Method for realizing quantitative risk management of green building technology based on supervised learning

Abstract

The invention relates to the technical field of risk management, and provides a method for realizing green building technical risk quantitative management based on supervised learning, and the method comprises the steps: recognizing construction elements in a construction region, querying construction rules corresponding to the construction elements, and collecting factor data corresponding to risk factors in the construction region; performing quantitative processing on the factor data to obtain factor quantitative values, calculating factor variation values corresponding to the risk factors, and calculating risk probabilities corresponding to the risk factors; calculating factor sensitivity among the key risk factors, calculating factor risk values among the key risk factors, and setting risk levels corresponding to the key risk factors; and calculating a risk impact effect of the key risk factor on the construction task, collecting element cost data corresponding to the construction element, performing risk quantification processing on the key risk factor to obtain a quantification risk factor, and generating a risk quantification scheme of the construction area. The invention aims to improve the efficiency of quantitative management of green building technical risks.

Description

Technical Field

[0001] The present invention relates to the technical field of risk management, and in particular to a method for implementing green building technology risk quantification management based on supervised learning. Background Art

[0002] Green building technology refers to the construction technology that maximizes resource conservation (energy saving, land saving, water saving, material saving), protects the environment and reduces pollution throughout the entire life cycle of the building, provides people with healthy, applicable and efficient use space, and coexists harmoniously with nature. With the continuous improvement of environmental awareness and the requirement of sustainable development, the application of green building technology in the construction industry is becoming more and more extensive. However, the application of green building technology also brings some risks, such as technical risks, economic risks, environmental risks, etc. In order to effectively manage these risks, it is necessary to quantify the risks of green building technology. Supervised learning is to build a model by learning a set of training data with known inputs (features) and corresponding known outputs (target values). It can make accurate predictions on new input data, and then learn the complex relationship between risks.

[0003] However, most of the existing green building technology risk quantification management methods are based on the establishment of a risk indicator system, quantifying the probability and loss of risk factors through questionnaires, and then applying various models to obtain the final risk level. Traditional research methods regard risk factors as independent individuals, ignoring the role of mutual influence between risk factors, thereby reducing the accuracy of risk quantification and making risk quantification management inefficient. Summary of the invention

[0004] The present invention provides a method for realizing green building technology risk quantification management based on supervised learning, and its main purpose is to improve the efficiency of green building technology risk quantification management.

[0005] To achieve the above purpose, the present invention provides a method for quantitative management of green building technology risks based on supervised learning, comprising:

[0006] Acquire a construction area corresponding to a green building, identify construction elements within the construction area, query construction rules corresponding to the construction elements, determine risk factors corresponding to the construction elements in combination with the construction rules, and collect factor data corresponding to the risk factors within the construction area;

[0007] Quantitatively process the factor data to obtain factor quantitative values, calculate the factor variation values ​​corresponding to the risk factors according to the factor quantitative values, and calculate the risk probability corresponding to the risk factors according to a preset risk database, and screen out key risk factors from the risk factors by combining the factor variation values ​​and the risk probability;

[0008] Calculating the factor sensitivity between the key risk factors, calculating the factor risk value between the key risk factors according to the factor sensitivity, and setting the risk level corresponding to the key risk factor according to the factor risk value;

[0009] Query the construction tasks corresponding to the construction area, calculate the risk impact effects of the key risk factors on the construction tasks, collect the element cost data corresponding to the construction elements, combine the element cost data and the risk impact effects, use a pre-trained supervised learning regression model to quantify the risks of the key risk factors, obtain quantitative risk factors, combine the quantitative risk factors and the risk level, and generate a risk quantification plan for the construction area.

[0010] Optionally, the determining the risk factors corresponding to the construction elements in combination with the construction rules includes:

[0011] Identifying regular character information in the construction rules, performing word segmentation processing on the regular character information to obtain regular character word segmentations;

[0012] Calculate the importance of the word segmentation corresponding to the regular character word segmentation;

[0013] Extracting key rule segmentations in the rule character segmentations according to the segmentation importance;

[0014] Match the risk factors corresponding to the key rule segmentation in the preset risk set table to obtain the rule risk factors;

[0015] According to the rule risk factors, the risk factors corresponding to the construction elements are obtained.

[0016] Optionally, the calculating the word segmentation importance corresponding to the regular character word segmentation includes:

[0017] Performing semantic analysis on the regular characters to obtain regular semantics;

[0018] Extracting semantic elements corresponding to the semantics of the rule, and counting element frequencies corresponding to the semantic elements;

[0019] In combination with the element frequency, a semantic entropy value corresponding to the semantics of the rule is calculated;

[0020] The word segmentation importance corresponding to the regular character word segmentation is determined according to the semantic entropy value.

[0021] Optionally, the quantitative processing of the factor data to obtain the factor quantitative value includes:

[0022] Performing data cleaning on the factor data to obtain target factor data;

[0023] Performing standardization processing on the target factor data to obtain standard factor data;

[0024] Encoding the standard factor data to obtain factor encoding values;

[0025] The factor coding value is quantitatively converted to obtain the factor quantitative value.

[0026] Optionally, calculating the factor variation value corresponding to the risk factor according to the factor quantitative value includes:

[0027] Counting the number of samples corresponding to the quantitative values ​​of the factors;

[0028] Calculate the quantitative mean corresponding to the risk factor by combining the quantitative value of the factor and the sample size;

[0029] Combining the quantitative mean, the sample size and the quantitative value of the factor, the factor variation value corresponding to the risk factor is calculated by the following formula:

[0030]

[0031] Among them, A represents the factor variation value corresponding to the risk factor, B a Indicates the ath quantitative value in the quantitative value of the factor, represents the quantitative mean, a represents the serial number corresponding to the quantitative value of the factor, and q represents the number of samples.

[0032] Optionally, the calculating the risk probability corresponding to the risk factor according to a preset risk database includes:

[0033] Extracting risk data corresponding to the risk factor from a preset risk database;

[0034] Filter out risk loss data and risk frequency data from the risk data;

[0035] Calculating the risk probability corresponding to the risk factor according to the risk frequency data;

[0036] Calculating the risk severity corresponding to the risk factor according to the risk loss data;

[0037] The risk probability corresponding to the risk factor is calculated by combining the risk severity and the risk probability.

[0038] Optionally, the calculating the factor sensitivity between the key risk factors includes:

[0039] Extracting factor attributes corresponding to the key risk factors, and calculating attribute priorities corresponding to the factor attributes;

[0040] According to the attribute priority, the factor attributes are subjected to attribute screening processing to obtain target factor attributes;

[0041] Calculating the attribute association strength between the target factor attributes;

[0042] The factor sensitivity between the key risk factors is calculated according to the attribute association strength.

[0043] Optionally, the calculating the attribute priority corresponding to the factor attribute includes:

[0044] Performing vector conversion on the factor attributes to obtain a factor vector;

[0045] Performing dimensionality reduction processing on the factor vector to obtain a dimensionality reduction factor vector;

[0046] Extracting a dimension reduction feature vector from the dimension reduction factor vector, and calculating a vector eigenvalue corresponding to the feature vector;

[0047] According to the vector eigenvalue, the attribute priority corresponding to the factor attribute is calculated by the following formula:

[0048]

[0049] Among them, D represents the attribute priority corresponding to the factor attribute, F j represents the jth eigenvalue in the vector eigenvalue, j represents the sequence number of the vector eigenvalue, and r represents the number of vector eigenvalues.

[0050] Optionally, the calculating the risk impact effect of the key risk factor on the construction task includes:

[0051] Decomposing the construction task to obtain construction subtasks;

[0052] Analyze the risk correlation between the key risk factors and the construction subtasks;

[0053] According to the risk correlation, construct a risk correlation matrix between the key risk factors and the construction subtasks;

[0054] The risk impact effect of the key risk factors on the construction task is calculated according to the risk association matrix.

[0055] A system for implementing green building technology risk quantification management based on supervised learning, characterized in that the system comprises:

[0056] A data collection module is used to obtain a construction area corresponding to a green building, identify construction elements in the construction area, query construction rules corresponding to the construction elements, determine risk factors corresponding to the construction elements in combination with the construction rules, and collect factor data corresponding to the risk factors in the construction area;

[0057] A key risk factor screening module is used to perform quantitative processing on the factor data to obtain the factor quantitative value, calculate the factor variation value corresponding to the risk factor according to the factor quantitative value, and calculate the risk probability corresponding to the risk factor according to a preset risk database, and screen out key risk factors from the risk factors by combining the factor variation value and the risk probability;

[0058] A risk level setting module, used to calculate the factor sensitivity between the key risk factors, calculate the factor risk value between the key risk factors according to the factor sensitivity, and set the risk level corresponding to the key risk factor according to the factor risk value;

[0059] The risk quantification module is used to query the construction tasks corresponding to the construction area, calculate the risk impact effects of the key risk factors on the construction tasks, collect the element cost data corresponding to the construction elements, combine the element cost data and the risk impact effects, and use a pre-trained supervised learning regression model to perform risk quantification processing on the key risk factors to obtain quantitative risk factors, and combine the quantitative risk factors and the risk level to generate a risk quantification plan for the construction area.

[0060] The present invention determines the risk factors corresponding to the construction elements by combining the construction rules, and can discover the risk factors existing in the construction area, so as to facilitate the subsequent effective risk management of the construction area. The present invention can better understand the information expressed by the data by quantitatively processing the factor data, and provide a basis for the calculation of subsequent factor variation values. The present invention calculates the factor sensitivity between the key risk factors, and can obtain the degree of mutual influence between the key risk factors, and then understand the mutual correlation and dependency between different factors in the key risk factors. The present invention calculates the risk impact effect of the key risk factors on the construction task, and can obtain the degree of influence caused by the key risk factors, and then provide a basis for subsequent risk quantification processing to improve the management efficiency of air volume quantification. Therefore, the method for implementing green building technology risk quantification management based on supervised learning provided by the embodiment of the present invention can improve the efficiency of green building technology risk quantification management. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1A flowchart of a method for implementing green building technology risk quantification management based on supervised learning provided by an embodiment of the present invention;

[0062] Figure 2 A functional module diagram of a system for implementing green building technology risk quantification management based on supervised learning provided in one embodiment of the present invention.

[0063] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0064] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0065] The embodiment of the present application provides a method for quantitative management of green building technology risks based on supervised learning. In the embodiment of the present application, the execution subject of the method for quantitative management of green building technology risks based on supervised learning includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for quantitative management of green building technology risks based on supervised learning can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0066] Reference Figure 1 FIG. 1 is a flow chart of a method for implementing green building technology risk quantification management based on supervised learning according to an embodiment of the present invention. In this embodiment, the method for implementing green building technology risk quantification management based on supervised learning includes steps S1-S5.

[0067] S1. Obtain a construction area corresponding to a green building, identify construction elements within the construction area, query construction rules corresponding to the construction elements, determine risk factors corresponding to the construction elements in combination with the construction rules, and collect factor data corresponding to the risk factors within the construction area.

[0068] The present invention determines the risk factors corresponding to the construction elements in combination with the construction rules, and can discover the risk factors existing in the construction area, so as to facilitate the subsequent effective risk management of the construction area, wherein the construction area is a specific range or location for construction activities, the construction elements are objects existing in the construction area, the construction rules are specific requirements that the construction elements should follow during the construction process, and the risk factors are factors or variables that have a potential negative impact on the activities or decisions of the construction elements.

[0069] As an embodiment of the present invention, the risk factors corresponding to the construction elements are determined in combination with the construction rules, including: identifying the rule character information in the construction rules, performing word segmentation processing on the rule character information to obtain rule character word segmentations, calculating the word segmentation importance corresponding to the rule character word segmentations, extracting the key rule word segmentations in the rule character word segmentations according to the word segmentation importance, matching the risk factors corresponding to the key rule word segmentations in a preset risk set table to obtain rule risk factors, and obtaining the risk factors corresponding to the construction elements according to the rule risk factors.

[0070] Among them, the rule character information is the text information in the construction rules, the rule character segmentation is the constituent words in the rule character information, the segmentation importance represents the corresponding importance of the rule character segmentation, the key rule segmentation is the representative segmentation in the rule character segmentation, and the preset risk set table is a table constructed by all collected risk factors.

[0071] Furthermore, the recognition of the rule character information in the construction rules can be achieved through OCR recognition technology; the word segmentation processing of the rule character information can be achieved through a word segmentation tool, and the word segmentation tool is compiled by a script language, such as JS script language; the word segmentation importance, analyzes the category corresponding to the rule character word segmentation, and selects the word segmentation with the largest word segmentation importance in the same category as the key rule word segmentation; matching the risk factors corresponding to the key rule word segmentation in the preset risk set table can be achieved through a similarity algorithm, calculating the similarity between each risk in the risk set table and the key rule word segmentation, and when the similarity is greater than a preset threshold, the corresponding risk is used as a rule risk factor.

[0072] Further, as an optional embodiment of the present invention, the calculation of the word segmentation importance corresponding to the regular character word segmentation includes: performing semantic analysis on the regular characters to obtain regular semantics, extracting semantic elements corresponding to the regular semantics, counting element frequencies corresponding to the semantic elements, calculating the semantic entropy value corresponding to the regular semantics based on the element frequencies, and determining the word segmentation importance corresponding to the regular character word segmentation based on the semantic entropy value.

[0073] Among them, the rule semantics is the meaning and explanation corresponding to the rule characters, the semantic elements are the components corresponding to the rule semantics, the element frequency is the probability of the occurrence of the semantic elements, and the semantic entropy value represents the uncertainty of the rule semantics. The higher the uncertainty, the lower the corresponding semantic importance. Furthermore, the semantic analysis of the rule characters can be achieved through semantic analysis; the semantic elements corresponding to the rule semantics can be extracted through part-of-speech tagging, and the elements in the rule semantics are extracted according to the part-of-speech; the semantic entropy value corresponding to the rule semantics can be obtained by combining the Shannon entropy algorithm with the element frequency calculation, and the semantic entropy values ​​are summed to obtain the total entropy value, and the total entropy value is converted into a percentage expression to obtain the word segmentation importance corresponding to the rule character word segmentation.

[0074] The present invention collects factor data corresponding to the risk factors in the construction area to facilitate subsequent quantitative processing of the risk factors. Furthermore, the collection of factor data corresponding to the risk factors in the construction area can be achieved through various sensors, such as temperature, humidity and humidity sensors.

[0075] S2. Quantitatively process the factor data to obtain factor quantitative values, calculate the factor variation values ​​corresponding to the risk factors based on the factor quantitative values, and calculate the risk probability corresponding to the risk factors based on a preset risk database, and screen out key risk factors from the risk factors by combining the factor variation values ​​and the risk probability.

[0076] The present invention can better understand the information expressed by the data by quantitatively processing the factor data, and provide a basis for the subsequent calculation of the factor variation value, wherein the factor quantitative value is the expression value corresponding to the factor data.

[0077] As an embodiment of the present invention, the quantitative processing of the factor data to obtain the factor quantitative value includes: performing data cleaning on the factor data to obtain target factor data, standardizing the target factor data to obtain standard factor data, encoding the standard factor data to obtain factor coding values, and quantitatively converting the factor coding values ​​to obtain factor quantitative values.

[0078] Among them, the target factor data is the data obtained after removing invalid data and abnormal data in the factor data, the standard factor data is the data obtained after eliminating the differences between the target factor data, and the factor coding value is the expression value obtained after encoding the standard factor data. Furthermore, the data cleaning of the factor data can be achieved through the box plot method; the standardization processing of the target factor data can be achieved through the Z-score standardization method; the encoding processing of the standard factor data can be achieved through the one-hot encoding technology; the quantitative conversion processing of the factor coding value can be achieved through a conversion function, and the conversion function is written in a programming language.

[0079] The present invention calculates the factor variation value corresponding to the risk factor based on the quantitative value of the factor, and can obtain the degree of change of the risk factor, and then understand the stability of the risk factor, so as to facilitate the screening and processing of key risk factors, wherein the factor variation value is the degree of change corresponding to the risk factor, and the lower the degree of change, the more stable the risk factor.

[0080] As an embodiment of the present invention, the factor variation value corresponding to the risk factor is calculated according to the factor quantitative value, including: counting the number of samples corresponding to the factor quantitative value, combining the factor quantitative value and the sample number, calculating the quantitative mean corresponding to the risk factor, combining the quantitative mean, the sample number and the factor quantitative value, and calculating the factor variation value corresponding to the risk factor by the following formula:

[0081]

[0082] Among them, A represents the factor variation value corresponding to the risk factor, B a Indicates the ath quantitative value in the quantitative value of the factor, represents the quantitative mean, a represents the serial number corresponding to the quantitative value of the factor, and q represents the number of samples.

[0083] Among them, the sample size is the number of quantitative values ​​of the factor corresponding to each risk factor. Furthermore, the sample size corresponding to the quantitative value of the factor can be obtained by statistical methods; the calculation of the quantitative mean corresponding to the risk factor can be achieved by the average function.

[0084] The present invention calculates the risk probability corresponding to the risk factor based on a preset risk database, and can understand the possibility of the occurrence of the risk factor, thereby improving the accuracy of subsequent screening of key risk factors, wherein the key risk factor is a high-incidence risk among the risk factors. Further, screening out the key risk factors from the risk factors can be achieved by a weighted average method, and weights corresponding to the factor variation value and the risk probability are respectively assigned. The weights can be assigned by a hierarchical analysis method, and the factor variation value and the risk probability are weighted and summed in combination with the corresponding weights to obtain a comprehensive score corresponding to the risk factor. According to the comprehensive score, the key risk factors are screened out from the risk factors.

[0085] As an embodiment of the present invention, the risk probability corresponding to the risk factor is calculated based on a preset risk database, including: extracting risk data corresponding to the risk factor from the preset risk database, screening out risk loss data and risk frequency data in the risk data, calculating the risk probability corresponding to the risk factor based on the risk frequency data, calculating the risk severity corresponding to the risk factor based on the risk loss data, and calculating the risk probability corresponding to the risk factor in combination with the risk severity and the risk probability.

[0086] Among them, the risk data is the data corresponding to the risk factor, the risk loss data is the relevant data on the economic loss caused by the risk factor in the risk data, the risk frequency data is the data recording the number of occurrences of the risk factor in the risk data, and the risk severity indicates the severity caused by the risk factor. Furthermore, the risk probability corresponding to the risk factor can be obtained by dividing the number of occurrences of each factor in the risk frequency data by the total number of occurrences; based on the risk loss data, the risk loss corresponding to the risk factor is determined, and the risk severity can be obtained by dividing the risk loss corresponding to each risk factor by the total risk loss; the risk probability corresponding to the risk factor can be obtained by multiplying the risk severity by the risk probability.

[0087] S3. Calculate the factor sensitivity between the key risk factors, calculate the factor risk value between the key risk factors according to the factor sensitivity, and set the risk level corresponding to the key risk factor according to the factor risk value.

[0088] By calculating the factor sensitivity between the key risk factors, the present invention can obtain the mutual influence degree between the key risk factors, and further understand the mutual correlation and dependency relationship between different factors in the key risk factors.

[0089] As an embodiment of the present invention, the calculation of the factor sensitivity between the key risk factors includes: extracting the factor attributes corresponding to the key risk factors, calculating the attribute priorities corresponding to the factor attributes, performing attribute screening processing on the factor attributes according to the attribute priorities to obtain target factor attributes, calculating the attribute association strength between the target factor attributes, and calculating the factor sensitivity between the key risk factors according to the attribute association strength.

[0090] Among them, the factor attribute is the characteristic attribute corresponding to the key risk factor, the attribute priority indicates the importance of the factor in the factor attribute relative to other factors, and the attribute association strength indicates the closeness of the association relationship between the target factor attributes.

[0091] Furthermore, the extraction of factor attributes corresponding to the key risk factors can be achieved through principal component analysis; the attribute priority is compared with a preset threshold, the preset threshold can be set to 0.8, or it can be set according to the actual application scenario. When the attribute priority is less than the preset threshold, the factor attributes are subjected to attribute screening processing to obtain the target factor attributes; the attribute association strengths are summed to obtain the total association strength, and the ratio of each attribute association strength to the total association strength is calculated to obtain the factor sensitivity between the key risk factors.

[0092] Optionally, as an optional embodiment of the present invention, the calculating the attribute priority corresponding to the factor attribute includes: performing vector conversion on the factor attribute to obtain a factor vector, performing dimensionality reduction processing on the factor vector to obtain a reduced-dimensionality factor vector, extracting a reduced-dimensionality feature vector in the reduced-dimensionality factor vector, calculating a vector eigenvalue corresponding to the feature vector, and calculating the attribute priority corresponding to the factor attribute according to the vector eigenvalue by the following formula:

[0093]

[0094] Among them, D represents the attribute priority corresponding to the factor attribute, F j represents the jth eigenvalue in the vector eigenvalue, j represents the sequence number of the vector eigenvalue, and r represents the number of vector eigenvalues.

[0095] Among them, the reduced dimension factor vector is the vector obtained after removing the redundant vectors in the factor vector, the reduced dimension feature vector is the characterization vector in the reduced dimension factor vector, and the vector eigenvalue is the linear value corresponding to the reduced dimension feature vector. Furthermore, the vector conversion of the factor attributes can be achieved through a vector algorithm, such as the word2vec algorithm; the dimensionality reduction processing of the factor vector can be achieved through an automatic encoder; the extraction of the reduced dimension feature vector in the reduced dimension factor vector can be achieved through a linear discriminant analysis algorithm; the calculation of the vector eigenvalue corresponding to the feature vector can be achieved through an eigenvalue decomposition algorithm.

[0096] Further, as an optional embodiment of the present invention, the calculation of the attribute association strength between the target factor attributes includes: collecting attribute observation data corresponding to the target factor attributes, calculating the attribute observation value corresponding to the target factor attribute based on the attribute observation data, calculating the observation mean corresponding to the target factor attribute based on the attribute observation value, and combining the attribute observation value and the observation mean to calculate the attribute association strength between the target factor attributes by the following formula:

[0097]

[0098] Among them, W represents the attribute association strength between the target factor attributes, H b represents the bth attribute observation value of the Hth attribute in the target factor attribute, G b represents the bth attribute observation value in the Gth attribute of the target factor attribute, and They represent the observed means corresponding to the Hth attribute and the Gth attribute in the target factor attributes respectively.

[0099] Among them, the attribute observation data is the specific description and recorded data corresponding to the target factor attributes, and the attribute observation value is the value of the target factor calculated based on the attribute observation data. Furthermore, the attribute observation data corresponding to the target factor attributes can be extracted from an existing database, and the database is a storage repository constructed by pre-collected data.

[0100] The present invention sets the risk level corresponding to the key risk factor according to the risk value of the factor, and the severity of the key risk factor can be understood through the risk level, which is convenient for the subsequent risk quantification processing of the construction area, wherein the risk level is the severity corresponding to the key risk factor. Furthermore, the risk ratio of each risk value in the factor risk value to the total risk value can be calculated, and the risk level corresponding to the key risk factor can be set according to the risk ratio. The risk level can be set to three levels: high, medium and low.

[0101] S4. Query the construction tasks corresponding to the construction area, calculate the risk impact effect of the key risk factors on the construction tasks, collect the element cost data corresponding to the construction elements, combine the element cost data and the risk impact effect, use a pre-trained supervised learning regression model to quantify the risk of the key risk factors, obtain quantitative risk factors, combine the quantitative risk factors and the risk level, and generate a risk quantification plan for the construction area.

[0102] The present invention can obtain the degree of influence caused by the key risk factors by calculating the risk impact effect of the key risk factors on the construction task, thereby providing a basis for subsequent risk quantification processing to improve the management efficiency of air volume quantification, wherein the risk impact effect represents the degree of influence caused by the key risk factors on the construction task.

[0103] As an embodiment of the present invention, the calculation of the risk impact effect of the key risk factors on the construction task includes: decomposing the construction task to obtain construction subtasks, analyzing the risk correlation between the key risk factors and the construction subtasks, constructing a risk correlation matrix between the key risk factors and the construction subtasks based on the risk correlation, and calculating the risk impact effect of the key risk factors on the construction task based on the risk correlation matrix.

[0104] Among them, the construction subtask is a more specific task corresponding to the construction task, the risk correlation represents the correlation between the key risk factors and the construction subtask, and the risk correlation matrix is ​​a square matrix constructed between the key risk factors and the construction subtasks. Each row in the matrix represents a key risk factor, and each column represents a construction subtask.

[0105] Furthermore, the decomposition processing of the construction task can be achieved through the work breakdown structure method, and the construction task is decomposed according to a hierarchical structure to form sub-tasks at different levels; the risk correlation between the key risk factors and the construction sub-tasks can be analyzed by querying the occurrence conditions corresponding to the key risk factors and analyzing the causal relationship between the occurrence conditions and the tasks of the construction sub-tasks; the risk correlation matrix between the key risk factors and the construction sub-tasks can be implemented by a matrix function, such as a zero matrix function; through the risk correlation matrix, the degree of influence of the key risk factors on the construction tasks is determined, and the probability of occurrence of the key risk factors is analyzed, and the degree of influence and the probability of occurrence are multiplied to obtain the risk impact effect of the key risk factors on the construction tasks.

[0106] The present invention combines the element cost data and the risk impact effect, and uses a pre-trained supervised learning regression model to perform risk quantification processing on the key risk factors, thereby concretizing the risk of the construction area, so as to effectively manage the risk of the construction area, wherein the element cost data is the cost data corresponding to the construction element, and the supervised learning regression model is a machine learning model that can quantify the key risk factors of the green building construction area, thereby concretizing the risk of the construction area, and the quantitative risk factor is a risk factor expressed by a measurable numerical value of the key risk factor. Further, the element cost data corresponding to the construction element can be collected from the official website of the manufacturer of the construction element; the steps of using the pre-trained supervised learning regression model to perform risk quantification processing on the key risk factors are: collecting key risk factor related data of the current green building construction area, ensuring the accuracy and completeness of the data, preliminarily organizing the data to make it meet the requirements of the model input, and extracting representative features from the key risk factor data. For example, if the key risk factor is the skill level of the construction personnel, the extractable features include years of work experience, the number of relevant certificates, the number of trainings, etc., and the extracted features are standardized to have a unified scale so that the model can be better processed. Common standardization methods can be used, such as mean normalization. The neurons in the hidden layer receive signals from the input layer and process them through activation functions. Common activation functions include ReLU (Rectified Linear Unit) and Sigmoid. The role of the activation function is to introduce nonlinearity so that the model can learn complex relationships. The output of the neuron is calculated based on the input signal and the current weight. During the training process, the model will continuously adjust the weights to minimize the prediction error. The weights are updated according to the gradient of the loss function through the back propagation algorithm. The output layer usually has only one neuron for outputting the risk quantification value. For regression problems, a linear activation function can be used to directly output a continuous numerical value to represent the risk level of the key risk factor, thereby obtaining the quantitative risk factor; the quantitative risk factor and the risk level are visualized to generate a risk quantification plan for the construction area.

[0107] The present invention determines the risk factors corresponding to the construction elements by combining the construction rules, and can discover the risk factors existing in the construction area, so as to facilitate the subsequent effective risk management of the construction area. The present invention can better understand the information expressed by the data by quantitatively processing the factor data, and provide a basis for the calculation of subsequent factor variation values. The present invention calculates the factor sensitivity between the key risk factors, and can obtain the degree of mutual influence between the key risk factors, and then understand the mutual correlation and dependency between different factors in the key risk factors. The present invention calculates the risk impact effect of the key risk factors on the construction task, and can obtain the degree of influence caused by the key risk factors, and then provide a basis for subsequent risk quantification processing to improve the management efficiency of air volume quantification. Therefore, the method for implementing green building technology risk quantification management based on supervised learning provided by the embodiment of the present invention can improve the efficiency of green building technology risk quantification management.

[0108] like Figure 2 , which is a functional module diagram of a system for implementing green building technology risk quantification management based on supervised learning provided by an embodiment of the present invention.

[0109] The system 100 for implementing green building technology risk quantification management based on supervised learning described in the present invention can be installed in an electronic device. According to the functions implemented, the system 100 for implementing green building technology risk quantification management based on supervised learning can include a data acquisition module 101, a key risk factor screening module 102, a risk level setting module 103 and a risk quantification module 104. The module described in the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0110] In this embodiment, the functions of each module / unit are as follows:

[0111] The data collection module 101 is used to obtain the construction area corresponding to the green building, identify the construction elements in the construction area, query the construction rules corresponding to the construction elements, determine the risk factors corresponding to the construction elements in combination with the construction rules, and collect factor data corresponding to the risk factors in the construction area;

[0112] The key risk factor screening module 102 is used to perform quantitative processing on the factor data to obtain the factor quantitative value, calculate the factor variation value corresponding to the risk factor according to the factor quantitative value, and calculate the risk probability corresponding to the risk factor according to a preset risk database, and screen out the key risk factors from the risk factors by combining the factor variation value and the risk probability;

[0113] The risk level setting module 103 is used to calculate the factor sensitivity between the key risk factors, calculate the factor risk value between the key risk factors according to the factor sensitivity, and set the risk level corresponding to the key risk factor according to the factor risk value;

[0114] The risk quantification module 104 is used to query the construction tasks corresponding to the construction area, calculate the risk impact effects of the key risk factors on the construction tasks, collect the element cost data corresponding to the construction elements, combine the element cost data and the risk impact effects, and use a pre-trained supervised learning regression model to perform risk quantification processing on the key risk factors to obtain quantitative risk factors, and combine the quantitative risk factors and the risk level to generate a risk quantification plan for the construction area.

[0115] In detail, each module described in the system 100 for implementing green building technology risk quantification management based on supervised learning described in the embodiment of the present application is used in the same manner as described above. Figure 1 The technical means described in the method of realizing quantitative management of green building technology risks based on supervised learning are the same as those in the method and can produce the same technical effects, so they will not be repeated here.

[0116] In the several embodiments provided by the present invention, it should be understood that the provided methods and systems can be implemented in other ways. For example, the method embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for quantitative management of green building technology risks based on supervised learning, characterized in that: The method comprises: Acquire a construction area corresponding to a green building, identify construction elements within the construction area, query construction rules corresponding to the construction elements, determine risk factors corresponding to the construction elements in combination with the construction rules, and collect factor data corresponding to the risk factors within the construction area; Quantitatively process the factor data to obtain factor quantitative values, calculate the factor variation values ​​corresponding to the risk factors according to the factor quantitative values, and calculate the risk probability corresponding to the risk factors according to a preset risk database, and screen out key risk factors from the risk factors by combining the factor variation values ​​and the risk probability; Calculating the factor sensitivity between the key risk factors, calculating the factor risk value between the key risk factors according to the factor sensitivity, and setting the risk level corresponding to the key risk factor according to the factor risk value; Query the construction tasks corresponding to the construction area, calculate the risk impact effects of the key risk factors on the construction tasks, collect the element cost data corresponding to the construction elements, combine the element cost data and the risk impact effects, use a pre-trained supervised learning regression model to quantify the risks of the key risk factors, obtain quantitative risk factors, combine the quantitative risk factors and the risk level, and generate a risk quantification plan for the construction area.

2. A method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 1, characterized in that: Determining the risk factors corresponding to the construction elements in combination with the construction rules includes: Identifying regular character information in the construction rules, performing word segmentation processing on the regular character information to obtain regular character word segmentations; Calculate the importance of the word segmentation corresponding to the regular character word segmentation; Extracting key rule segmentations in the rule character segmentations according to the segmentation importance; Match the risk factors corresponding to the key rule segmentation in the preset risk set table to obtain the rule risk factors; According to the rule risk factors, the risk factors corresponding to the construction elements are obtained.

3. A method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 2, characterized in that: The calculating the importance of the word segmentation corresponding to the regular character word segmentation includes: Performing semantic analysis on the regular characters to obtain regular semantics; Extracting semantic elements corresponding to the semantics of the rule, and counting element frequencies corresponding to the semantic elements; In combination with the element frequency, a semantic entropy value corresponding to the semantics of the rule is calculated; The word segmentation importance corresponding to the regular character word segmentation is determined according to the semantic entropy value.

4. The method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 1, characterized in that: The quantitative processing of the factor data to obtain the factor quantitative value includes: Performing data cleaning on the factor data to obtain target factor data; Performing standardization processing on the target factor data to obtain standard factor data; Encoding the standard factor data to obtain factor encoding values; The factor coding value is quantitatively converted to obtain the factor quantitative value.

5. The method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 1, characterized in that: Calculating the factor variation value corresponding to the risk factor according to the factor quantitative value includes: Counting the number of samples corresponding to the quantitative values ​​of the factors; Calculate the quantitative mean corresponding to the risk factor by combining the quantitative value of the factor and the sample size; Combining the quantitative mean, the sample size and the quantitative value of the factor, the factor variation value corresponding to the risk factor is calculated by the following formula: Among them, A represents the factor variation value corresponding to the risk factor, B a Indicates the ath quantitative value in the quantitative value of the factor, represents the quantitative mean, a represents the serial number corresponding to the quantitative value of the factor, and q represents the number of samples.

6. A method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 1, characterized in that: The step of calculating the risk probability corresponding to the risk factor according to a preset risk database includes: Extracting risk data corresponding to the risk factor from a preset risk database; Filter out risk loss data and risk frequency data from the risk data; Calculating the risk probability corresponding to the risk factor according to the risk frequency data; Calculating the risk severity corresponding to the risk factor according to the risk loss data; The risk probability corresponding to the risk factor is calculated by combining the risk severity and the risk probability.

7. The method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 1, characterized in that: The calculating the factor sensitivity between the key risk factors comprises: Extracting factor attributes corresponding to the key risk factors, and calculating attribute priorities corresponding to the factor attributes; According to the attribute priority, the factor attributes are subjected to attribute screening processing to obtain target factor attributes; Calculating the attribute association strength between the target factor attributes; The factor sensitivity between the key risk factors is calculated according to the attribute association strength.

8. A method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 7, characterized in that: The calculating the attribute priority corresponding to the factor attribute includes: Performing vector conversion on the factor attributes to obtain a factor vector; Performing dimensionality reduction processing on the factor vector to obtain a dimensionality reduction factor vector; Extracting a dimension reduction feature vector from the dimension reduction factor vector, and calculating a vector eigenvalue corresponding to the feature vector; According to the vector eigenvalue, the attribute priority corresponding to the factor attribute is calculated by the following formula: Among them, D represents the attribute priority corresponding to the factor attribute, F j represents the jth eigenvalue in the vector eigenvalue, j represents the sequence number of the vector eigenvalue, and r represents the number of vector eigenvalues.

9. The method for implementing green building technology risk quantification management based on supervised learning as claimed in claim 1, characterized in that: The calculating the risk impact effect of the key risk factors on the construction task includes: Decomposing the construction task to obtain construction subtasks; Analyze the risk correlation between the key risk factors and the construction subtasks; According to the risk correlation, construct a risk correlation matrix between the key risk factors and the construction subtasks; The risk impact effect of the key risk factors on the construction task is calculated according to the risk association matrix.

10. A system for quantitative management of green building technology risks based on supervised learning, characterized in that: The system comprises: A data collection module is used to obtain a construction area corresponding to a green building, identify construction elements in the construction area, query construction rules corresponding to the construction elements, determine risk factors corresponding to the construction elements in combination with the construction rules, and collect factor data corresponding to the risk factors in the construction area; A key risk factor screening module is used to perform quantitative processing on the factor data to obtain the factor quantitative value, calculate the factor variation value corresponding to the risk factor according to the factor quantitative value, and calculate the risk probability corresponding to the risk factor according to a preset risk database, and screen out key risk factors from the risk factors by combining the factor variation value and the risk probability; A risk level setting module, used to calculate the factor sensitivity between the key risk factors, calculate the factor risk value between the key risk factors according to the factor sensitivity, and set the risk level corresponding to the key risk factor according to the factor risk value; The risk quantification module is used to query the construction tasks corresponding to the construction area, calculate the risk impact effects of the key risk factors on the construction tasks, collect the element cost data corresponding to the construction elements, combine the element cost data and the risk impact effects, and use a pre-trained supervised learning regression model to perform risk quantification processing on the key risk factors to obtain quantitative risk factors, and combine the quantitative risk factors and the risk level to generate a risk quantification plan for the construction area.

Images