for building materials can be at least one of the following: carbon emission data for reinforcing steel, concrete, and I-beams. This carbon emission data for building materials can be the amount of building materials used per unit area after dividing the total amount of building materials used by the building area.

[0057] Finally, after retrieving the construction carbon emission data required for carbon emission protocol calculation from the engineering management system based on the estimated carbon emission protocol, the anomaly data identification device performs data verification, unit verification, and data missingness verification on multiple construction carbon emission data sets, which helps improve the validity of the construction carbon emission data. Specifically, if the construction carbon emission data is at least one of water consumption carbon emission data or electricity consumption carbon emission data, then data verification, unit verification, and data missingness verification need to be performed on either the water consumption carbon emission data or the electricity consumption carbon emission data separately.

[0058] Step 20: Perform binary clustering on the multiple carbon emission data of the building construction to obtain two carbon emission data clusters, and obtain the sum of binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster.

[0059] It's important to understand that binary clustering divides multiple construction carbon emission data sets into two clusters. Construction carbon emission data within the same cluster exhibit similarities, while data in different clusters show differences. In other words, after binary clustering, data with high similarity are grouped together in the same cluster, while data with low similarity are separated as much as possible.

[0060] Binary clustering was performed on multiple building construction carbon emission data sets, resulting in two carbon emission data clusters and two centroids for each cluster. The two carbon emission data clusters consist of a first carbon emission data cluster and a second carbon emission data cluster. The two centroids consist of a first centroid and a second centroid. The first centroid is located within the first carbon emission data cluster; the second centroid is located within the second carbon emission data cluster.

[0061] The distances between each data point in the first carbon emission data cluster and the first centroid are summed to obtain the first distance sum. The distances between each data point in the second carbon emission data cluster and the second centroid are summed to obtain the second distance sum. By summing the first and second distance sums, the binary classification distance sum is obtained.

[0062] Step 30: Perform single-class clustering on the multiple building construction carbon emission data to obtain centroids, and obtain the single-class distance between each building construction carbon emission data and the centroids.

[0063] It's important to understand that single-class clustering involves iteratively calculating the centroids of multiple building construction carbon emission data points. After obtaining the centroids through iterative calculation, the sum of single-class distances between each building construction carbon emission data point and the centroid can be obtained.

[0064] Furthermore, as an embodiment, the binary classification distance sum is the sum of the squares of the distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster; and / or the unclassification distance sum is the sum of the squares of the distances between each data point in the carbon emission data cluster and the centroid of the cluster.

[0065] Understandably, the distance between each data point in the calculated carbon emission data cluster and the centroid of the carbon emission data cluster is calculated using the sum of squares, rather than the direct distance, which helps to highlight abnormal carbon emission data.

[0066] Step 40: If the sum of the binary classification distances is less than the sum of the unclassification distances, then the cluster with the fewest data among the two carbon emission data clusters is taken as the building construction carbon emission data, and the process returns to execute steps S20 and 30 until the cluster with the fewest data among the two carbon emission data clusters contains only a single building construction carbon emission data.

[0067] Step 50: Identify the individual building construction carbon emission data as the abnormal carbon emission data.

[0068] It's important to understand that outlier carbon emission data refers to carbon emission data that deviates from the average value by more than twice among multiple construction carbon emission data sets. Therefore, theoretically, in the absence of outliers, the sum of the primary distances should be less than the sum of the secondary distances. Conversely, if the sum of the secondary distances is less than the sum of the primary distances, indicating the presence of outlier carbon emission data among multiple construction carbon emission data sets, then since the outlier data deviates from the centroid, it will reside in the cluster with the fewest data points among the two carbon emission data clusters. Therefore, the cluster with the fewest data points among the two carbon emission data clusters will be considered the most recent construction carbon emission data.

[0069] Then repeat steps S20 and S30, that is, repeatedly perform binary clustering and single clustering on the data clusters with abnormal data to continuously narrow down the range of data clusters with abnormal data until the cluster with the fewest data in the two carbon emission data clusters has only a single building construction carbon emission data, that is, lock the abnormal data.

[0070] In this embodiment, since many factors affect the identification of abnormal carbon emission data during the construction process in the construction industry, common prediction algorithms cannot support the needs of identifying abnormal carbon emission data. However, the abnormal data identification method proposed in this application uses the k-means algorithm to perform single-class clustering and binary clustering on multiple construction carbon emission data. Since abnormal data will affect the clustering results, the clusters with abnormal data can be identified by comparing the single-class clustering and binary clustering. Then, the above processing is continued on the clusters with abnormal data until the abnormal carbon emission data is classified from the multiple construction carbon emission data, thereby effectively identifying abnormal carbon emission data.

[0071] Furthermore, as one implementation method, step 20 specifically includes:

[0072] Step 21: Determine the average value of multiple carbon emission data points from building construction.

[0073] Step 22: Determine a data point on each side of the average value point to obtain the initial centroid of the cluster.

[0074] Step 23: Perform clustering processing on the building construction carbon emission data based on the initial centroid of the cluster.

[0075] It's important to understand that the average point is the location of the average value of multiple construction carbon emission data points in the data space. The initial centroid of the cluster is a data point determined on both sides of the average point before the first iteration of the multiple construction carbon emission data. After the initial centroids are determined, the construction carbon emission data is clustered, which is essentially updating and iterating the initial centroids.

[0076] In this embodiment, a data point is determined on both sides of the average value point to obtain the initial centroid of the cluster, which can optimize the selection of the initial centroid of the cluster.

[0077] Furthermore, as one implementation method, step 23 specifically includes:

[0078] Step 231: Cluster the carbon emission data of the building construction to obtain two initial carbon emission data clusters.

[0079] Step 232: Determine whether the centroids of the two initial carbon emission data clusters are further away from the average value point relative to the initial centroid of the cluster.

[0080] Step 233: If at least one of the two centroids is farther away from the mean point than the initial centroid of the cluster, then two new data points are randomly selected as the initial centroids of the cluster, and steps S231 and S232 are returned to be executed until the clustering iteration is completed.

[0081] It is important to understand that, under normal circumstances, after clustering multiple construction carbon emission data, the centroids of the two initial carbon emission data clusters gradually move closer to the average value relative to the initial cluster centroid. If at least one of the centroids of the two initial carbon emission data clusters is far from the average value relative to the initial cluster centroid, or both are far away, it indicates that there is abnormal carbon emission data in the multiple construction carbon emission data.

[0082] For example, if one of the centroids of two initial carbon emission data clusters is far from the average point relative to the initial centroid of the cluster (i.e., far away on one side), it indicates that there is at least one anomalous carbon emission data point among the multiple construction carbon emission data points. If both centroids of two initial carbon emission data clusters are far from the average point relative to the initial centroid of the cluster (i.e., far away on both sides), it indicates that there are at least two anomalous carbon emission data points among the multiple construction carbon emission data points.

[0083] In this embodiment, if at least one centroid point in the two initial data clusters moves away from one side or both sides, two new data points are randomly selected as the initial centroids for clustering. This avoids getting trapped in local optima and helps to optimize the selection of the initial centroids for clustering.

[0084] Furthermore, as one implementation method, step 22 specifically includes:

[0085] Step 221: Randomly select a data point within a preset range on both sides of the average value point to obtain the initial centroid of the cluster.

[0086] It is important to understand that the preset range is a range that the user has set in advance.

[0087] For example, when the carbon emission data of building construction is electricity cost, the average point is 100 degrees, and a data point is randomly selected in a preset range of 90 degrees to 100 degrees with a value step of 10 degrees, and a data point is randomly selected in the range of 100 degrees to 110 degrees.

[0088] In this embodiment, a data point is randomly selected within a preset range on both sides of the average value point to obtain the initial centroid of the cluster, which is beneficial for the selection of the initial centroid of the cluster.

[0089] This application provides a second embodiment of an abnormal data identification method. (Refer to...) Figure 3 , Figure 3 A flowchart illustrating a second embodiment of the abnormal data identification method of this application is shown.

[0090] In this embodiment, after step S50, the method further includes:

[0091] Step S51: Based on the characteristics of the multiple construction carbon emission data, classify the multiple construction carbon emission data to obtain site-related carbon emission data, vehicle-related carbon emission data, and material-related carbon emission data.

[0092] Step S52: Delete the abnormal carbon emission data from the construction site-related carbon emission data, the vehicle-related carbon emission data, and the material-related carbon emission data to obtain the target construction site-related carbon emission data, the target vehicle-related carbon emission data, and the target material-related carbon emission data.

[0093] Step S53: Determine the carbon emission protocols corresponding to the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, and the carbon emission data related to the target material.

[0094] Step S54: Calculate the carbon emission data based on the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, the carbon emission data related to the target material, and their respective carbon emission agreements.

[0095] It is important to understand that the characteristics of carbon emission data in construction can be the usage characteristics of the materials themselves during the construction period and the main characteristics of the materials themselves. Based on the characteristics of carbon emission data in construction and three scenarios (construction site related scenarios, vehicle related scenarios, and material related scenarios), the characteristics of multiple carbon emission data in construction are classified to obtain construction site related carbon emission data, vehicle related carbon emission data, and material related carbon emission data.

[0096] Carbon emissions data related to construction sites can include the amount of paper used on the site. Since construction sites do not vary much, the average amount of paper used can be used to estimate the carbon emissions data for paper usage on the site.

[0097] Vehicle-related carbon emission data can be the carbon emission data of vehicle gasoline. Since gasoline consumption is related to vehicle data, we can first calculate the average monthly gasoline consumption of each vehicle, and then multiply the vehicle data by the number of construction months to estimate the carbon emission data of gasoline consumed by the vehicle during the construction period.

[0098] Material-related carbon emission data can include carbon emission data for steel bars and concrete.

[0099] Based on relevant construction experience and combined with data closely related to carbon emission data of materials, such as construction site area, construction building area, number of above-ground floors, number of underground floors, total green area, CFA area (building area), plot ratio, and building height, a KNN dataset is constructed. The three most similar data can be selected for classification and prediction processing.

[0100] Furthermore, in practical work, some improvements were made to the KNN algorithm:

[0101] Improvements to the importance of material-related carbon emission data dimensions. For example, concrete carbon emission data is related to the construction area, building height, number of above-ground floors, and number of underground floors. Therefore, when training KNN on concrete carbon emission data (which includes both important and unimportant data), the coefficient for important data is R, while the coefficient for unimportant data is 1, effectively distinguishing the dimensions of concrete carbon emission data.

[0102] Improvements have been made to the final carbon emission data processing for the classification and prediction stage. Due to the numerous unique scenarios on the site, the final carbon emission data processing method has been changed from calculating the reciprocal values ​​and then summing them using a weighted average. It's important to note that a portion of the carbon emission data after classification and prediction is assembled using the MIC module; this part of the carbon emission data is provided by the manufacturer and can be calculated directly.

[0103] Identify the corresponding carbon emission protocols for the target construction site, target vehicles, and target materials. Once the protocols are defined, use Natural Language Processing (NLP) to determine if there are any anomalies in the correspondence between the types of building materials and the scope of the protocols.

[0104] Furthermore, as one implementation method, it is necessary to check for any anomalies in the correspondence between the types of building materials and the scope of the agreement. This is because the description of the scope of the agreement may include some common building materials within the statistical range, but it is impossible to include all of them.

[0105] Because there are many names for the same building material, and the system distinguishes between simplified and traditional Chinese characters, the system currently relies mainly on manual maintenance by business personnel. However, errors still exist, necessitating verification. Therefore, the overall approach to verifying the same building material is as follows: first, convert both simplified and traditional Chinese characters to simplified Chinese; then, use NLP word segmentation. Unlike the usual method of calculating similarity distance using segmented words, this method uses pinyin, combining a pinyin syllable with its tone to form a single character for distance calculation. The character with the smallest distance corresponds to its protocol range.

[0106] Carbon emission data is calculated based on carbon emission data related to the target construction site, target vehicles, and target materials, along with their respective carbon emission agreements. The agreements require the automatic counting of the types of building materials involved in each scope scenario, multiplying the quantity of building materials by the carbon neutrality factor, and summing the results for statistical calculation.

[0107] Based on the same concept, this application also provides an abnormal data identification device. (See reference...) Figure 4 , Figure 4 This application provides a schematic diagram of the structure of an abnormal data identification device, which specifically includes:

[0108] Module 400 is used to acquire carbon emission data from multiple building construction projects.

[0109] The first processing module 410 is used to perform binary clustering processing on the single building construction carbon emission data identified as abnormal carbon emission data to obtain two carbon emission data clusters, and to obtain the sum of binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster.

[0110] The second processing module 420 is used to perform single-class clustering on multiple building construction carbon emission data to obtain centroids and to obtain the sum of single-class distances between each building construction carbon emission data and the centroids.

[0111] The loop module 430 is configured to, if the sum of the binary classification distances is less than the sum of the unclassification distances, take the cluster with the fewest data points among the two carbon emission data clusters as new construction carbon emission data, return to perform binary clustering on the multiple new construction carbon emission data to obtain two carbon emission data clusters, and obtain the sum of the binary classification distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster; perform unclassification clustering on the multiple new construction carbon emission data to obtain the centroid, and obtain the sum of the unclassification distances between the multiple new construction carbon emission data and the centroid, until the cluster with the fewest data points among the two carbon emission data clusters contains only a single construction carbon emission data point;

[0112] The identification module 440 is used to identify the single building construction carbon emission data as the abnormal carbon emission data.

[0113] The technical solution of this embodiment, through the cooperation between various functional modules, addresses the issue that many factors affect the identification of abnormal carbon emission data during the construction process in the building industry. Common prediction algorithms cannot support the requirements for identifying abnormal carbon emission data. However, the abnormal data identification method proposed in this application involves performing single-class clustering and binary clustering on multiple building construction carbon emission data. Since abnormal data affects the clustering results, clusters containing abnormal data can be identified by comparing the single-class clustering and binary clustering processes. Then, the above processing is continued on the clusters containing abnormal data until the abnormal carbon emission data is classified from the multiple building construction carbon emission data, thereby effectively identifying abnormal carbon emission data.

[0114] Furthermore, embodiments of this application also propose a computer storage medium storing an abnormal data identification program. When executed by a processor, the abnormal data identification program implements the steps of the abnormal data identification method described above. Therefore, it will not be repeated here. Additionally, the beneficial effects of using the same method will not be repeated. For technical details not disclosed in the computer-readable storage medium embodiments of this application, please refer to the description of the method embodiments of this application. As an example, program instructions can be deployed to execute on a single computing device, or on multiple computing devices located at one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0115] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0116] It should also be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided in this application, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0117] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0118] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for identifying abnormal data, characterized in that, The method includes: Acquire multiple carbon emission data of building construction, wherein the carbon emission data of building construction includes at least one of carbon emission data of building materials, carbon emission data of water consumption, and carbon emission data of electricity consumption; Binary clustering is performed on multiple carbon emission data of building construction to obtain two carbon emission data clusters, and the sum of binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster is obtained. A single-class clustering process is performed on multiple building construction carbon emission data to obtain centroids, and the sum of single-class distances between each building construction carbon emission data and the centroids is obtained. If the sum of the binary distances is less than the sum of the unary distances, then the cluster with the fewest data points in the two carbon emission data clusters is taken as the new construction carbon emission data. The process of performing binary clustering on the multiple new construction carbon emission data is then repeated to obtain two carbon emission data clusters. The sum of the binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster is obtained. Unary clustering is then performed on the multiple new construction carbon emission data to obtain the centroid. The sum of the unary distances between the multiple new construction carbon emission data and the centroid is obtained. This process continues until the cluster with the fewest data points in the two carbon emission data clusters contains only a single construction carbon emission data point. The carbon emission data of the individual building construction is identified as abnormal carbon emission data; The binary clustering process for multiple building construction carbon emission data includes: Determine the average value point of multiple building construction carbon emission data; A data point is determined on each side of the average value point to obtain the initial centroid of the cluster; The carbon emission data of building construction is clustered based on the initial centroids of the clusters.

2. The abnormal data identification method according to claim 1, characterized in that, The clustering process for the building construction carbon emission data based on the initial centroids includes: Clustering the carbon emission data from the building construction yielded two initial carbon emission data clusters. Determine whether the centroids of the two initial carbon emission data clusters are further away from the average value point relative to the initial centroid of the cluster; If at least one of the two centroids is farther away from the mean point relative to the initial centroid of the cluster, then two new data points are randomly selected as the initial centroids of the cluster, and the clustering process on the carbon emission data of the building construction is returned to obtain two initial carbon emission data clusters; it is determined whether the centroids of the two initial carbon emission data clusters are farther away from the mean point relative to the initial centroid of the cluster, until the clustering iteration is completed.

3. The abnormal data identification method according to claim 1, characterized in that, The step of determining a data point on each side of the average point to obtain the initial centroid of the cluster includes: Randomly select a data point within a preset range on both sides of the average value point to obtain the initial centroid of the cluster.

4. The abnormal data identification method according to claim 1, characterized in that, The binary classification distance sum is the sum of the squares of the distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster; and / or The first classification distance sum is the sum of the squares of the distances between each data point in the carbon emission data cluster and the centroid of the carbon emission data cluster.

5. The abnormal data identification method according to claim 1, characterized in that, After identifying the single building construction carbon emission data as the abnormal carbon emission data, the method further includes: Based on the characteristics of the various construction carbon emission data, the various construction carbon emission data are classified to obtain construction site-related carbon emission data, vehicle-related carbon emission data, and material-related carbon emission data. From the carbon emission data related to the construction site, the carbon emission data related to the vehicle, and the carbon emission data related to the material, delete the abnormal carbon emission data of each type of carbon emission data to obtain the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, and the carbon emission data related to the target material. Determine the carbon emission protocols corresponding to the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, and the carbon emission data related to the target material; Carbon emission data is calculated based on the carbon emission data related to the target construction site, the carbon emission data related to the target vehicle, the carbon emission data related to the target material, and their respective carbon emission agreements.

6. An abnormal data identification device, characterized in that, The device includes: The acquisition module is used to acquire multiple carbon emission data of building construction, which includes at least one of carbon emission data of building materials, carbon emission data of water consumption, and carbon emission data of electricity consumption. The first processing module is used to perform binary clustering on the multiple construction carbon emission data to obtain two carbon emission data clusters, and to obtain the sum of binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster. The first processing module is also used to determine the average value point of the multiple construction carbon emission data; determine a data point on each side of the average value point to obtain the initial centroid of the cluster; and perform clustering on the construction carbon emission data according to the initial centroid of the cluster. The second processing module is used to perform single-class clustering on multiple building construction carbon emission data to obtain centroids, and to obtain the sum of single-class distances between each building construction carbon emission data and the centroids. The loop module is used to, if the sum of the binary distances is less than the sum of the unary distances, select the cluster with the fewest data points among the two carbon emission data clusters as the new construction carbon emission data, and return to perform binary clustering on the multiple new construction carbon emission data to obtain two carbon emission data clusters, and obtain the sum of the binary distances between each data point in each carbon emission data cluster and the centroid of the carbon emission data cluster; perform unary clustering on the multiple new construction carbon emission data to obtain the centroid, and obtain the sum of the unary distances between the multiple new construction carbon emission data and the centroid, until the cluster with the fewest data points among the two carbon emission data clusters contains only a single construction carbon emission data point; The identification module is used to identify the carbon emission data of a single building construction as abnormal carbon emission data.

7. An abnormal data identification device, characterized in that, include: A processor, a memory, and an anomaly data identification program stored in the memory, wherein the anomaly data identification program is executed by the processor to implement the steps of the anomaly data identification method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an abnormal data identification program, which, when executed by a processor, implements the abnormal data identification method as described in any one of claims 1 to 5.

Images