A multi-dimensional old building reconstruction effect evaluation method and device

Abstract

The application discloses a multi-dimensional old building reconstruction effect evaluation method and device, and relates to the field of building engineering quality evaluation.The method comprises the following steps: acquiring project characteristic parameters of a building to be evaluated and real-time evaluation data corresponding to each evaluation dimension of the building to be evaluated; inputting the project characteristic parameters into a dynamic weight prediction model to obtain target weights of each evaluation dimension; inputting the real-time evaluation data corresponding to each evaluation dimension into a scoring prediction model to obtain basic scores corresponding to each basic dimension and a core score corresponding to a core dimension; and performing weighted processing on the target weights of each evaluation dimension, the basic scores of the building to be evaluated and the core score of the core dimension to obtain a reconstruction effect score of the building to be evaluated.The application sets multiple evaluation dimensions for evaluation, can cover the life cycle of old building construction, the reconstruction score is more objective, and the application no longer depends on personal experience and judgment, so that the accuracy and authenticity are improved.

Description

Technical Field

[0001] This application relates to the field of building engineering quality evaluation technology, and in particular to a multi-dimensional method and apparatus for evaluating the renovation effect of old buildings. Background Technology

[0002] Currently, in the field of urban renewal and renovation of old residential areas, building evaluation work mainly relies on two types of methods: one is segmented acceptance evaluation, which involves conducting item-by-item testing of construction quality according to relevant standards, or completing the acceptance of physical projects in accordance with technical specifications, and then introducing economic benefit analysis (such as return on investment, cost-benefit ratio, etc.) to assess the economic feasibility of the project; the other is comprehensive evaluation based on expert experience, which is usually carried out by an evaluation team composed of industry experts, who score each item according to a qualitative evaluation checklist, and combine the results of resident satisfaction questionnaires to form a final comprehensive rating conclusion.

[0003] However, in current practice, the evaluation of urban renewal projects still faces three major pain points: ① The evaluation dimensions are relatively singular: existing methods mostly focus on the quality of engineering construction and economic benefits, lacking a systematic quantitative assessment of multi-dimensional goals such as improved building functions, environmental sustainability, and socio-cultural benefits after renovation; ② High subjectivity: traditional evaluation processes rely heavily on the personal experience and judgment of experts, lacking objective and unified algorithmic models and digital tools covering the entire project lifecycle, resulting in low consistency and weak comparability of results; ③ Limited applicability: current evaluation systems often fail to fully consider the characteristics and objective differences of different types of urban renewal projects, and the "one-size-fits-all" evaluation method cannot truly and comprehensively reflect the actual renovation effects of various projects. In summary, current methods for evaluating urban renewal projects suffer from low accuracy and authenticity, and are not applicable to different types of renovation projects. Summary of the Invention

[0004] The purpose of this application is to address at least one of the aforementioned technical deficiencies.

[0005] On the one hand, embodiments of this application provide a multi-dimensional method for evaluating the effects of urban renewal building renovation, the method including: Obtain the project characteristic parameters of the building to be evaluated and the real-time evaluation data of the building to be evaluated for each evaluation dimension. The evaluation dimensions include at least one basic dimension and a core dimension. The project feature parameters are input into the dynamic weight prediction model to obtain the target weight for each evaluation dimension. The real-time evaluation data of the building to be evaluated for each evaluation dimension is input into the scoring prediction model to obtain the basic score of the building to be evaluated for each basic dimension and the core score of the building to be evaluated for the core dimension. The renovation effect score of the building to be evaluated is obtained by weighting the target weight of each evaluation dimension, the basic score of the building to be evaluated corresponding to each basic dimension, and the core score of the building to be evaluated corresponding to the core dimension.

[0006] Optionally, the dynamic weight prediction model includes a project feature encoding layer, a weight prediction layer, and a conflict resolution normalization layer. The dynamic weight prediction model obtains the target weight for each evaluation dimension in the following ways: The project feature parameters corresponding to the building to be evaluated are input into the project feature coding layer for categorical feature coding to obtain the corresponding project feature vector. The project feature vector is input into the weight prediction layer, so that the weight prediction layer adjusts the initial weight of each evaluation dimension based on the project feature vector and the feature weight mapping rule, so as to obtain the initial target weight of each evaluation dimension. The conflict resolution normalization layer modifies and normalizes the initial target weights of each evaluation dimension to obtain the target weights of each evaluation dimension.

[0007] Optionally, the conflict resolution normalization layer performs a modified normalization process on the initial target weights of each evaluation dimension to obtain the target weights for each evaluation dimension, including: The system detects whether there are any anomalies in the initial target weights of each evaluation dimension. If there are anomalies, the system corrects the initial target weights of each evaluation dimension according to the preset conflict resolution strategy, thus obtaining the corrected target weights of each evaluation dimension. The target weights of each evaluation dimension are normalized to obtain the target weights of each evaluation dimension. Abnormal cases include logical conflicts or numerical anomalies. The sum of the target weights of each evaluation dimension is 1.

[0008] Optionally, the weight prediction layer obtains the initial target weights for each evaluation dimension in the following ways: Based on the project feature vector and feature weight mapping rules, determine the target weight adjustment instruction triggered by the project feature parameters; The initial weights of each evaluation dimension are adjusted according to the target weight adjustment instruction to obtain the initial target weights for each evaluation dimension.

[0009] Optional, the basic dimensions include project preparation completion rate, construction management success rate, project acceptance pass rate, and innovation application improvement rate, while the core dimensions include renovation implementation score rate. Each evaluation dimension includes at least one secondary indicator, and each secondary indicator includes at least one tertiary indicator. The real-time evaluation data is the evaluation data of the building to be evaluated corresponding to each tertiary indicator.

[0010] Optionally, the rating prediction model includes a base dimension rating sub-network, which comprises a base vector encoding layer and at least one parallel branch sub-network. The rating prediction model obtains the base score for each base dimension of the building to be evaluated in the following manner: The real-time evaluation data of the building to be evaluated for each basic dimension is input into the basic vector encoding layer for vector encoding to obtain the basic feature vector of each basic dimension. For each basic dimension, the basic feature vector corresponding to the basic dimension is input into the corresponding branch sub-network, so that the corresponding branch sub-network compares the basic feature vector with the scoring requirements of each tertiary indicator included in the basic scoring rules, determines the score corresponding to each tertiary indicator included in the basic dimension, and sums the scores corresponding to each tertiary indicator included in the basic dimension to obtain the basic score of the building to be evaluated corresponding to the basic dimension.

[0011] Optionally, the core dimension includes at least one tertiary classification. The rating prediction model also includes a core dimension rating sub-network, which includes a core vector encoding layer and a core rating sub-network. The rating prediction model obtains the core rating of the building to be evaluated corresponding to the core dimension in the following way: The real-time evaluation data of the building to be evaluated corresponding to the core dimension is input into the core vector encoding layer to obtain the core feature vector corresponding to the core dimension. The core feature vector is input into the core scoring subnetwork so that the core scoring subnetwork determines the score rate of the building to be evaluated for each of the three categories based on the core feature vector. The score rate of each of the three categories is then weighted and summed according to the weights corresponding to each category to obtain the core score of the building to be evaluated for the core dimension.

[0012] Optionally, the three-level classification includes a basic category, a refinement category, and an enhancement category. Each three-level category includes at least one differentiated evaluation item. The core scoring sub-network determines the score rate of the building to be evaluated for each three-level category based on the core feature vector, including: The target differentiated evaluation items and the differential score corresponding to each target differentiated evaluation item are determined based on the core feature vector of the building to be evaluated corresponding to the three-level classification. Based on the differential score of each target differentiation evaluation item and the number of target differentiation evaluation items, the score rate of the building to be evaluated corresponding to the three-level classification is determined.

[0013] Optionally, after obtaining the renovation effect score of the building to be evaluated, the following may also be included: Obtain historical renovation effect scores and display the renovation effect scores and historical renovation effect scores in the form of a time curve graph; An intelligent diagnostic report is generated based on the real-time evaluation data and renovation effect scores of the building to be evaluated for each evaluation dimension. The intelligent diagnostic report includes specific deduction items.

[0014] On the other hand, embodiments of this application provide a multi-dimensional evaluation device for the renovation effect of old buildings, including: The data acquisition module is used to acquire the project characteristic parameters of the building to be evaluated and the real-time evaluation data of the building to be evaluated corresponding to each evaluation dimension. The evaluation dimensions include at least one basic dimension and a core dimension. The weight determination module is used to input project feature parameters into the dynamic weight prediction model to obtain the target weight for each evaluation dimension; The scoring determination module is used to input the real-time evaluation data of the building to be evaluated for each evaluation dimension into the scoring prediction model to obtain the basic score of the building to be evaluated for each basic dimension and the core score of the building to be evaluated for the core dimension. The renovation effect scoring module is used to perform weighted processing based on the target weight of each evaluation dimension, the basic score of the building to be evaluated corresponding to each basic dimension, and the core score of the building to be evaluated corresponding to the core dimension, to obtain the renovation effect score of the building to be evaluated.

[0015] In another aspect, embodiments of this application provide an electronic device, including a processor and a memory: The memory is configured to store machine-readable instructions that, when executed by the processor, cause the processor to perform any one of the methods in a multi-dimensional evaluation method for the effectiveness of urban renewal building renovation.

[0016] The beneficial effects of the technical solutions provided in this application include at least the following: In this application, when determining the renovation effect score of a building to be evaluated, multiple quantitative indicators across various evaluation dimensions are used. This covers the entire construction lifecycle of the renovated building, making the final renovation score more objective. Furthermore, each evaluation dimension is divided into basic and core dimensions based on its contribution to the project. This better highlights the importance of each dimension, achieving intelligent shifting of the evaluation focus. A dynamic weight prediction model is used to obtain the target weight for each evaluation dimension, making it adaptable to various renovation projects with vastly different characteristics, from basic repairs to historical preservation, eliminating the evaluation bias of fixed weights for specific project types. Finally, a weighted summation based on the corresponding target weights yields the final renovation effect score. This eliminates reliance on expert experience and judgment, improving accuracy and authenticity. Moreover, it is adaptable to diverse construction conditions and is applicable to different types of renovation projects, demonstrating broader application prospects.

[0017] Furthermore, in this application, the weights of each evaluation dimension are obtained through project characteristic parameters and a dynamic weight prediction model. Since the project characteristic parameters characterize the attributes of the building being evaluated, the resulting weights are more consistent with the characteristics of the building, eliminating the evaluation bias of fixed weights for specific types of projects. In addition, because the weight allocation of the evaluation dimensions is dynamically adjusted through the project characteristic parameters corresponding to the building being evaluated and the dynamic weight prediction model, the weight allocation no longer relies on the subjective experience and judgment of experts, but is based on objective reasoning using a transparent and structured rule base. This reduces human arbitrariness, and the final determined target weights can be explained by tracing back to the rules they match, thus making the evaluation results more convincing. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a multi-dimensional method for evaluating the effects of urban renewal building renovation, provided as an embodiment of this application; Figure 2 A schematic diagram illustrating an evaluation dimension provided in an embodiment of this application; Figure 3 A schematic diagram of a multi-dimensional evaluation device for the renovation effect of old buildings provided in this application embodiment; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting the invention.

[0021] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0023] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0024] Specifically, such as Figure 1 As shown, the method may include: Step S101: Obtain the project feature parameters of the building to be evaluated and the real-time evaluation data of the building to be evaluated corresponding to each evaluation dimension. The evaluation dimensions include at least one basic dimension and a core dimension.

[0025] Optionally, each evaluation dimension can be divided into at least one basic dimension and one core dimension. The division of basic and core dimensions is determined based on the importance of the evaluation dimension to the project transformation. The real-time evaluation data obtained is the real-time evaluation data of the building to be evaluated for each basic and core dimension.

[0026] Step S102: Input the project feature parameters into the dynamic weight prediction model to obtain the target weight for each evaluation dimension.

[0027] Optionally, project characteristic parameters refer to attribute parameters that describe the essence of urban renewal projects. In optional embodiments of this application, project characteristic parameters include at least one of strategic attribute parameters, scale and resource parameters, and socio-ecological parameters. Strategic attribute parameters include project type and core renovation goals. Scale and resource parameters include investment scale level, construction period urgency, and technology application complexity. Socio-ecological parameters include resident structure characteristics and sensitivity to the surrounding environment.

[0028] Among them, project types can be divided into basic security type, quality improvement type, and historical style type, etc.; core renovation goals can be divided into safety reinforcement, energy conservation and consumption reduction, and humanistic care, etc.; policy guidance level can be divided into ordinary projects, key demonstration projects, etc.; scale and resource parameters can include investment scale level, construction period urgency and technology application complexity; resident structure characteristics can refer to the aging rate; and surrounding environmental sensitivity can refer to factors such as proximity to schools or hospitals, etc.

[0029] In practical applications, the acquired project feature parameters can be input into a dynamic weight prediction model to obtain the target weight for each evaluation dimension. The target weight for each evaluation dimension obtained at this point includes the target weight for each basic dimension and the target weight for each core dimension. Since the core dimension is relatively more important than the basic dimension in old building renovation projects in practical applications, the target weight for the core dimension is relatively larger than the target weight for the basic dimension.

[0030] In optional embodiments of this application, the dynamic weight prediction model includes a project feature encoding layer, a weight prediction layer, and a conflict resolution normalization layer. The dynamic weight prediction model obtains the target weight for each evaluation dimension in the following ways: The project feature parameters corresponding to the building to be evaluated are input into the project feature coding layer for categorical feature coding to obtain the corresponding project feature vector. The project feature vector is input into the weight prediction layer, so that the weight prediction layer adjusts the initial weight of each evaluation dimension based on the project feature vector and the feature weight mapping rule, so as to obtain the initial target weight of each evaluation dimension. The conflict resolution normalization layer modifies and normalizes the initial target weights of each evaluation dimension to obtain the target weights of each evaluation dimension.

[0031] Optionally, the dynamic weight prediction model may include a project feature encoding layer, a weight prediction layer, and a conflict resolution normalization layer. In practical applications, the project feature parameters corresponding to the building to be evaluated can be input into the project feature encoding layer. This project feature encoding layer uses an embedding layer or one-hot encoding to map each type of project feature parameter into a meaningful dense vector, enabling the model to learn the semantic relationships between different project types. Then, the vectors obtained from mapping each type of project feature parameter are concatenated to obtain the corresponding project feature vector.

[0032] Furthermore, a weight prediction layer composed of one or more fully connected layers adjusts the initial weights of each evaluation dimension based on the project feature vector and feature weight mapping rules to obtain the initial target weights of each evaluation dimension.

[0033] In optional embodiments of this application, the weight prediction layer obtains the initial target weight for each evaluation dimension in the following ways: Based on the project feature vector and feature weight mapping rules, determine the target weight adjustment instruction triggered by the project feature parameters; The initial weights of each evaluation dimension are adjusted according to the target weight adjustment instruction to obtain the initial target weights for each evaluation dimension.

[0034] The feature weight mapping rule represents the project feature vector requirements corresponding to each weight adjustment instruction, and each mapping rule is a "condition-action" pair. For example, the condition part of a "condition-action" pair defines a specific state or combination of one or more project feature vectors, such as IF Project Type = "Historical Appearance Type" and "Cultural Heritage" of the core transformation goal. In this case, the action part specifies which evaluation dimensions should have their weights adjusted upward or downward when the condition is met, and indicates the direction and relative magnitude of the adjustment, such as increasing the weight of the innovation application dimension by 0.15 and decreasing the baseline weight of the "Improvement" indicator in the transformation implementation dimension by 0.10.

[0035] Correspondingly, the initial weights of each pre-configured evaluation dimension can be obtained. Then, the project feature vectors are matched with the conditional parts of each mapping rule to identify all mapping rules that satisfy the conditions. The action part of the satisfied mapping rule is the determined target weight adjustment instruction. Finally, the initial weights of each evaluation dimension are adjusted cumulatively according to the target weight adjustment instruction, i.e., the corresponding values ​​are increased or decreased, thus obtaining the initial target weight of each evaluation dimension. Further, the initial target weights of each evaluation dimension are input into the conflict resolution normalization layer to perform correction and normalization processing on the initial target weights of each evaluation dimension, thus obtaining the target weight of each evaluation dimension.

[0036] In an optional embodiment of this application, the conflict resolution normalization layer performs a modified normalization process on the initial target weights of each evaluation dimension to obtain the target weights of each evaluation dimension, including: The system detects whether there are any anomalies in the initial target weights of each evaluation dimension. If there are anomalies, the system corrects the initial target weights of each evaluation dimension according to the preset conflict resolution strategy, thus obtaining the corrected target weights of each evaluation dimension. The target weights of each evaluation dimension are normalized to obtain the target weights of each evaluation dimension. Abnormal cases include logical conflicts or numerical anomalies. The sum of the target weights of each evaluation dimension is 1.

[0037] Optionally, the conflict resolution normalization layer can be composed of a Softmax activation function. After obtaining the initial target weights for each evaluation dimension, the conflict resolution normalization layer can detect whether there are any abnormalities in the initial target weights for each evaluation dimension, that is, check whether there are logical conflicts or exceed reasonable ranges in the adjusted weights. For example, if the weight of a certain dimension is adjusted to a negative value or multiple rule adjustments contradict each other, the initial target weights of each evaluation dimension are corrected according to the preset conflict resolution strategy to obtain the corrected target weights for each evaluation dimension. The conflict resolution strategy can be set according to actual requirements, and this embodiment does not limit it. For example, the "nearest rule priority" or "expert preset priority" strategy can be used to correct the initial target weights, thereby ensuring the rationality and consistency of the obtained weight set.

[0038] Furthermore, the weight coefficients of each dimension after adjustment and conflict resolution are normalized to ensure that their sum is strictly equal to 1. This ensures the mathematical rigor of the output dynamic weight set, maintains the non-linear mapping relationship, avoids logical conflicts, and can then be directly used for weighted calculation.

[0039] Optionally, the weight prediction layer in this application is implemented based on a multilayer perceptron. In this case, the process of obtaining the target weight of each evaluation dimension through the dynamic weight prediction model can be expressed by the following formula: W = Softmax(Z) Z = (V_project)= A^(3)( L^(3)( A^(2)( L^(2)( A^(1)( L^(1)(V_project) ) ) ) ) Where W is the target weight for each evaluation dimension, and V_project is the project feature vector obtained by feature encoding the project feature parameters. It is a universal function approximator responsible for complex feature inference. Softmax is the activation function, L^(1) is the first linear transformation layer, A^(1) is the first activation function, L^(2) is the second linear transformation layer, A^(2) is the second activation function, L^(3) is the third linear transformation layer, and A^(3) is the third activation function.

[0040] In this application, the weight of each evaluation dimension is obtained through project characteristic parameters and a dynamic weight prediction model. Since the project characteristic parameters characterize the attributes of the building being evaluated, the resulting weights are more consistent with the characteristics of the building, eliminating the evaluation bias of fixed weights for specific types of projects. Furthermore, because the weight allocation of the evaluation dimensions is dynamically adjusted through the project characteristic parameters corresponding to the building being evaluated and the dynamic weight prediction model, the weight allocation no longer relies on the subjective experience and judgment of experts, but is based on objective reasoning using a transparent and structured rule base. This reduces human arbitrariness, and the final determined target weights can be explained by tracing back to the rules they match, thus making the evaluation results more convincing.

[0041] In optional embodiments of this application, the basic dimensions include project preparation completion rate, construction management success rate, project acceptance pass rate, and innovation application improvement rate, while the core dimensions include renovation implementation score rate. Each evaluation dimension includes at least one secondary indicator, and each secondary indicator includes at least one tertiary indicator. The real-time evaluation data is the evaluation data of the building to be evaluated corresponding to each tertiary indicator.

[0042] Optionally, taking into account the evaluation of the effects of urban renewal buildings, the evaluation dimensions in the multi-dimensional indicator library, namely project preparation completeness rate, construction management success rate, project acceptance pass rate, and innovation application improvement rate, are divided into basic dimensions, and the renovation implementation score rate is taken as the core dimension, as follows: Figure 2 As shown in the table. In practical applications, depending on the type of evaluation data, each evaluation dimension, i.e., the primary indicator, can be further divided into at least one secondary indicator, and each secondary indicator can be further divided into at least one tertiary indicator. For example, when the evaluation dimensions are project preparation completeness rate, construction management success rate, project acceptance pass rate, innovation application improvement rate, and renovation implementation score rate, the specific division of the secondary and tertiary indicators for each evaluation dimension and the target weight for each evaluation dimension can be shown in the table below:

[0043] Step S103: Input the real-time evaluation data of the building to be evaluated for each evaluation dimension into the scoring prediction model to obtain the basic score of the building to be evaluated for each basic dimension and the core score of the building to be evaluated for the core dimension.

Claims

1. A multi-dimensional method for evaluating the effects of urban renewal building renovation, characterized in that, include: Obtain the project characteristic parameters of the building to be evaluated and the real-time evaluation data of the building to be evaluated corresponding to each of the evaluation dimensions, wherein the evaluation dimensions include at least one basic dimension and a core dimension; The project feature parameters are input into the dynamic weight prediction model to obtain the target weight for each evaluation dimension. The real-time evaluation data of the building to be evaluated corresponding to each of the evaluation dimensions is input into the scoring prediction model to obtain the basic score of the building to be evaluated corresponding to each of the basic dimensions and the core score of the building to be evaluated corresponding to the core dimension. The renovation effect score of the building to be evaluated is obtained by weighting the target weight of each evaluation dimension, the base score of the building to be evaluated corresponding to each of the basic dimensions, and the core score of the building to be evaluated corresponding to the core dimension.

2. The method according to claim 1, characterized in that, The dynamic weight prediction model includes a project feature encoding layer, a weight prediction layer, and a conflict resolution normalization layer. The dynamic weight prediction model obtains the target weight for each evaluation dimension through the following methods: The project feature parameters corresponding to the building to be evaluated are input into the project feature encoding layer for categorical feature encoding to obtain the corresponding project feature vector. The project feature vector is input into the weight prediction layer, so that the weight prediction layer adjusts the initial weight of each evaluation dimension based on the project feature vector and the feature weight mapping rule, so as to obtain the initial target weight of each evaluation dimension. The conflict resolution normalization layer performs a correction and normalization process on the initial target weights of each evaluation dimension to obtain the target weights of each evaluation dimension.

3. The method according to claim 2, characterized in that, The conflict resolution normalization layer performs a correction and normalization process on the initial target weights of each evaluation dimension to obtain the target weights of each evaluation dimension, including: If any abnormality exists in the initial target weight of each evaluation dimension, the initial target weight of each evaluation dimension is corrected according to the preset conflict resolution strategy to obtain the corrected target weight of each evaluation dimension. The target weights of each evaluation dimension are normalized to obtain the target weights of each evaluation dimension. The abnormal situations include logical conflicts or numerical anomalies. The sum of the target weights of each evaluation dimension is 1.

4. The method according to claim 2, characterized in that, The weight prediction layer obtains the initial target weights for each evaluation dimension in the following ways: Based on the project feature vector and feature weight mapping rules, determine the target weight adjustment instruction triggered by the project feature parameters; The initial weights of each evaluation dimension are adjusted according to the target weight adjustment instruction to obtain the initial target weights of each evaluation dimension.

5. The method according to claim 1, characterized in that, The basic dimensions include project preparation completion rate, construction management success rate, project acceptance pass rate, and innovation application improvement rate. The core dimensions include renovation implementation score rate. Each evaluation dimension includes at least one secondary indicator, and each secondary indicator includes at least one tertiary indicator. The real-time evaluation data is the evaluation data of the building to be evaluated corresponding to each of the tertiary indicators.

6. The method according to claim 5, characterized in that, The rating prediction model includes a basic dimension rating sub-network, which comprises a basic vector encoding layer and at least one parallel branch sub-network. The rating prediction model obtains the basic rating of the building to be evaluated for each of the basic dimensions in the following manner: The real-time evaluation data of the building to be evaluated corresponding to each of the basic dimensions is input into the basic vector encoding layer for vector encoding to obtain the basic feature vector of each of the basic dimensions; For each of the basic dimensions, the basic feature vector corresponding to the basic dimension is input into the corresponding branch sub-network, so that the corresponding branch sub-network compares the basic feature vector with the scoring requirements corresponding to each of the three-level indicators included in the basic scoring rules, determines the score corresponding to each of the three-level indicators included in the basic dimension, and sums the scores corresponding to each of the three-level indicators included in the basic dimension to obtain the basic score of the building to be evaluated corresponding to the basic dimension.

7. The method according to claim 1, characterized in that, The core dimension includes at least one tertiary classification. The rating prediction model further includes a core dimension rating sub-network, which includes a core vector encoding layer and a core rating sub-network. The rating prediction model obtains the core rating of the building to be evaluated corresponding to the core dimension in the following manner: The real-time evaluation data of the building to be evaluated corresponding to the core dimension is input into the core vector encoding layer to obtain the core feature vector corresponding to the core dimension; The core feature vector is input into the core scoring subnetwork, so that the core scoring subnetwork determines the score rate of the building to be evaluated corresponding to each of the three-level categories based on the core feature vector, and performs weighted summation of the score rates of each of the three-level categories according to the weights corresponding to each of the three-level categories to obtain the core score of the building to be evaluated corresponding to the core dimension.

8. The method according to claim 6, characterized in that, The three-level classification includes a basic category, a refinement category, and an enhancement category. Each of the three-level categories includes at least one differentiated evaluation item. The core scoring sub-network determines the score rate of the building to be evaluated corresponding to each of the three-level categories based on the core feature vector, including: The target differentiated evaluation items and the differential score corresponding to each target differentiated evaluation item are determined based on the core feature vector of the building to be evaluated corresponding to the three-level classification. Based on the differential score of each of the target differentiated evaluation items and the number of the target differentiated evaluation items, the score rate of the building to be evaluated corresponding to the three-level classification is determined.

9. The method according to claim 1, characterized in that, After obtaining the renovation effect score of the building to be evaluated, the process also includes: Obtain historical renovation effect scores and display the renovation effect scores and historical renovation effect scores in the form of a time curve graph; An intelligent diagnostic report is generated based on the real-time evaluation data of the building to be evaluated corresponding to each of the evaluation dimensions and the renovation effect score. The intelligent diagnostic report includes specific deduction items.

10. A multi-dimensional evaluation method and device for the renovation effect of old buildings, characterized in that, include: The data acquisition module is used to acquire project characteristic parameters of the building to be evaluated and real-time evaluation data of the building to be evaluated corresponding to each of the evaluation dimensions. The evaluation dimensions include at least one basic dimension and a core dimension. The weight determination module is used to input the project feature parameters into the dynamic weight prediction model to obtain the target weight of each evaluation dimension. The scoring determination module is used to input the real-time evaluation data of the building to be evaluated corresponding to each of the evaluation dimensions into the scoring prediction model to obtain the basic score of the building to be evaluated corresponding to each of the basic dimensions and the core score of the building to be evaluated corresponding to the core dimension. The renovation effect scoring module is used to perform weighted processing based on the target weight of each evaluation dimension, the basic score of the building to be evaluated corresponding to each of the basic dimensions, and the core score of the building to be evaluated corresponding to the core dimension, to obtain the renovation effect score of the building to be evaluated.

Images