A construction project planning site selection and land use preliminary evaluation method and system
Through multi-source data acquisition and deep learning algorithms, site selection optimization and dynamic prediction models are constructed, combined with expert systems, and comprehensive evaluation is carried out, which solves the problems of data integration and model fusion in the existing technology, and realizes scientific and accurate decision-making support for site selection of construction projects. .
Patent Information
- Application Number
- CN202410996034.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-07-24
AI Technical Summary
The existing technology faces the problems of different formats and uneven quality in the planning and site selection and land use pre-examination of construction projects, and it is difficult to build a scientific and reasonable site selection optimization model and dynamic prediction model. There are challenges in the integration and comprehensive evaluation of output results of different models.
The basic data of construction project site selection is collected through multiple data sources, preprocessing and deep learning algorithm mining, multi-dimensional site selection optimization model and dynamic prediction model, and comprehensive evaluation is used for expert systems, fuzzy logic and neural networks to generate visualization Site selection recommendation report.
It realizes scientific, accurate and comprehensive evaluation of site selection plans, improves the scientificity and efficiency of decision-making, provides flexible and humanized decision-making assistance, and enhances the forward-looking and robustness of site selection plans.
Smart Images

Figure CN118521012B_ABST
Abstract
Description
Technical Field
[0001] The invention proposes a construction project planning site selection and land use pre-examination evaluation method and system, belonging to the technical field of national land space planning and natural resource management. Background Art
[0002] In recent years, with the widespread application of technologies such as big data, cloud computing, and artificial intelligence, new ideas and methods have been provided for construction project planning, site selection, and land use pre-examination. By integrating multi-source heterogeneous data, such as geospatial data, socioeconomic data, and environmental monitoring data, a comprehensive and detailed basic database for site selection can be constructed to provide solid data support for scientific decision-making. At the same time, the application of advanced algorithms such as deep learning and machine learning makes it possible to conduct in-depth mining and intelligent analysis of massive data, automatically extract key indicators, assess potential risks, and predict future trends, thereby significantly improving the scientificity and accuracy of site selection decisions.
[0003] However, existing technologies still face many challenges in practical applications. On the one hand, the collection, integration and processing of multi-source data need to solve problems such as different data formats and uneven quality; on the other hand, how to build a scientific and reasonable site selection optimization model and dynamic prediction model based on the specific characteristics and construction goals of the project, and how to effectively integrate and comprehensively evaluate the output results of different models are still technical problems that need to be solved urgently. Summary of the invention
[0004] The present invention provides a construction project planning site selection and land use pre-evaluation method and system to solve the problems mentioned in the above background technology:
[0005] The present invention provides a construction project planning site selection and land use pre-evaluation method, the method comprising:
[0006] Collecting basic data of multiple construction project site selection through multiple data sources, storing the basic data in cloud space, screening out a subset of the basic data of project site selection, and preprocessing the data subset;
[0007] Through deep learning algorithms, the pre-processed data subsets are deeply mined to determine the weight of each indicator. Based on the project characteristics and construction goals, a multi-dimensional indicator site selection optimization model is constructed. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set.
[0008] Based on basic data, a dynamic prediction model is constructed, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the dynamic prediction model is used to predict the performance of site selection plans under different scenarios;
[0009] The output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through expert system, fuzzy logic and neural network to generate a visual site selection recommendation evaluation report.
[0010] Furthermore, the basic data of site selection for multiple construction projects are collected through multiple data sources, and the basic data are stored in a cloud space, and a subset of the basic data of project site selection is screened out, and the data subset is preprocessed, including:
[0011] Determine the basic data that needs to be collected, and select a data source to collect the basic data based on the basic data that needs to be collected;
[0012] Determine the frequency and time period of collection, store the collected basic data in a cloud space, and the cloud space receives and processes the basic data;
[0013] The cloud space stores the processed basic data into a structured database;
[0014] According to the requirements of project site selection, a subset of project site selection data is screened out, and the data subset is preprocessed, and the preprocessing includes data cleaning, data conversion, feature engineering and data integration.
[0015] Furthermore, the basic data include natural geographic data and ecological environment data; the data sources include integrated geographic information systems, remote sensing satellites, the Internet and sensor networks.
[0016] Furthermore, the determining of the frequency and time period of collection, storing the collected basic data in a cloud space, the cloud space receiving the basic data, and processing the basic data include:
[0017] Determine the basic data to be collected and the frequency of collection according to project requirements and data timeliness; Collect basic data according to the basic data to be collected and the frequency of collection;
[0018] The collected basic data are processed accordingly, and the results of the corresponding processing are transmitted to the cloud space respectively; after the cloud space receives the corresponding processed basic data, the received data is processed again.
[0019] Furthermore, the collected basic data are processed accordingly, and the results of the corresponding processing are transmitted to the cloud space respectively; after the cloud space receives the corresponding processed basic data, the received data is processed again, including:
[0020] Processing the collected basic data respectively, including natural geographical data processing and ecological environment data processing;
[0021] The results of the corresponding processing of the basic data are compressed respectively and transmitted to the cloud space through a multi-channel transmission protocol, and the cloud space processes the received data again.
[0022] Furthermore, the pre-processed data subset is deeply mined by the deep learning algorithm to determine the weight of each indicator, and a multi-dimensional indicator site selection optimization model is constructed based on the project characteristics and construction goals. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set, including:
[0023] Use deep learning algorithms to extract features from preprocessed data subsets, and select feature subsets based on the importance of features and the degree of influence on the site selection of construction projects;
[0024] Evaluate the importance of indicators; determine the weight of each indicator based on the importance of the evaluation indicators; define a multi-dimensional site optimization objective function based on the construction goals;
[0025] According to business needs and project characteristics, set the constraints of the site selection optimization model, integrate the objective function, constraints and evaluation index system, and build a site selection optimization model;
[0026] Using multi-objective optimization algorithm, according to the constraints of the site selection optimization model, the Pareto optimal solution set of each evaluation index is obtained;
[0027] The obtained Pareto optimal solution set is interpreted and analyzed, and the recommended site locations and the scores of the recommended site locations on various evaluation indicators are extracted.
[0028] Furthermore, the method of constructing a dynamic prediction model based on basic data, presetting development scenarios, inputting the preset development scenarios into the dynamic prediction model, and using the dynamic prediction model to predict the performance of site selection plans under different scenarios includes:
[0029] Obtain basic data, clean the basic data, integrate data from different data sources into a unified data set, extract features, and build a dynamic prediction model;
[0030] The data set is divided into a training set, a test set and a validation set. The dynamic prediction model is trained with the training set, tested with the test set and predicted with the validation set.
[0031] According to the actual situation and predicted needs, set different development scenarios and set relevant parameters under each development scenario;
[0032] Generate corresponding scenario data according to the preset development scenario, input the generated scenario data into the dynamic prediction model to predict the site selection plan;
[0033] Analyze the predicted results of site selection plans under different scenarios, and evaluate their feasibility and advantages and disadvantages; and optimize and adjust the site selection plan based on the predicted results and actual needs.
[0034] Furthermore, the output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through an expert system, fuzzy logic and neural network to generate a visual site selection recommendation evaluation report, including:
[0035] Integrate the output results of the site optimization model and the dynamic prediction model to obtain an integrated data set, and perform feature engineering on the integrated data set;
[0036] Select a comprehensive evaluation model and train it by integrating the data set; use the cross-validation method to verify the accuracy and generalization ability of the comprehensive evaluation model;
[0037] Use the trained comprehensive evaluation model to evaluate different site selection options and generate evaluation results; and visualize the evaluation results.
[0038] The present invention proposes a system for implementing any of the above-mentioned methods for site selection and land use preliminary evaluation of construction project planning, the system comprising:
[0039] Data acquisition module: collects basic data of site selection for multiple construction projects through multiple data sources, stores the basic data in the cloud space, selects a subset of basic data for project site selection, and pre-processes the data subset;
[0040] Site selection optimization module: The pre-processed data subset is deeply mined through deep learning algorithms to determine the weight of each indicator. Based on the project characteristics and construction goals, a site selection optimization model with multi-dimensional indicators is constructed. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set.
[0041] Dynamic prediction module: Based on basic data, a dynamic prediction model is constructed, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the dynamic prediction model is used to predict the performance of site selection plans under different scenarios;
[0042] Fusion module: Fusion the output results of the site selection optimization model and the dynamic prediction model, conduct comprehensive evaluation through expert system, fuzzy logic and neural network, and generate a visual site selection recommendation evaluation report.
[0043] Beneficial effects of the present invention: The site selection evaluation method proposed in the present invention makes full use of multiple data sources, including integrated geographic information systems, remote sensing satellites, the Internet and sensor networks, and can collect all-round basic data covering natural geography. Data is collected and stored in cloud space in a timely and quantitative manner, ensuring the security and convenience of the data. At the same time, the quality and availability of the original data are improved through data cleaning, conversion, feature engineering and data integration in the preprocessing stage; the deep learning algorithm is used to perform deep mining on the preprocessed data subset, which can automatically extract key features and quantify indicator weights, and help to objectively and impartially reflect the importance of various factors to site selection. By constructing a site selection optimization model with multi-dimensional indicators and combining a multi-objective optimization algorithm to solve the Pareto optimal solution set, it can effectively balance multiple conflicting goals, thereby obtaining the most cost-effective and adaptable site selection plan; the constructed dynamic prediction model can simulate and predict future development trends based on basic data, and consider the impact of different preset development scenarios on the site selection plan. This not only enhances the foresight of the site selection plan, but also can cope with the challenges brought by future uncertainties and ensure the robustness and adaptability of the site selection strategy; the results of the site selection optimization model and the dynamic prediction model are deeply integrated and comprehensively evaluated through expert systems, fuzzy logic and neural networks, which not only ensures the scientificity and accuracy of the evaluation results, but also provides flexible and humanized decision-making assistance. The final generated visual site selection recommendation evaluation report is conducive to decision makers to intuitively understand the advantages and disadvantages of each site selection plan and make scientific decisions quickly. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 It is a step diagram of the method of the present invention;
[0045] Figure 2 This is a system module diagram of the present invention. DETAILED DESCRIPTION
[0046] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0047] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. The embodiments described are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.
[0049] One embodiment of the present invention, as Figure 1 As shown, a construction project planning site selection and land use preliminary evaluation method, the method comprising:
[0050] S1: Collect basic data of site selection for multiple construction projects through multiple data sources, store the basic data in cloud space, filter out a subset of basic data for project site selection, and pre-process the data subset;
[0051] S2: Use deep learning algorithms to conduct in-depth mining on the preprocessed data subsets, determine the weights of each indicator, and build a multi-dimensional indicator site selection optimization model based on project characteristics and construction goals. Input data into the site selection optimization model to obtain the Pareto optimal solution set;
[0052] S3: Based on the basic data, a dynamic prediction model is constructed, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the dynamic prediction model is used to predict the performance of the site selection scheme under different scenarios;
[0053] S4: The output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through expert system, fuzzy logic and neural network to generate a visual site selection recommendation evaluation report.
[0054] The working principle of the above technical solution is: using multiple data sources, such as integrated geographic information system (GIS), remote sensing satellites and Internet of Things devices, to collect basic data required for the site selection of multiple construction projects; the basic data includes data on natural geography (such as topography, climate conditions, geological structure) and ecological environment. The collected data needs to be preprocessed, including data cleaning, denoising and format conversion, to ensure the accuracy and availability of the data. A deep learning algorithm is used to deeply mine the preprocessed data to extract high-level features that reflect the key influencing factors of the site selection of construction projects; in this process, the weights of each indicator are determined, and a multi-dimensional site selection optimization model is constructed according to the characteristics and construction goals of the project. The preprocessed data subset is input into the site selection optimization model, and the Pareto optimal solution set is obtained using the model. The optimal solution set includes recommended site selection locations and their scores on various evaluation indicators, as well as multi-dimensional indicators. Based on the collected basic data, a dynamic prediction model is constructed to preset different development scenarios. These preset development scenarios are input into the dynamic prediction model, and the model is used to predict the performance of the site selection scheme under different scenarios. The output results of the site selection optimization model and the dynamic prediction model are fused. Comprehensive evaluation is performed using expert systems, fuzzy logic, neural networks and other methods to generate a visual site selection recommendation evaluation report. This report can provide decision makers with intuitive and comprehensive site selection recommendations and help them make the most appropriate decision.
[0055] The effects of the above technical solutions are as follows: through the collection and comprehensive utilization of multiple data sources, including GIS, remote sensing satellites and Internet of Things devices, comprehensive basic data on the site selection of construction projects are obtained, covering information on multiple aspects such as natural geography and ecological environment, providing rich data support for site selection evaluation; deep learning algorithms are used to deeply mine the data, extract advanced features that reflect the key influencing factors of site selection, and construct a site selection optimization model of multi-dimensional indicators by determining the weights of each indicator, providing scientific analysis methods and tool support for site selection decisions; a dynamic prediction model is constructed based on basic data, which can preset different development scenarios and predict the performance of site selection plans under different scenarios, thereby improving the flexibility and pertinence of site selection decisions; the output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through expert systems, fuzzy logic and neural networks, etc., to generate a visual site selection recommendation evaluation report, providing decision makers with intuitive and comprehensive site selection recommendations, and improving the comprehensiveness, accuracy and scientificity of site selection evaluation through the comprehensive utilization of multiple data sources and advanced algorithm technologies, providing decision makers with more diversified and reliable site selection decision support, which helps to improve decision efficiency and reduce decision risks.
[0056] In one embodiment of the present invention, the basic data of site selection for multiple construction projects are collected through multiple data sources, the basic data are stored in a cloud space, a subset of the basic data of project site selection is screened out, and the data subset is preprocessed, including:
[0057] Determine the basic data that needs to be collected, and select a data source to collect the basic data based on the basic data that needs to be collected;
[0058] Determine the frequency and time period of collection, store the collected basic data in a cloud space, and the cloud space receives and processes the basic data;
[0059] The cloud space stores the processed basic data into a structured database;
[0060] According to the requirements of project site selection, a subset of project site selection data is screened out, and the data subset is preprocessed, and the preprocessing includes data cleaning, data conversion, feature engineering and data integration.
[0061] The basic data include natural geographic data and ecological environment data; the data sources include integrated geographic information systems, remote sensing satellites, the Internet and sensor networks.
[0062] The working principle of the above technical solution is: determine the basic data required for the site selection of the construction project, and select the appropriate data source for collection based on the determined basic data requirements. For example, use the integrated geographic information system to obtain natural geographic data, and use the sensor network to obtain ecological environment data; determine the frequency and time period of data collection based on the data update frequency and the needs of site selection decisions; for example: natural geographic data:
[0063] Frequency: Once a month,
[0064] Time period: early to mid-month,
[0065] Reason: Physical geographic data changes relatively slowly, and collecting them monthly can capture seasonal and long-term changes while avoiding waste of resources caused by frequent collection.
[0066] Ecological environment data:
[0067] Frequency: Once a week,
[0068] Time period: Every weekend,
[0069] Reason: Ecological and environmental data are greatly affected by seasons and climate. Collecting data once a week can better capture short-term changes and detect environmental anomalies in a timely manner.
[0070] The collected basic data is stored in the cloud space to ensure data security and accessibility; after receiving the collected basic data, the cloud space performs necessary data processing, including data cleaning, conversion and integration, to ensure data quality and consistency; the processed basic data is stored in a structured database for subsequent data analysis and mining. Commonly used databases include MySQL, PostgreSQL, etc.; according to the site selection requirements, the data subset related to the site selection is selected from the entire basic data set; the site selection data subset is further preprocessed, including data cleaning, conversion, feature engineering and data integration, to prepare for subsequent site selection analysis and modeling.
[0071] The effects of the above technical solutions are as follows: by collecting basic data through multiple data sources, more comprehensive and accurate information can be obtained, thereby improving the scientificity and reliability of site selection decisions; the collected basic data are stored in cloud space, which has high availability, scalability and security, and is convenient for data management and access; the collected basic data are processed and preprocessed, including cleaning, conversion, feature engineering and data integration, which improves the quality and availability of the data and provides a good data foundation for subsequent site selection analysis and modeling; the processed basic data are stored in a structured database, which facilitates data management, query and analysis, and improves the efficiency of data organization and retrieval; according to the needs of project site selection, the data subsets related to the site selection are screened out, which reduces the redundancy and noise of the data and improves the efficiency and accuracy of the site selection analysis; the comprehensive use of multiple data sources and processed data subsets provides a scientific basis and technical support for project site selection decisions, helps reduce site selection risks, and improves the success rate and benefits of projects.
[0072] In one embodiment of the present invention, determining the frequency and time period of collection, storing the collected basic data in a cloud space, and the cloud space receiving the basic data and processing the basic data include:
[0073] Determine the basic data to be collected and the frequency of collection according to project requirements and data timeliness; Collect basic data according to the basic data to be collected and the frequency of collection;
[0074] The collected basic data are processed accordingly, and the results of the corresponding processing are transmitted to the cloud space respectively; after the cloud space receives the corresponding processed basic data, the received data is processed again.
[0075] Determine the basic data to be collected and the collection frequency according to the project requirements and data timeliness; collect the basic data according to the basic data to be collected and the collection frequency, including:
[0076] Processing the collected basic data respectively, including natural geographical data processing and ecological environment data processing;
[0077] The results of the corresponding processing of the basic data are compressed respectively, and transmitted to the cloud space through a multi-channel transmission protocol. The cloud space processes the received data again, and the re-processing includes decompression.
[0078] The working principle of the above technical solution is: according to the project requirements and data timeliness, determine the basic data types that need to be collected, as well as the collection frequency and time period; according to the determined needs, collect basic data. The collection process may involve equipment and technologies such as sensors and data interfaces.
[0079] Physical geographic data processing: including data cleaning, integration and spatial analysis processing, such as terrain analysis, landscape analysis and spatial interpolation.
[0080] Ecological environment data processing: Data is collected through sensors and transmitted to edge devices for preliminary processing, such as data cleaning and simple analysis.
[0081] The processed basic data are transmitted to the cloud space separately, using a multi-channel transmission protocol, and the transmitted data is compressed to reduce transmission time and cost; after the cloud space receives the transmitted data, it is processed again, which may include decompression, data format conversion, data integration and other operations to ensure the integrity and availability of the data.
[0082] The effects of the above technical solutions are: according to project requirements and data timeliness, determine the basic data to be collected and the collection frequency to ensure that the collected data meets the project requirements and improve the validity and availability of the data; comprehensively process the collected basic data, including natural geographic data and ecological environmental data, to ensure the accuracy and completeness of the data; perform spatial analysis and processing of natural geographic data, including terrain analysis, landscape analysis and spatial interpolation, to provide support for subsequent spatial analysis and decision-making; use edge devices to perform preliminary processing of ecological environmental data to improve data transmission efficiency and reduce the cost and time of data transmission; use multi-channel transmission protocols and data compression technology to improve the efficiency and stability of data transmission and ensure the timeliness and integrity of data; the cloud space re-processes the received data, including decompression and data format conversion, to improve the quality and availability of data.
[0083] In one embodiment of the present invention, the deep learning algorithm is used to perform deep mining on the preprocessed data subset, determine the weight of each indicator, and build a multi-dimensional indicator site selection optimization model based on project characteristics and construction goals, input data into the site selection optimization model, and obtain the Pareto optimal solution set, including:
[0084] Use deep learning algorithms to extract features from preprocessed data subsets, and select feature subsets based on the importance of features and the degree of influence on the site selection of construction projects;
[0085] Assess the importance of indicators; determine the weight of each indicator based on the importance of the evaluation indicators; define a multi-dimensional site selection optimization objective function based on the construction goals; the objective function takes into account various evaluation indicators and their weights; for example, the weight of the ecological protection red line is 0.15, and if it meets the requirements, it will be above 85 points, if it basically meets the requirements, it will be 60-85 points, and if it does not meet the requirements, it will be below 60 points.
[0086] According to business needs and project characteristics, set the constraints of the site selection optimization model, integrate the objective function, constraints and evaluation index system, and build a site selection optimization model;
[0087] Using multi-objective optimization algorithm, according to the constraints of the site selection optimization model, the Pareto optimal solution set of each evaluation index is obtained;
[0088] The obtained Pareto optimal solution set is interpreted and analyzed, and the recommended site locations and the scores of the recommended site locations on various evaluation indicators are extracted.
[0089] The working principle of the above technical solution is: use deep learning algorithms to extract features from preprocessed data subsets and select according to the importance of features. It is implemented through deep learning models such as convolutional neural networks and recurrent neural networks to identify key features in the data; based on the characteristics and goals of the site selection project, the importance of each indicator is evaluated and the weight of each indicator is determined. It is achieved by analyzing the contribution of indicators to the success of the project; according to the construction goals, the weight of each indicator is taken into account to define a multi-dimensional site selection optimization objective function. According to business needs and project characteristics, set constraints for the site selection optimization model, such as technical feasibility. These constraints can limit the scope of site selection and ensure the feasibility of the site selection plan; comprehensively consider the objective function, constraints and evaluation indicator system to build a site selection optimization model. This can be a mathematical model for determining the best site selection plan; use multi-objective optimization algorithms, such as genetic algorithms, particle swarm optimization algorithms, etc., to obtain the Pareto optimal solution set of each evaluation indicator under the premise of considering constraints. These algorithms can find a balance between multiple evaluation indicators and come up with the optimal site selection plan; interpret and analyze the obtained Pareto optimal solution set, and extract recommended site locations and their scores on various evaluation indicators.
[0090] The effects of the above technical solution are: by mining data features through deep learning algorithms, combined with the evaluation of indicator importance and weight determination, and multi-dimensional site optimization objective functions, the advantages and disadvantages of potential sites can be more accurately identified and evaluated; this technical solution takes into account multiple factors such as environmental benefits and risk control, making site selection decisions more comprehensive and integrated; using multi-objective optimization algorithms, it is possible to obtain the Pareto optimal solution set of various evaluation indicators under the premise of considering constraints, thereby improving the efficiency and accuracy of site selection decisions; interpreting and analyzing the obtained Pareto optimal solution set can clearly explain why certain locations are recommended, enhancing the interpretability and credibility of the decision; by considering risk control factors, the technical solution can help reduce the risks of site selection decisions and increase the possibility of project success.
[0091] In one embodiment of the present invention, the method of extracting features from a preprocessed data subset using a deep learning algorithm and selecting a feature subset based on the importance of the features and the degree of influence on the site selection of a construction project includes:
[0092] The outliers in the data set are detected using the isolation forest algorithm. For the detected outliers, they are retained, deleted, or replaced according to their business significance. If the outliers have practical significance (such as data under extreme weather conditions), they are retained and marked separately. Otherwise, they are replaced using the mean, median, or interpolation method based on neighboring points.
[0093] For numerical features that obey normal distribution or approximate normal distribution, use the Z-score standardization method, subtract the mean from the feature value and divide it by the standard deviation, so that the data distribution has zero mean and unit variance;
[0094] For numerical features with different dimensions or ranges, the minimum-maximum normalization method is used to scale the feature values to the interval [0,1] to eliminate the impact of the dimension on model training;
[0095] Analyze the distribution and pattern of missing values to determine whether they are missing at random, missing completely at random, or missing not at random;
[0096] A filling strategy is selected according to the nature of the missing values and the data structure. The filling strategies include: for time series data, forward filling, backward filling or interpolation may be used; for categorical variables, mode filling may be used; for numerical variables, mean or median filling or prediction filling based on machine learning methods such as K nearest neighbors and random forests may be used.
[0097] Select an appropriate deep learning model based on the data type and business scenario. For example, for image data (such as satellite image analysis and site selection), a convolutional neural network (CNN) can be used; for time series data (such as historical meteorological data), a long short-term memory network (LSTM) or a gated recurrent unit (GRU) can be used.
[0098] Automatically extract high-level abstract features from raw data by training deep learning models;
[0099] Using Gradient Boosted Tree (GBDT) and Random Forest (RF), combined with the features extracted by the deep learning model, the importance score of each feature in the model prediction is calculated;
[0100] All features extracted by the deep learning model are used as the initial feature set, and an ensemble learning algorithm (such as GBDT) is used as the base model. In each iteration, the least important features are removed from the feature set according to the feature importance score, and the model is retrained. This process is repeated until the preset number of features is reached or the model performance is no longer significantly improved.
[0101] Fuse data from different sources to generate new composite features, explore the interactions between features, and capture the nonlinear relationship between features by generating feature interaction terms (such as products, polynomials, etc.);
[0102] The K-fold cross-validation method is used to evaluate the impact of different feature subsets on model performance. The data set is divided into K subsets, and K-1 subsets are used for training each time, and the remaining subset is used for testing. This process is repeated K times, and the average performance indicators (such as accuracy, recall, F1 score, etc.) are calculated.
[0103] The optimal feature subset is selected based on the cross-validation results.
[0104] The working principle of the above technical solution is as follows: Isolation forest algorithm is used to identify outliers in the data set. The processing method of outliers is determined according to their business meaning, such as retaining and marking, deleting or replacing. The replacement strategy may include using the mean, median or interpolation method based on neighboring points to ensure the integrity and accuracy of the data; for numerical features, appropriate standardization or normalization methods are selected according to their distribution characteristics. Standardization is applicable to data with normal distribution or approximate normal distribution, and the dimension effect is eliminated by subtracting the mean and dividing by the standard deviation; normalization is applicable to data with different dimensions or ranges, scaling it to the interval [0,1]; analyze the distribution and pattern of missing values and select appropriate filling strategies. For time series data, forward filling, backward filling or interpolation may be used; for categorical variables and numerical variables, mode filling and mean, median filling, or prediction filling based on machine learning methods such as K nearest neighbors and random forests are used respectively; select appropriate deep learning models according to data types and business scenarios. For example, CNN is used for image data, and LSTM or GRU is used for time series data. These models can automatically extract high-level abstract features from raw data and capture complex patterns and relationships in the data. By training deep learning models and optimizing model parameters using algorithms such as back propagation, the models can extract features useful for site selection decisions from raw data. Ensemble learning algorithms such as GBDT and RF are used to combine the features extracted by deep learning models to calculate the importance score of each feature in model prediction. These scores reflect the contribution of features to model performance. Using ensemble learning algorithms (such as GBDT) as the base model, the least important features are gradually removed in an iterative manner, and the model is retrained to evaluate the performance of the remaining feature sets. This process is repeated until the preset number of features is reached or the model performance is no longer significantly improved. Data from different sources are fused to generate new composite features. At the same time, the interactions between features are explored, and nonlinear relationships are captured by generating feature interaction terms to further improve the model's prediction and generalization capabilities. The K-fold cross-validation method is used to evaluate the impact of different feature subsets on model performance. By dividing the data set into multiple subsets and taking turns as training and test sets, the average performance index is calculated to evaluate the effectiveness of feature subsets. The optimal feature subset is selected based on the cross-validation results. This subset can reduce computational costs and overfitting risks while maintaining high prediction accuracy, providing the best feature combination for subsequent site selection optimization model construction.
[0105] The effects of the above technical solutions are as follows: outliers are detected and processed by the isolation forest algorithm to ensure that the data used in the model training process is cleaner and more accurate. Retaining outliers with practical significance and marking them separately helps the model learn a more comprehensive data distribution and improve the generalization ability of the model. Standardizing and normalizing numerical features eliminates the impact of different dimensions and ranges on model training, allowing the model to more fairly evaluate the importance of each feature. Reasonable filling of missing values reduces the impact of missing data on model performance, and selecting appropriate filling strategies according to different types of data further improves data quality. High-level abstract features are automatically extracted from raw data using deep learning models. These features can capture complex patterns and relationships in the data, avoiding the tediousness and subjectivity of manually designing features in traditional methods. Combined with ensemble learning algorithms (such as GBDT, RF) to calculate feature importance scores, and select features based on the scores, effectively reducing feature dimensions and improving model training efficiency and prediction performance. At the same time, the feature subsets are gradually optimized by the recursive feature elimination (RFE) method to ensure that the final selected feature subsets are both concise and effective. The data from different sources are integrated to generate new composite features, which enriches the feature dimension and improves the modeling ability of the model for complex problems. The interaction between features is explored, and the nonlinear relationship between features is captured by generating feature interaction terms, which further improves the prediction accuracy and interpretability of the model. The K-fold cross-validation method is used to evaluate the impact of different feature subsets on model performance, ensuring the fairness and reliability of model evaluation. Through multiple training and testing, the optimal feature subset can be selected more accurately. The selection of the optimal feature subset not only improves the prediction performance of the model (such as the improvement of indicators such as accuracy, recall, and F1 score), but also reduces the computational cost and overfitting risk of the model, making the model more efficient and stable in practical applications. The combination of deep learning models and ensemble learning algorithms not only improves the prediction ability of the model, but also provides an in-depth understanding of the importance of features. This helps business personnel understand which factors have the greatest impact on the site selection of construction projects, so as to make more scientific and reasonable decisions. At the same time, the generated composite features and feature interaction terms also provide business personnel with more business insights, helping them to discover new market opportunities or potential risks.
[0106] In one embodiment of the present invention, the dynamic prediction model is constructed based on basic data, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the performance of site selection schemes under different scenarios is predicted using the dynamic prediction model, including:
[0107] Obtain basic data, clean the basic data, integrate data from different data sources into a unified data set, extract features, and build a dynamic prediction model;
[0108] The data set is divided into a training set, a test set and a validation set. The dynamic prediction model is trained with the training set, tested with the test set and predicted with the validation set.
[0109] According to the actual situation and predicted needs, set different development scenarios and set relevant parameters under each development scenario;
[0110] Generate corresponding scenario data according to the preset development scenario, input the generated scenario data into the dynamic prediction model to predict the site selection plan;
[0111] Analyze the predicted results of site selection plans under different scenarios, and evaluate their feasibility and advantages and disadvantages; and optimize and adjust the site selection plan based on the predicted results and actual needs.
[0112] The working principle of the above technical solution is as follows: first, obtain the necessary basic data from various data sources, then clean and preprocess these data to ensure the integrity and accuracy of the data; based on the cleaned data, extract and select features to build a dynamic model suitable for prediction, which may adopt time series model, machine learning model, etc. These models will be used to predict future development scenarios; divide the data set into training set, test set and validation set, use the training set to train the dynamic prediction model, evaluate and tune the model through the test set, and finally use the validation set to verify and confirm the model; according to actual needs and prediction goals, set different development scenarios and determine the relevant parameters under each scenario. Then generate corresponding scenario data based on these parameters, including data such as land use at future time points; input the generated scenario data into the dynamic prediction model, use the model to predict the site selection scheme under different scenarios, and obtain the site selection scheme performance under various scenarios; analyze and compare the prediction results of the site selection scheme under different scenarios, evaluate its feasibility and advantages and disadvantages, and find the optimal site selection scheme; optimize and adjust the site selection scheme based on the prediction results and actual needs to make it more in line with the actual situation and expected goals.
[0113] The effects of the above technical solutions are as follows: by establishing a dynamic prediction model and combining different development scenarios, scientific site selection plan predictions and evaluations can be provided to decision makers to help them make more accurate and reliable decisions; by predicting and evaluating site selection plans under different scenarios, potential risk factors can be identified and corresponding avoidance strategies can be taken to reduce uncertainty and risks during project implementation; by analyzing development trends and demand changes under different scenarios, resources can be better planned and utilized, and resource utilization efficiency can be improved; by predicting and evaluating the performance of different site selection plans, the optimal plan can be selected to avoid unnecessary investment and waste of resources, thereby achieving cost savings and maximizing benefits; through the application of automated data processing and prediction models, the efficiency and accuracy of decision-making can be greatly improved, the interference of human subjective factors can be reduced, and the decision-making process can be accelerated.
[0114] In one embodiment of the present invention, the prediction results of the site selection schemes under different scenarios are analyzed to evaluate the feasibility and advantages and disadvantages; and the site selection schemes are optimized and adjusted according to the prediction results and actual needs, including:
[0115] Establish a comprehensive quantitative evaluation indicator system, including but not limited to environmental impact, technical feasibility and other dimensions; ensure that each dimension has clear quantitative standards and calculation methods;
[0116] Summarize the site selection prediction results under different development scenarios and make horizontal and vertical comparisons;
[0117] Conduct sensitivity analysis on key parameters to assess the impact of changes in these parameters on the predicted results of the site selection plan;
[0118] Based on the quantitative evaluation indicators and prediction results, conduct a comprehensive assessment of the feasibility of each site selection plan and determine whether the plan is feasible;
[0119] Based on the feasibility assessment, further analyze the advantages and disadvantages of each site selection plan;
[0120] Consider the cross-impact between different site selection options, such as resource competition, environmental synergy, etc.; evaluate the synergy or conflict between the options by simulating the prediction results under different combinations of options.
[0121] According to the feasibility assessment and advantages and disadvantages analysis results, clarify the goal of optimizing the site selection plan;
[0122] Develop specific optimization strategies for each optimization goal;
[0123] Apply optimization strategies to dynamic forecasting models, re-perform scenario forecasts and result analysis; evaluate optimization effects based on new forecast results, and make iterative adjustments as needed;
[0124] The analysis results and optimization strategies are organized into a decision support report, which clearly presents the performance of site selection plans under different development scenarios, analysis of advantages and disadvantages, optimization suggestions and predictions of their implementation effects.
[0125] Use data visualization technology to intuitively display the analysis results in the form of charts, maps, etc.
[0126] The working principle of the above technical solution is as follows: construct a comprehensive set of quantitative evaluation indicators, which is the basis of the entire analysis process. The system covers multiple dimensions such as environmental impact and technical feasibility, and each dimension has clear quantitative standards and calculation methods. These indicators are used for the subsequent objective and comprehensive evaluation of the site selection plan; the site selection plan prediction results under different development scenarios are summarized and compared horizontally and vertically. The horizontal comparison reveals the performance differences under different scenarios at the same time point, which helps to understand the impact of different external environments on the site selection plan; the vertical comparison shows the trend of change over time under the same scenario, which helps to predict the long-term development of the site selection plan; sensitivity analysis is performed on the key parameters that affect the site selection decision, and the degree of influence of these parameter changes on the prediction results is evaluated. This step helps to identify the key factors that affect the site selection decision and provide decision makers with insights on which factors need to be focused on and adjusted; based on the quantitative evaluation indicators and prediction results, the feasibility of each site selection plan is comprehensively evaluated. At the same time, the advantages and disadvantages of each plan are analyzed in depth to clarify the positive benefits and potential opportunities it brings, as well as the challenges and risks it may face. This step provides decision makers with an in-depth understanding of the practical feasibility and potential value of each plan; considering the mutual influence between site selection plans, such as resource competition and environmental synergy, the synergy or conflict between the plans is evaluated by simulating the prediction results under different combinations of plans. This step helps decision makers to have a more comprehensive understanding of the interaction between the plans, so as to make more scientific and reasonable decisions; according to the feasibility evaluation and advantages and disadvantages analysis results, the goal of site selection plan optimization is clarified. Specific optimization strategies are formulated for each optimization goal, and the feasibility and cost-effectiveness of strategy implementation are considered. The optimization strategy is applied to the dynamic prediction model for re-prediction and analysis, the optimization effect is evaluated according to the new prediction results, and iterative adjustments are made as needed to ensure the effectiveness and pertinence of the optimization strategy; the analysis results and optimization strategies are organized into a decision support report, which clearly presents the site selection plan performance, advantages and disadvantages analysis, optimization suggestions and implementation effect prediction under different development scenarios. Data visualization technology is used to intuitively display the analysis results in the form of charts, maps, etc., to help decision makers understand the information and laws behind complex data more quickly and make more accurate and scientific decisions.
[0127] The effect of the above technical solution is: by establishing a comprehensive quantitative evaluation index system, and predicting, comparing, analyzing and evaluating based on this system, it can ensure that the site selection decision process is more scientific and systematic. Quantitative data and methods reduce the impact of subjective judgment and improve the accuracy of decision-making; this comprehensive evaluation method helps to discover potential problems and opportunities, so as to make more comprehensive and reasonable decisions; by simulating the performance of site selection schemes under different development scenarios and making horizontal and vertical comparisons, this technical solution can help decision makers better understand and predict future environmental changes. At the same time, the sensitivity analysis of key parameters can identify the key factors affecting decision-making and provide decision makers with flexible strategies to cope with changes; based on feasibility assessment and advantages and disadvantages analysis, this technical solution can clarify the goal of site selection scheme optimization and formulate specific optimization strategies. These strategies aim to reduce costs, improve benefits, etc., which help optimize resource allocation and reduce decision-making risks; through data visualization technology, the analysis results are intuitively displayed in the form of charts, maps, etc., which can help decision makers understand the information and laws behind complex data more quickly. This intuitive display method improves decision-making efficiency and transparency, making the decision-making process more open and traceable; the technical solution involves cooperation in multiple departments and fields, such as environment, technology, etc. When evaluating the site selection plan, the technical solution pays special attention to environmental impact and sustainability. By ensuring that the site selection plan meets environmental protection standards and sustainable development requirements, it can support the long-term development of the enterprise.
[0128] In one embodiment of the present invention, the output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through an expert system, fuzzy logic, and a neural network to generate a visual site selection recommendation evaluation report, including:
[0129] Integrate the output results of the site optimization model and the dynamic prediction model to obtain an integrated data set, and perform feature engineering on the integrated data set;
[0130] Select an existing comprehensive evaluation model; train the comprehensive evaluation model by integrating the data set; use the cross-validation method to verify the accuracy and generalization ability of the comprehensive evaluation model;
[0131] Use the trained comprehensive evaluation model to evaluate different site selection options and generate evaluation results; and visualize the evaluation results.
[0132] The working principle of the above technical solution is: integrate the output results of the site selection optimization model and the dynamic prediction model into a unified data set. In this step, feature engineering processing may be required, including feature selection, feature transformation, etc., to ensure the quality and integrity of the data set; the comprehensive evaluation model can adopt a variety of methods, such as expert system, fuzzy logic and neural network. These methods can be used in combination to comprehensively consider the influence of various factors and generate the final evaluation results; after selecting a suitable comprehensive evaluation model, the model is trained using the integrated data set. During the training process, methods such as cross-validation can be used to evaluate the accuracy and generalization ability of the model to ensure that the model has good predictive ability; use the trained comprehensive evaluation model to evaluate different site selection schemes. This includes inputting the relevant data of each scheme into the model and generating the corresponding evaluation results; the final evaluation results can be displayed in a visual way so that decision makers can more intuitively understand and compare the advantages and disadvantages of each site selection scheme. Visual reports can include charts, maps and other forms to make the evaluation results clearer and easier to understand.
[0133] The effects of the above technical solution are as follows: by integrating the output results of the site selection optimization model and the dynamic prediction model, a variety of factors affecting site selection are comprehensively considered, including geographical location, traffic conditions, market demand, etc., so as to generate more comprehensive and accurate site selection suggestions; through feature engineering processing and training of the comprehensive evaluation model, the original data can be effectively processed and the most representative and influential features can be extracted, thereby improving the accuracy and reliability of the site selection model; the comprehensive evaluation model can comprehensively consider the weights and influence of various factors through methods such as expert systems, fuzzy logic and neural networks, reducing subjective bias in the decision-making process and reducing decision-making risks; through the visualization of the site selection recommendation evaluation report, decision makers can intuitively understand the evaluation results of each site selection plan, and compare and analyze them. This helps decision makers make decisions more quickly and convey the decision basis to relevant stakeholders; the training and verification process of the comprehensive evaluation model, as well as the visual display of the evaluation results, can help decision makers make site selection decisions more quickly and accurately, and improve decision-making efficiency.
[0134] In one embodiment of the present invention, the output results of the site selection optimization model and the dynamic prediction model are integrated to obtain an integrated data set, and feature engineering is performed on the integrated data set, including:
[0135] Match and integrate the Pareto optimal solution set output by the site selection optimization model with the prediction results of the dynamic prediction model under different development scenarios to obtain a comprehensive data set including site selection optimization results and future prediction trends; and verify the comprehensive data set, including data consistency check, outlier processing and missing value filling;
[0136] Based on data integration, deep feature engineering is performed, which includes data cleaning, conversion, integration, polynomial feature and interactive feature generation, and mining is performed based on the hidden correlations and deep-level features between deep feature engineering data;
[0137] Through time series analysis technology (such as ARIMA, LSTM, etc.), the time series data in the dynamic prediction model is further mined to extract time trend characteristics;
[0138] Use spatial analysis techniques (such as GIS spatial interpolation and spatial autocorrelation analysis) to process geospatial data and identify spatial distribution patterns and spatial heterogeneity;
[0139] Conduct sensitivity analysis for different development scenarios output by the dynamic prediction model to evaluate the degree and direction of impact of changes in key indicators under different scenarios on site selection results;
[0140] By constructing a situational sensitivity matrix, the advantages and disadvantages of site selection options under different scenarios are quantified, providing decision makers with flexible response strategies in uncertain environments;
[0141] Combined with dynamic prediction results, a quantitative assessment of risks and opportunities of site selection options is conducted, and machine learning algorithms (such as random forests and gradient boosting trees) are used to identify key factors affecting the success of site selection and their potential risk points.
[0142] The working principle of the above technical solution is to match and integrate the Pareto optimal solution set output by the site selection optimization model with the prediction results of the dynamic prediction model under different development scenarios. Subsequently, the comprehensive data set is strictly verified, including data consistency check, outlier processing and missing value filling; on the basis of data integration, deep feature engineering is carried out to explore the hidden correlations and deep-level features between data. Through these processes, the expressiveness of data and the prediction ability of the model can be significantly improved. Time series analysis techniques (such as ARIMA model, LSTM neural network, etc.) are used to further mine the time series data in the dynamic prediction model. These techniques can capture the time trend characteristics in the data, predict future changes, and provide insights into the time dimension for site selection decisions. With the help of spatial analysis techniques (such as GIS spatial interpolation, spatial autocorrelation analysis, etc.), geospatial data is processed. These techniques can identify spatial distribution patterns and spatial heterogeneity, reveal the important impact of geographical location on site selection decisions, and help decision makers better understand the advantages and disadvantages of different site selection options in the spatial dimension. Sensitivity analysis is carried out for different development scenarios output by the dynamic prediction model. By constructing a scenario sensitivity matrix, the advantages and disadvantages of site selection options under different scenarios can be quantified, allowing decision makers to flexibly respond to future changes. Combined with dynamic prediction results, the risks and opportunities of site selection options can be quantitatively evaluated. Using machine learning algorithms (such as random forests, gradient boosting trees, etc.), key factors affecting the success of site selection and their potential risk points can be identified.
[0143] The effect of the above technical solution is: by integrating the output results of the site selection optimization model and the dynamic prediction model, a comprehensive data set containing the site selection optimization results and future prediction trends is formed. This comprehensive data set provides decision makers with a more comprehensive and systematic information basis, which helps to make more scientific and reasonable decisions; the data consistency check, outlier processing and missing value filling of the comprehensive data set significantly improve the quality and accuracy of the data. Reduce the risk of decision-making errors caused by data errors or missing data; through deep feature engineering methods such as data cleaning, conversion, integration, polynomial features and interactive feature generation, the hidden correlations and deep-level features between data are excavated. It helps to discover the key factors affecting site selection and their interactions, and provide more refined insights for decision-making; using time series analysis techniques such as ARIMA and LSTM, the time series data in the dynamic prediction model is deeply mined to extract time trend characteristics. It helps to predict the changing trend of the environment and provide a reference for the time dimension for site selection decisions; through spatial analysis techniques such as GIS spatial interpolation and spatial autocorrelation analysis, the geographic spatial data is processed to identify spatial distribution patterns and spatial heterogeneity. It helps to evaluate the impact of geographical location on site selection and discover potential spatial advantages and disadvantages; conduct sensitivity analysis on changes in key indicators under different development scenarios to evaluate the degree and direction of their impact on site selection results. This helps decision makers understand the risks and opportunities under different scenarios and provide a basis for formulating flexible response strategies; by constructing a situational sensitivity matrix, quantify the advantages and disadvantages of site selection plans under different scenarios. It helps decision makers quickly select the best solution in different scenarios and improve the flexibility and adaptability of decision-making; combine dynamic prediction results to quantify the risks and opportunities of site selection plans. Use machine learning algorithms to identify key factors that affect the success of site selection and their potential risk points, and provide decision makers with risk warnings and opportunity grasping references.
[0144] One embodiment of the present invention, as Figure 2 As shown, a system for implementing the above-mentioned construction project planning site selection and land use preliminary evaluation method, the system includes:
[0145] Data acquisition module: collects basic data of site selection for multiple construction projects through multiple data sources, stores the basic data in the cloud space, selects a subset of basic data for project site selection, and pre-processes the data subset;
[0146] Site selection optimization module: The pre-processed data subset is deeply mined through deep learning algorithms to determine the weight of each indicator. Based on the project characteristics and construction goals, a site selection optimization model with multi-dimensional indicators is constructed. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set.
[0147] Dynamic prediction module: Based on basic data, a dynamic prediction model is constructed, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the dynamic prediction model is used to predict the performance of site selection plans under different scenarios;
[0148] Fusion module: Fusion the output results of the site selection optimization model and the dynamic prediction model, conduct comprehensive evaluation through expert system, fuzzy logic and neural network, and generate a visual site selection recommendation evaluation report.
[0149] The working principle of the above technical solution is: using multiple data sources, such as integrated geographic information system (GIS), remote sensing satellites and Internet of Things devices, to collect basic data required for the site selection of multiple construction projects; the basic data includes data on physical geography (such as topography, climate conditions, geological structure) and ecological environment. The collected data needs to be preprocessed, including data cleaning, denoising and format conversion, to ensure the accuracy and availability of the data. The preprocessed data is deeply mined using a deep learning algorithm to extract high-level features that reflect the key influencing factors of the site selection of the construction project; in this process, the weights of each indicator are determined, and a multi-dimensional site selection optimization model is constructed based on the characteristics and construction goals of the project. The preprocessed data subset is input into the site selection optimization model, and the Pareto optimal solution set is obtained using the model. The optimal solution set includes recommended site locations and their scores on various evaluation indicators, as well as multi-dimensional indicators such as environmental benefits and risk control. Based on the collected basic data, a dynamic prediction model is constructed, and different development scenarios can be preset. These preset development scenarios are input into the dynamic prediction model, and the model is used to predict the performance of the site selection scheme under different scenarios. The output results of the site selection optimization model and the dynamic prediction model are integrated. Expert systems, fuzzy logic, neural networks and other methods are used for comprehensive evaluation to generate a visual site selection recommendation evaluation report. This report can provide decision makers with intuitive and comprehensive site selection recommendations and help them make the most appropriate decision.
[0150] The effects of the above technical solutions are as follows: through the collection and comprehensive utilization of multiple data sources, including GIS, remote sensing satellites and Internet of Things devices, comprehensive basic data on the site selection of construction projects are obtained, covering information on multiple aspects such as natural geography, providing rich data support for site selection evaluation; deep learning algorithms are used to deeply mine the data, extract advanced features that reflect the key influencing factors of site selection, and construct a site selection optimization model of multi-dimensional indicators by determining the weights of each indicator, providing scientific analysis methods and tool support for site selection decisions; a dynamic prediction model is constructed based on basic data, which can preset different development scenarios and predict the performance of site selection plans under different scenarios, thereby improving the flexibility and pertinence of site selection decisions; the output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through expert systems, fuzzy logic and neural networks, etc., to generate a visual site selection recommendation evaluation report, providing decision makers with intuitive and comprehensive site selection recommendations, and improving the scientificity and efficiency of decision-making; by comprehensively utilizing multiple data sources and advanced algorithm technologies, the comprehensiveness, accuracy and scientificity of site selection evaluation are improved, providing decision makers with more diversified and reliable site selection decision support, which helps to improve decision efficiency and reduce decision risks.
[0151] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A method for site selection and land use pre-evaluation of construction project planning, characterized in that: The method comprises: Collecting basic data on site selection for multiple construction projects through multiple data sources, storing the basic data in a cloud space, screening out a subset of the basic data on site selection for the project, and preprocessing the data subset; Through deep learning algorithms, the pre-processed data subsets are deeply mined to determine the weight of each indicator. Based on the project characteristics and construction goals, a multi-dimensional indicator site selection optimization model is constructed. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set. Based on basic data, a dynamic prediction model is constructed, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the dynamic prediction model is used to predict the performance of site selection plans under different scenarios; The output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through expert systems, fuzzy logic and neural networks to generate a visual site selection recommendation evaluation report; The method collects basic data of site selection for multiple construction projects through multiple data sources, stores the basic data in a cloud space, selects a subset of basic data for project site selection, and pre-processes the data subset, including: Determine the basic data that needs to be collected, and select a data source to collect the basic data based on the basic data that needs to be collected; Determine the frequency and time period of collection, store the collected basic data in a cloud space, and the cloud space receives and processes the basic data; The cloud space stores the processed basic data into a structured database; According to the needs of project site selection, a subset of project site selection data is screened out, and the data subset is preprocessed, wherein the preprocessing includes data cleaning, data conversion, feature engineering, and data integration; The basic data include natural geographic data and ecological environment data; the data sources include integrated geographic information systems, remote sensing satellites, the Internet and sensor networks; The deep learning algorithm is used to deeply mine the pre-processed data subset, determine the weight of each indicator, and build a multi-dimensional indicator site selection optimization model based on project characteristics and construction goals. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set, including: Use deep learning algorithms to extract features from preprocessed data subsets, and select feature subsets based on the importance of features and the degree of influence on the site selection of construction projects; Evaluate the importance of indicators; determine the weight of each indicator based on the importance of the evaluation indicators; define a multi-dimensional site optimization objective function based on the construction goals; According to business needs and project characteristics, set the constraints of the site selection optimization model, integrate the objective function, constraints and evaluation index system, and build a site selection optimization model; Using multi-objective optimization algorithm, according to the constraints of the site selection optimization model, the Pareto optimal solution set of each evaluation index is obtained; The obtained Pareto optimal solution set is explained and analyzed, and the recommended site locations and the scores of the recommended site locations on various evaluation indicators are extracted; The method comprises: constructing a dynamic prediction model based on basic data, presetting development scenarios, inputting the presetting development scenarios into the dynamic prediction model, and using the dynamic prediction model to predict the performance of the site selection scheme under different scenarios; including: Obtain basic data, clean the basic data, integrate data from different data sources into a unified data set, extract features, and build a dynamic prediction model; The data set is divided into a training set, a test set and a validation set. The dynamic prediction model is trained with the training set, tested with the test set and predicted with the validation set. According to the forecast demand, set different development scenarios and set relevant parameters under each development scenario; Generate corresponding scenario data according to the preset development scenario, input the generated scenario data into the dynamic prediction model to predict the site selection plan; Analyze the predicted results of site selection plans under different scenarios, evaluate their feasibility and advantages and disadvantages; and optimize and adjust the site selection plans based on the predicted results and actual needs; The output results of the site selection optimization model and the dynamic prediction model are integrated, and a comprehensive evaluation is performed through an expert system, fuzzy logic and neural network to generate a visual site selection recommendation evaluation report, including: Integrate the output results of the site optimization model and the dynamic prediction model to obtain an integrated data set, and perform feature engineering on the integrated data set; Select a comprehensive evaluation model and train it by integrating the data set; use the cross-validation method to verify the accuracy and generalization ability of the comprehensive evaluation model; Use the trained comprehensive evaluation model to evaluate different site selection options and generate evaluation results; and visualize the evaluation results; The output results of the site selection optimization model and the dynamic prediction model are integrated to obtain an integrated data set, and feature engineering is performed on the integrated data set, including: Match and integrate the Pareto optimal solution set output by the site selection optimization model with the prediction results of the dynamic prediction model under different development scenarios to obtain a comprehensive data set containing site selection optimization results and future prediction trends; and verify the comprehensive data set; Based on data integration, deep feature engineering is carried out to mine the hidden correlations and deep features between deep feature engineering data; Through time series analysis, the time series data in the dynamic prediction model is further mined to extract time trend characteristics; Use spatial analysis techniques to process geospatial data and identify spatial distribution patterns and spatial heterogeneity based on the processing results; Conduct sensitivity analysis for different development scenarios output by the dynamic prediction model to evaluate the degree and direction of impact of changes in key indicators under different scenarios on site selection results; By constructing a situational sensitivity matrix, the advantages and disadvantages of site selection options under different scenarios are quantified, providing decision makers with flexible response strategies in uncertain environments; Combined with dynamic prediction results, a quantitative assessment of risks and opportunities of site selection options is conducted, and machine learning algorithms are used to identify key factors affecting the success of site selection and their potential risk points.
2. A construction project planning site selection and land use preliminary evaluation method according to claim 1, characterized in that: The determining of the frequency and time period of collection, storing the collected basic data in a cloud space, the cloud space receiving the basic data, and processing the basic data includes: Determine the basic data to be collected and the frequency of collection according to project requirements and data timeliness; Collect basic data according to the basic data to be collected and the frequency of collection; The collected basic data are processed accordingly, and the results of the corresponding processing are transmitted to the cloud space respectively; after the cloud space receives the corresponding processed basic data, the received data is processed again.
3. A construction project planning site selection and land use preliminary evaluation method according to claim 2, characterized in that: The collected basic data are processed accordingly, and the results of the processing are transmitted to the cloud space respectively; After receiving the corresponding processed basic data, the cloud space processes the received data again, including: Processing the collected basic data respectively, including natural geographical data processing and ecological environment data processing; The results of the corresponding processing of the basic data are compressed respectively and transmitted to the cloud space through a multi-channel transmission protocol, and the cloud space processes the received data again.
4. A system for implementing a construction project planning site selection and land use pre-evaluation method as claimed in claim 1, characterized in that: The system comprises: Data acquisition module: collects basic data of site selection for multiple construction projects through multiple data sources, stores the basic data in the cloud space, selects a subset of basic data for project site selection, and pre-processes the data subset; Site selection optimization module: The pre-processed data subset is deeply mined through deep learning algorithms to determine the weight of each indicator. Based on the project characteristics and construction goals, a site selection optimization model with multi-dimensional indicators is constructed. The input data is input into the site selection optimization model to obtain the Pareto optimal solution set. Dynamic prediction module: Based on basic data, a dynamic prediction model is constructed, development scenarios are preset, the preset development scenarios are input into the dynamic prediction model, and the dynamic prediction model is used to predict the performance of site selection plans under different scenarios; Fusion module: Fusion the output results of the site selection optimization model and the dynamic prediction model, conduct comprehensive evaluation through expert system, fuzzy logic and neural network, and generate a visual site selection recommendation evaluation report.
