Building low-carbon operation management method based on data analysis

Abstract

The invention relates to the technical field of smart communities, and discloses a building low-carbon operation management method based on data analysis, which comprises the following steps: acquiring building data, preprocessing the data to obtain a multi-dimensional data set, performing feature extraction by adopting a principal component analysis method aiming at the multi-dimensional data set, and calculating a feature value and a feature vector of a covariance matrix to obtain a building low-carbon operation management result. Selecting the first k principal components with relatively large feature values as new feature spaces to obtain a feature data set after dimension reduction, taking the feature data set after dimension reduction as input, constructing a credit rating prediction model by using a random forest algorithm, calculating importance scores of all features to rating results, and sorting the results according to the importance scores of the features to obtain a credit rating result. And selecting the first n features with higher scores as key feature factors, and constructing a credit evaluation feature vector. According to the method, a more accurate and comprehensive evaluation result is provided for building credit rating, and green building development and realization of energy conservation and emission reduction targets are effectively supported.

Description

Technical Field

[0001] The present invention relates to the technical field of smart communities, and in particular to a low-carbon building operation management method based on data analysis. Background Art

[0002] In the low-carbon operation and management of buildings, how to identify the key factors that have a greater impact on building energy consumption and carbon emissions from massive building operation data is a key technical issue. Since building operations involve multiple dimensions such as energy consumption, equipment operation, and personnel activities, the operating characteristics of different buildings vary greatly, resulting in complex factors affecting building energy consumption and carbon emissions. In addition, building operation data usually presents high-dimensional and nonlinear characteristics, which increases the difficulty of data analysis and feature extraction. Therefore, there is an urgent need for an effective data analysis method that can automatically learn to extract key features from building operation data and quantify the impact of each feature on building energy consumption and carbon emissions. At the same time, due to the differences in the operating characteristics of different buildings, the extracted key features may change over time. Data analysis methods are required to adapt to this dynamic change, continuously learn and optimize feature vectors, in order to ensure the accuracy and effectiveness of low-carbon operation and management of buildings. Summary of the invention

[0003] The purpose of the present invention is to overcome one or more of the above-mentioned existing technical problems and provide a building low-carbon operation management method based on data analysis.

[0004] To achieve the above objectives, the present invention provides a building low-carbon operation management method based on data analysis, comprising: Obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, remove missing values ​​and outliers, and perform data standardization to obtain a standardized multidimensional data set; For the standardized multidimensional data set, the principal component analysis method is used for feature extraction. By calculating the eigenvalues ​​and eigenvectors of the covariance matrix, the first k principal components with larger eigenvalues ​​are selected as the new feature space to obtain the feature data set after dimensionality reduction. The feature data set after dimensionality reduction is used as input, and a credit rating prediction model is constructed using the random forest algorithm. The importance score of each feature to the rating result is calculated through ensemble learning of decision trees. According to the ranking results of feature importance scores, the top n features with higher scores are selected as key feature factors to construct the credit assessment feature vector; Continuously collect building operation data and evaluation feedback data, and regularly update the feature vector. If the amount of evaluation feedback data reaches the preset threshold, the feature vector update process is triggered to recalculate the feature importance score and update the key feature factor. The building operation data is mined and analyzed to identify potential feature factors. If the correlation between the new feature and the existing feature vector is lower than the preset threshold, it will be included in the feature pool.

[0005] According to one aspect of the present invention, building energy consumption data, carbon emission data and energy-saving measures data are obtained, and the data are preprocessed to determine whether there are missing values ​​and outliers in the data. If there are missing values ​​and outliers, the missing values ​​and outlier data are removed; According to the pre-processed building energy consumption data, carbon emission data and energy-saving measures data, the data of each dimension is standardized by using the data standardization algorithm to obtain a standardized multi-dimensional data set; For the standardized multidimensional data set, the principal component analysis algorithm is used to extract the main features of the data set, reduce the data dimension, and obtain the reduced-dimensional data set; According to the data set after dimensionality reduction, the K-means clustering algorithm is used to perform cluster analysis on the data, and the data is divided into several clusters according to the clustering results to obtain a cluster set; For each cluster, an energy consumption prediction model is constructed based on building energy consumption data, carbon emission data and energy-saving measures data to predict the energy consumption and carbon emission of the building in the future. According to the energy consumption prediction results and carbon emission prediction results of the clusters, the energy-saving potential of each cluster is judged, the priority of energy-saving transformation is determined, and the energy-saving transformation plan is obtained; For the energy-saving transformation plan, a multi-objective optimization algorithm is used to optimize the parameters of energy-saving measures and obtain the optimal energy-saving transformation implementation plan.

[0006] According to one aspect of the present invention, a covariance matrix is ​​constructed based on the standardized multidimensional data set, and the covariance matrix is ​​decomposed using a singular value decomposition algorithm to obtain eigenvalues ​​and corresponding eigenvectors; According to the size of the eigenvalue, sort the eigenvalues ​​from large to small, and select the eigenvectors corresponding to the first k largest eigenvalues ​​as the principal component vectors; For the standardized multidimensional data set, an inner product operation is performed on each data sample and the principal component vector to obtain the projection coordinates of the sample in the principal component space; According to the corresponding relationship between the projection coordinates and the principal component vector, the coordinate representation of each sample in the principal component space is obtained to obtain the feature data set after dimensionality reduction; For the feature data set after dimensionality reduction, the k-means clustering algorithm is used to cluster the data. The position of each cluster center is optimized iteratively to minimize the sum of square distances from the sample to the cluster center until the cluster center no longer changes, and the final clustering result is obtained. According to the clustering results, samples of different categories are distinguished. For each category, the mean vector of samples within the category is calculated as the representative vector of the category; According to the distance relationship between representative vectors, the distance matrix between categories is constructed, and the representative vectors are clustered using a hierarchical clustering algorithm to obtain the hierarchical relationship between categories, thereby realizing the high-level semantic representation of the standardized multidimensional data set.

[0007] According to one aspect of the present invention, raw data related to credit rating is obtained, the data is preprocessed to remove noise and outliers, and feature selection and feature extraction are performed; Use principal component analysis and other dimensionality reduction algorithms to reduce the dimensionality of feature data and obtain a set of feature data after dimensionality reduction; The feature data set after dimension reduction is divided into a training set and a test set. The training set is used to train the random forest model, and the test set is used to evaluate the model performance. According to the preset random forest parameters, a random forest model is constructed, and the model is trained using the training set data to obtain a trained random forest model; Input the test set data into the trained random forest model to perform credit rating prediction and obtain the prediction results; The importance score of each feature to the credit rating prediction result is calculated through the feature importance evaluation algorithm, and the features are sorted according to the importance score; According to the feature importance ranking results, features with higher importance scores are selected as key features for credit rating prediction.

[0008] According to one aspect of the present invention, original feature data related to credit assessment is obtained, and the features are preprocessed, including missing value processing, outlier processing and data normalization; A machine learning algorithm is used to train the credit assessment model based on the preprocessed feature data, and the importance score of each feature is calculated through the feature importance assessment method; Sort all features in descending order according to their importance scores, and select the top n features with higher scores as key feature factors; Based on the selected key feature factors, a credit assessment feature vector is constructed to form an optimized feature subset; The credit assessment model is trained based on the optimized feature vector to obtain the final credit risk prediction result; Compare the credit risk prediction results with the preset risk level threshold, determine the user's credit risk level, and output a credit assessment report.

[0009] According to one aspect of the present invention, building operation data and evaluation feedback data are obtained, and the data are cleaned and standardized according to a preset data format and data standard to obtain standardized building operation data and evaluation feedback data; For the standardized building operation data and evaluation feedback data, the principal component analysis algorithm is used to extract the main features of the data and obtain the feature vector representing the main features of the data; According to the feature vector, the importance score of each feature is calculated using the random forest algorithm, and the key feature factors are determined according to the importance score sorting; Continuously obtain the latest building operation data and evaluation feedback data to determine whether the data volume reaches the preset threshold. If the threshold is reached, the feature vector update process is triggered; In the feature vector update process, the latest building operation data and evaluation feedback data are merged with the original data, and the principal component analysis is re-performed to obtain the updated feature vector; According to the updated feature vector, the importance score of each feature is recalculated, and the key feature factors are updated according to the importance score sorting.

[0010] According to one aspect of the present invention, building operation data is acquired, and data mining algorithms are used to continuously analyze the data to identify new potential characteristic factors; Calculate the correlation between the new feature and the existing feature vector. If the correlation is lower than the preset threshold, the new feature is included in the feature pool. Based on the enriched feature pool, a credit assessment model is constructed. Obtain changes in business requirements. If new business requirements emerge, the model update mechanism is triggered; In response to new business needs, we continue to mine new potential feature factors from building operation data; The newly mined feature factors are correlated with the existing feature pool. If the correlation is lower than the preset threshold, they are included in the feature pool and the credit assessment model is retrained using the updated feature pool.

[0011] To achieve the above objectives, the present invention provides a building low-carbon operation management system based on data analysis, comprising: Data acquisition and processing module: obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, remove missing values ​​and outliers, and perform data standardization to obtain a standardized multidimensional data set; Feature data set acquisition module: For the standardized multidimensional data set, the principal component analysis method is used to extract features. By calculating the eigenvalues ​​and eigenvectors of the covariance matrix, the first k principal components with larger eigenvalues ​​are selected as the new feature space to obtain the feature data set after dimensionality reduction; Feature scoring module: takes the reduced feature data set as input, uses the random forest algorithm to build a credit rating prediction model, and calculates the importance score of each feature to the rating result through ensemble learning of decision trees; Feature vector construction module: According to the feature importance score sorting results, the top n features with higher scores are selected as key feature factors to construct the credit assessment feature vector; Key factor screening module: continuously collects building operation data and evaluation feedback data, and regularly updates the feature vector. If the amount of evaluation feedback data reaches the preset threshold, the feature vector update process is triggered, the feature importance score is recalculated, and the key feature factor is updated; Feature factor update module: mines and analyzes building operation data to identify potential feature factors. If the correlation between the new feature and the existing feature vector is lower than the preset threshold, it will be included in the feature pool.

[0012] To achieve the above-mentioned purpose, the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the above-mentioned low-carbon building operation management method based on data analysis is implemented.

[0013] To achieve the above-mentioned purpose, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned low-carbon operation and management method for a building based on data analysis is implemented.

[0014] Based on this, the beneficial effects of the present invention are: by collecting data on building energy consumption, carbon emissions and energy-saving measures, the data is preprocessed and standardized, and then the principal component analysis method is used to reduce the dimension and extract key features. Subsequently, the random forest algorithm is used to build a prediction model, calculate the feature importance score, select high-scoring features to build an evaluation vector, and realize a dynamic update mechanism. By continuously collecting operational data and evaluation feedback, regularly updating feature vectors, and mining new potential feature factors, the evaluation dimensions are continuously enriched. This adaptive method can improve the generalization ability of the model, adapt to changing business needs, provide more accurate and comprehensive evaluation results for building credit ratings, and effectively support the development of green buildings and the realization of energy conservation and emission reduction goals. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a flow chart of a building low-carbon operation management method based on data analysis according to an exemplary embodiment; Figure 2 The present invention is a flow chart of a building low-carbon operation management system based on data analysis according to an exemplary embodiment. DETAILED DESCRIPTION

[0016] The present invention will now be discussed with reference to exemplary embodiments. It should be understood that the embodiments discussed are only for enabling those skilled in the art to better understand and thereby implement the present invention, rather than implying any limitation on the scope of the present invention.

[0017] As used herein, the term “including” and variations thereof are to be interpreted as open-ended terms meaning “including but not limited to.” The term “based on” is to be interpreted as “based, at least in part, on,” and the terms “one embodiment” and “an embodiment” are to be interpreted as “at least one embodiment.”

[0018] According to one embodiment of the present invention, Figure 1 is a flow chart of a building low-carbon operation management method based on data analysis according to an exemplary embodiment. Figure 1 To achieve the above purpose, the present invention provides a building low-carbon operation management method based on data analysis, comprising: Obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, remove missing values ​​and outliers, and perform data standardization to obtain a standardized multidimensional data set; For the standardized multidimensional data set, the principal component analysis method is used for feature extraction. By calculating the eigenvalues ​​and eigenvectors of the covariance matrix, the first k principal components with larger eigenvalues ​​are selected as the new feature space to obtain the feature data set after dimensionality reduction. The feature data set after dimensionality reduction is used as input, and a credit rating prediction model is constructed using the random forest algorithm. The importance score of each feature to the rating result is calculated through ensemble learning of decision trees. According to the ranking results of feature importance scores, the top n features with higher scores are selected as key feature factors to construct the credit assessment feature vector; Continuously collect building operation data and evaluation feedback data, and regularly update the feature vector. If the amount of evaluation feedback data reaches the preset threshold, the feature vector update process is triggered to recalculate the feature importance score and update the key feature factor. The building operation data is mined and analyzed to identify potential feature factors. If the correlation between the new feature and the existing feature vector is lower than the preset threshold, it will be included in the feature pool.

[0019] According to one embodiment of the present invention, building energy consumption data, carbon emission data and energy-saving measures data are obtained, and pre-processed to determine whether there are missing values ​​and outliers in the data. If there are missing values ​​and outliers, the missing values ​​and outlier data are removed; According to the pre-processed building energy consumption data, carbon emission data and energy-saving measures data, the data of each dimension is standardized by using the data standardization algorithm to obtain a standardized multi-dimensional data set; For the standardized multidimensional data set, the principal component analysis algorithm is used to extract the main features of the data set, reduce the data dimension, and obtain the reduced-dimensional data set; According to the data set after dimensionality reduction, the K-means clustering algorithm is used to perform cluster analysis on the data, and the data is divided into several clusters according to the clustering results to obtain a cluster set; For each cluster, an energy consumption prediction model is constructed based on building energy consumption data, carbon emission data and energy-saving measures data to predict the energy consumption and carbon emission of the building in the future. According to the energy consumption prediction results and carbon emission prediction results of the clusters, the energy-saving potential of each cluster is judged, the priority of energy-saving transformation is determined, and the energy-saving transformation plan is obtained; For the energy-saving transformation plan, a multi-objective optimization algorithm is used to optimize the parameters of energy-saving measures and obtain the optimal energy-saving transformation implementation plan.

[0020] According to one embodiment of the present invention, a covariance matrix is ​​constructed based on the standardized multidimensional data set, and the covariance matrix is ​​decomposed using a singular value decomposition algorithm to obtain eigenvalues ​​and corresponding eigenvectors; According to the size of the eigenvalue, sort the eigenvalues ​​from large to small, and select the eigenvectors corresponding to the first k largest eigenvalues ​​as the principal component vectors; For the standardized multidimensional data set, an inner product operation is performed on each data sample and the principal component vector to obtain the projection coordinates of the sample in the principal component space; According to the corresponding relationship between the projection coordinates and the principal component vector, the coordinate representation of each sample in the principal component space is obtained to obtain the feature data set after dimensionality reduction; For the feature data set after dimensionality reduction, the k-means clustering algorithm is used to cluster the data. The position of each cluster center is optimized iteratively to minimize the sum of square distances from the sample to the cluster center until the cluster center no longer changes, and the final clustering result is obtained. According to the clustering results, samples of different categories are distinguished. For each category, the mean vector of samples within the category is calculated as the representative vector of the category; According to the distance relationship between representative vectors, the distance matrix between categories is constructed, and the representative vectors are clustered using a hierarchical clustering algorithm to obtain the hierarchical relationship between categories, thereby realizing the high-level semantic representation of the standardized multidimensional data set.

[0021] According to one embodiment of the present invention, raw data related to credit rating is obtained, the data is preprocessed to remove noise and outliers, and feature selection and feature extraction are performed; Use principal component analysis and other dimensionality reduction algorithms to reduce the dimensionality of feature data and obtain a set of feature data after dimensionality reduction; The feature data set after dimension reduction is divided into a training set and a test set. The training set is used to train the random forest model, and the test set is used to evaluate the model performance. According to the preset random forest parameters, a random forest model is constructed, and the model is trained using the training set data to obtain a trained random forest model; Input the test set data into the trained random forest model to perform credit rating prediction and obtain the prediction results; The importance score of each feature to the credit rating prediction result is calculated through the feature importance evaluation algorithm, and the features are sorted according to the importance score; According to the feature importance ranking results, features with higher importance scores are selected as key features for credit rating prediction.

[0022] According to one embodiment of the present invention, original feature data related to credit assessment is obtained, and the features are preprocessed, including missing value processing, outlier processing and data normalization; A machine learning algorithm is used to train the credit assessment model based on the preprocessed feature data, and the importance score of each feature is calculated through the feature importance assessment method; Sort all features in descending order according to their importance scores, and select the top n features with higher scores as key feature factors; Based on the selected key feature factors, a credit assessment feature vector is constructed to form an optimized feature subset; The credit assessment model is trained based on the optimized feature vector to obtain the final credit risk prediction result; Compare the credit risk prediction results with the preset risk level threshold, determine the user's credit risk level, and output a credit assessment report.

[0023] According to one embodiment of the present invention, building operation data and evaluation feedback data are obtained, and the data are cleaned and standardized according to a preset data format and data standard to obtain standardized building operation data and evaluation feedback data; For the standardized building operation data and evaluation feedback data, the principal component analysis algorithm is used to extract the main features of the data and obtain the feature vector representing the main features of the data; According to the feature vector, the importance score of each feature is calculated using the random forest algorithm, and the key feature factors are determined according to the importance score sorting; Continuously obtain the latest building operation data and evaluation feedback data to determine whether the data volume reaches the preset threshold. If the threshold is reached, the feature vector update process is triggered; In the feature vector update process, the latest building operation data and evaluation feedback data are merged with the original data, and the principal component analysis is re-performed to obtain the updated feature vector; According to the updated feature vector, the importance score of each feature is recalculated, and the key feature factors are updated according to the importance score sorting.

[0024] According to one embodiment of the present invention, building operation data is obtained, and data mining algorithms are used to continuously analyze the data to identify new potential characteristic factors; Calculate the correlation between the new feature and the existing feature vector. If the correlation is lower than the preset threshold, the new feature is included in the feature pool. Based on the enriched feature pool, a credit assessment model is constructed. Obtain changes in business requirements. If new business requirements emerge, the model update mechanism is triggered; In response to new business needs, we continue to mine new potential feature factors from building operation data; The newly mined feature factors are correlated with the existing feature pool. If the correlation is lower than the preset threshold, they are included in the feature pool and the credit assessment model is retrained using the updated feature pool.

[0025] According to one embodiment of the present invention, after obtaining building energy consumption data, carbon emission data and energy-saving measures data, firstly, a box plot and other methods are used to preprocess the data to determine whether there are missing values ​​and outliers. If it is found that the missing value exceeds 10% or the outlier exceeds 5%, the corresponding data record is removed. Then, the preprocessed data is standardized by the minimum-maximum standardization algorithm, and each dimension data is mapped to the interval [0,1] to eliminate the dimension effect. Then, the principal component analysis algorithm is used to extract features from the standardized multidimensional data, and the first k principal components with a cumulative variance contribution rate of 85% are selected to achieve data dimensionality reduction. On this basis, the K-means clustering algorithm is used to perform cluster analysis on the reduced-dimensional data, and the optimal number of clusters is determined to be 4 by the elbow rule, and the data is divided into 4 cluster clusters. For each cluster cluster, energy consumption data and carbon emission data are selected as feature variables, and energy-saving measures data are selected as label variables. A BP neural network energy consumption prediction model is constructed, and a 5-fold cross-validation method is used for model training and evaluation, and the average prediction accuracy reaches more than 90%. According to the prediction results, the energy-saving potential of each cluster is analyzed. The energy-saving potential is ranked from high to low in cluster 2, cluster 4, cluster 1 and cluster 3, and the priority of energy-saving transformation is determined accordingly. Finally, for cluster 2 with the highest priority, a multi-objective particle swarm optimization algorithm is used to optimize the parameters of eight energy-saving measures, such as exterior wall insulation materials and window performance parameters, under the constraints of meeting the requirements of building use functions and comfort, to obtain the optimal energy-saving transformation implementation plan.

[0026] According to one embodiment of the present invention, a multidimensional data set is first standardized so that the mean value of the data in each dimension is 0 and the variance is 1. Then a covariance matrix is ​​constructed based on the standardized data, and the matrix is ​​decomposed using a singular value decomposition algorithm to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvalues ​​are sorted from large to small, and the eigenvectors corresponding to the first k largest eigenvalues ​​are selected as principal component vectors, where the value of k can be determined according to the cumulative contribution rate, for example, the first few principal components with a cumulative contribution rate of 80% are taken. Then, the standardized data sample is inner-producted with the principal component vector to obtain the projection coordinates of the sample in the principal component space, and the feature data set after dimensionality reduction can be obtained according to the corresponding relationship between the coordinates and the principal component vector. On this basis, the k-means clustering algorithm is used to cluster the data after dimensionality reduction, and the position of the cluster center is iteratively optimized so that the sum of the squares of the distances from the sample to the cluster center is minimized. The initial position of the cluster center can be randomly selected, and its position is continuously updated during the iteration process until the center no longer changes, and the final clustering result can be obtained. The number of clustering categories can be determined by the elbow method and other methods, such as taking the number of categories corresponding to the inflection point of the clustering evaluation index SSE. According to the clustering results, samples of different categories can be distinguished, and the mean vector of samples within each category is calculated as the representative of the category. Finally, based on the distance relationship between the representative vectors of each category, the distance matrix between categories is constructed, and the representative vectors are clustered using a hierarchical clustering algorithm to obtain the hierarchical relationship between categories. Hierarchical clustering can use different distance metrics and connection methods such as the shortest distance method, the longest distance method, and the average distance method to generate a cluster pedigree diagram. By analyzing the cluster pedigree diagram, a high-level semantic representation of the standardized multidimensional data set can be achieved, revealing the internal structure and relationship of the data. For example, according to the hierarchical structure of the cluster pedigree diagram, the categories can be divided layer by layer from top to bottom to obtain a semantic tree representation, which reflects the semantic organizational relationship of the data at different granularities and provides richer information for subsequent data interpretation and application.

[0027] According to one embodiment of the present invention, the original building operation data is obtained from the energy consumption monitoring system, equipment operation records, and traffic statistics of the building, and the evaluation feedback data of the building users are collected through questionnaires, interviews, and the like. The data is cleaned using Python's Pandas library, records with missing values ​​exceeding 20% ​​are removed, outliers are corrected or deleted, and standardized according to the preset data schema and converted into a unified data format. Next, the PCA class in the scikit-learn library is used to perform principal component analysis on the standardized data, and the number of principal components is set to 50% of the data dimension, and the main features of the data are extracted and the corresponding feature vectors are obtained. Then, the random forest algorithm is used to calculate the importance score of each feature, and a random forest model is constructed by setting parameters such as n_estimators=100 and max_depth=10. The feature importance is obtained using the feature_importances_ attribute of the model, and the features are sorted from high to low according to the importance score, and the top 20% of the features are selected as key feature factors. In order to adapt to changes in the building operation status, the feature vector update process is triggered whenever the amount of new data reaches 30% of the original data. The new data is merged with the original data, and the principal component analysis and feature importance calculation are re-performed to update the feature vector and key feature factors. Finally, the updated feature vector and key feature factors are applied to building operation management and evaluation feedback analysis. For example, by monitoring the changing trend of key feature factors, problems in building operation can be discovered and solved in a timely manner; operation management strategies can be adjusted according to the importance of different features to optimize resource allocation; combined with evaluation feedback data, the correlation between building user satisfaction and key feature factors can be analyzed, and targeted improvement measures can be proposed to continuously improve the level of building operation management and user experience.

[0028] According to one embodiment of the present invention, a large amount of data is first collected from the building operation process, including energy consumption data, equipment operation data, environmental monitoring data, etc., and the data volume reaches TB level. Then, the Apriori association rule mining algorithm is used to analyze the data. By setting the minimum support to 05 and the minimum confidence to 8, 20 potential characteristic factors such as equipment failure and ambient temperature, energy consumption abnormality and equipment aging are identified. Then, the Pearson correlation coefficient is used to calculate the correlation between the new feature and the existing 100 features. If the correlation coefficient is less than 1, the new feature is included in the feature pool. Based on the enriched 120 features, the random forest algorithm is used to build a credit assessment model, which is trained by 10-fold cross validation. The accuracy of the model is increased from the original 85% to 93%. When business needs change and the safety of the building needs to be considered, the model update mechanism is triggered. 10 new characteristic factors such as the number of fire alarms and the intact rate of fire protection facilities are mined from the massive building operation data. After calculating, the Pearson correlation coefficients of these new features and the existing feature pool are all less than 1, so they are included in the feature pool to form 130 features. Finally, the gradient boosting decision tree algorithm was used to retrain the credit assessment model using the updated feature pool, and the model hyperparameters were optimized through grid search, so that the model's accuracy in building safety assessment reached 95%. The overall performance was further improved, adapting to new business needs and achieving continuous optimization of the model.

[0029] In addition, in order to achieve the above-mentioned purpose, the present invention also provides a building low-carbon operation management system based on data analysis. Figure 2 is a flow chart of a building low-carbon operation management system based on data analysis according to an exemplary embodiment. Figure 2 As shown, a building low-carbon operation management system based on data analysis in the present invention includes: Data acquisition and processing module: obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, remove missing values ​​and outliers, and perform data standardization to obtain a standardized multidimensional data set; Feature data set acquisition module: For the standardized multidimensional data set, the principal component analysis method is used to extract features. By calculating the eigenvalues ​​and eigenvectors of the covariance matrix, the first k principal components with larger eigenvalues ​​are selected as the new feature space to obtain the feature data set after dimensionality reduction; Feature scoring module: takes the reduced feature data set as input, uses the random forest algorithm to build a credit rating prediction model, and calculates the importance score of each feature to the rating result through ensemble learning of decision trees; Feature vector construction module: According to the feature importance score sorting results, the top n features with higher scores are selected as key feature factors to construct the credit assessment feature vector; Key factor screening module: continuously collects building operation data and evaluation feedback data, and regularly updates the feature vector. If the amount of evaluation feedback data reaches the preset threshold, the feature vector update process is triggered, the feature importance score is recalculated, and the key feature factor is updated; Feature factor update module: mines and analyzes building operation data to identify potential feature factors. If the correlation between the new feature and the existing feature vector is lower than the preset threshold, it will be included in the feature pool.

[0030] In order to achieve the above-mentioned purpose of the invention, the present invention also provides an electronic device, which includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the above-mentioned low-carbon building operation and management method based on data analysis is implemented.

[0031] In order to achieve the above-mentioned purpose of the invention, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned low-carbon building operation management method based on data analysis is implemented.

[0032] Based on this, the beneficial effect of the present invention is that by collecting data on building energy consumption, carbon emissions and energy-saving measures, the data is preprocessed and standardized, and then the principal component analysis method is used to reduce the dimension and extract key features. Subsequently, the random forest algorithm is used to build a prediction model, calculate the feature importance score, select high-scoring features to build an evaluation vector, and realize a dynamic update mechanism. By continuously collecting operational data and evaluation feedback, regularly updating feature vectors, and mining new potential feature factors, the evaluation dimension is continuously enriched. This adaptive method can improve the generalization ability of the model, adapt to changing business needs, provide more accurate and comprehensive evaluation results for building credit ratings, and effectively support the development of green buildings and the realization of energy conservation and emission reduction goals.

[0033] Those of ordinary skill in the art will appreciate that the modules and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0034] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and equipment can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0035] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0036] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0037] In addition, each functional module in the embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0038] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the energy-saving signal sending / receiving method of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0039] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.

[0040] It should be understood that the size of the serial numbers of each step in the content of the invention and the embodiments of the present invention does not absolutely mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

Claims

1. A building low-carbon operation management method based on data analysis, characterized in that: include: Obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, remove missing values ​​and outliers, and perform data standardization to obtain a standardized multidimensional data set; For the standardized multidimensional data set, the principal component analysis method is used for feature extraction. By calculating the eigenvalues ​​and eigenvectors of the covariance matrix, the first k principal components with larger eigenvalues ​​are selected as the new feature space to obtain the feature data set after dimensionality reduction. The feature data set after dimensionality reduction is used as input, and a credit rating prediction model is constructed using the random forest algorithm. The importance score of each feature to the rating result is calculated through ensemble learning of decision trees. According to the ranking results of feature importance scores, the first n features with higher scores are selected as key feature factors to construct the credit assessment feature vector; Continuously collect building operation data and evaluation feedback data, and regularly update the feature vector. If the amount of evaluation feedback data reaches the preset threshold, the feature vector update process is triggered to recalculate the feature importance score and update the key feature factor. The building operation data is mined and analyzed to identify potential feature factors. If the correlation between the new feature and the existing feature vector is lower than the preset threshold, it will be included in the feature pool.

2. A building low-carbon operation management method based on data analysis as claimed in claim 1, characterized in that: Obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, and determine whether there are missing values ​​and outliers in the data. If there are missing values ​​and outliers, remove the missing values ​​and outlier data; According to the pre-processed building energy consumption data, carbon emission data and energy-saving measures data, the data of each dimension is standardized by using the data standardization algorithm to obtain a standardized multi-dimensional data set; For the standardized multidimensional data set, the principal component analysis algorithm is used to extract the main features of the data set, reduce the data dimension, and obtain the reduced-dimensional data set; According to the data set after dimensionality reduction, the K-means clustering algorithm is used to perform cluster analysis on the data, and the data is divided into several clusters according to the clustering results to obtain a cluster set; For each cluster, an energy consumption prediction model is constructed based on building energy consumption data, carbon emission data and energy-saving measures data to predict the energy consumption and carbon emission of the building in the future. According to the energy consumption prediction results and carbon emission prediction results of the clusters, the energy-saving potential of each cluster is judged, the priority of energy-saving transformation is determined, and the energy-saving transformation plan is obtained; For the energy-saving transformation plan, a multi-objective optimization algorithm is used to optimize the parameters of energy-saving measures and obtain the optimal energy-saving transformation implementation plan.

3. A building low-carbon operation management method based on data analysis as claimed in claim 2, characterized in that: According to the standardized multidimensional data set, a covariance matrix is ​​constructed, and the covariance matrix is ​​decomposed using a singular value decomposition algorithm to obtain eigenvalues ​​and corresponding eigenvectors; According to the size of the eigenvalue, sort the eigenvalues ​​from large to small, and select the eigenvectors corresponding to the first k largest eigenvalues ​​as the principal component vectors; For the standardized multidimensional data set, an inner product operation is performed on each data sample and the principal component vector to obtain the projection coordinates of the sample in the principal component space; According to the corresponding relationship between the projection coordinates and the principal component vector, the coordinate representation of each sample in the principal component space is obtained to obtain the feature data set after dimensionality reduction; For the feature data set after dimensionality reduction, the k-means clustering algorithm is used to cluster the data. The position of each cluster center is optimized iteratively to minimize the sum of square distances from the sample to the cluster center until the cluster center no longer changes, and the final clustering result is obtained. According to the clustering results, samples of different categories are distinguished. For each category, the mean vector of samples within the category is calculated as the representative vector of the category; According to the distance relationship between representative vectors, the distance matrix between categories is constructed, and the representative vectors are clustered using a hierarchical clustering algorithm to obtain the hierarchical relationship between categories, thereby realizing the high-level semantic representation of the standardized multidimensional data set.

4. A building low-carbon operation management method based on data analysis as claimed in claim 3, characterized in that: Obtain raw data related to credit ratings, preprocess the data, remove noise and outliers, and perform feature selection and feature extraction; Use principal component analysis and other dimensionality reduction algorithms to reduce the dimensionality of feature data and obtain a set of feature data after dimensionality reduction; The feature data set after dimension reduction is divided into a training set and a test set. The training set is used to train the random forest model, and the test set is used to evaluate the model performance. According to the preset random forest parameters, a random forest model is constructed, and the model is trained using the training set data to obtain a trained random forest model; Input the test set data into the trained random forest model to perform credit rating prediction and obtain the prediction results; The importance score of each feature to the credit rating prediction result is calculated through the feature importance evaluation algorithm, and the features are sorted according to the importance score; According to the feature importance ranking results, features with higher importance scores are selected as key features for credit rating prediction.

5. A building low-carbon operation management method based on data analysis as claimed in claim 4, characterized in that: Obtain original feature data related to credit assessment and preprocess the features, including missing value processing, outlier processing and data normalization; A machine learning algorithm is used to train the credit assessment model based on the preprocessed feature data, and the importance score of each feature is calculated through the feature importance assessment method; Sort all features in descending order according to their importance scores, and select the top n features with higher scores as key feature factors; Based on the selected key feature factors, a credit assessment feature vector is constructed to form an optimized feature subset; The credit assessment model is trained based on the optimized feature vector to obtain the final credit risk prediction result; Compare the credit risk prediction results with the preset risk level threshold, determine the user's credit risk level, and output a credit assessment report.

6. A building low-carbon operation management method based on data analysis as claimed in claim 5, characterized in that: Obtain building operation data and evaluation feedback data, clean and standardize the data according to the preset data format and data standards, and obtain standardized building operation data and evaluation feedback data; For the standardized building operation data and evaluation feedback data, the principal component analysis algorithm is used to extract the main features of the data and obtain the feature vector representing the main features of the data; According to the feature vector, the importance score of each feature is calculated using the random forest algorithm, and the key feature factors are determined according to the importance score sorting; Continuously obtain the latest building operation data and evaluation feedback data to determine whether the data volume reaches the preset threshold. If the threshold is reached, the feature vector update process is triggered; In the feature vector update process, the latest building operation data and evaluation feedback data are merged with the original data, and the principal component analysis is re-performed to obtain the updated feature vector; According to the updated feature vector, the importance score of each feature is recalculated, and the key feature factors are updated according to the importance score sorting.

7. A building low-carbon operation management method based on data analysis as claimed in claim 6, characterized in that: Obtain building operation data and continuously analyze the data using data mining algorithms to identify new potential characteristic factors; Calculate the correlation between the new feature and the existing feature vector. If the correlation is lower than the preset threshold, the new feature is included in the feature pool. Based on the enriched feature pool, a credit assessment model is constructed. Obtain changes in business requirements. If new business requirements emerge, the model update mechanism is triggered; In response to new business needs, we continue to mine new potential feature factors from building operation data; The newly mined feature factors are correlated with the existing feature pool. If the correlation is lower than the preset threshold, they are included in the feature pool and the credit assessment model is retrained using the updated feature pool.

8. A building low-carbon operation management system based on data analysis, characterized in that: include: Data acquisition and processing module: obtain building energy consumption data, carbon emission data and energy-saving measures data, pre-process the data, remove missing values ​​and outliers, and perform data standardization to obtain a standardized multidimensional data set; Feature data set acquisition module: For the standardized multidimensional data set, the principal component analysis method is used to extract features. By calculating the eigenvalues ​​and eigenvectors of the covariance matrix, the first k principal components with larger eigenvalues ​​are selected as the new feature space to obtain the feature data set after dimensionality reduction; Feature scoring module: takes the reduced feature data set as input, uses the random forest algorithm to build a credit rating prediction model, and calculates the importance score of each feature to the rating result through ensemble learning of decision trees; Feature vector construction module: According to the feature importance score sorting results, the top n features with higher scores are selected as key feature factors to construct the credit assessment feature vector; Key factor screening module: continuously collects building operation data and evaluation feedback data, and regularly updates the feature vector. If the amount of evaluation feedback data reaches the preset threshold, the feature vector update process is triggered, the feature importance score is recalculated, and the key feature factor is updated; Feature factor update module: mines and analyzes building operation data to identify potential feature factors. If the correlation between the new feature and the existing feature vector is lower than the preset threshold, it will be included in the feature pool.

9. An electronic device, characterized in that: The invention comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, a data analysis-based low-carbon operation and management method for a building is implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for low-carbon building operation and management based on data analysis as described in any one of claims 1 to 7 is implemented.

Images