A multi-dimensional evaluation method and system for new urban construction integration projects

Abstract

The present invention discloses a new urban construction integration project evaluation method and system based on multiple dimensions, which relates to the field of urban construction evaluation and multidimensional data modeling and analysis technology, including collecting the original value and score of the same indicator in multiple new urban construction integration projects, and constructing an indicator scoring point set. The score values ​​in the indicator scoring point set are normalized, and the influence of outliers is eliminated. Multi-class scoring function models are fitted and the goodness of fit is calculated, and the standard scoring function form is determined according to the goodness of fit. A scoring function template is established based on the standard scoring function form and function parameters, and stored in a scoring function template library. The method of the present invention realizes a scientific, unified, and reusable quantitative evaluation mechanism for new urban construction integration projects, effectively overcoming the problems of strong subjectivity of scoring rules, difficulty in migration, and inconsistent evaluation standards in existing methods, and improving the objectivity, flexibility and intelligence of the evaluation system.

Description

Technical Field

[0001] The present invention relates to the technical field of urban construction evaluation and multidimensional data modeling and analysis, and specifically to a new urban construction integration project evaluation method and system based on multidimensionality. Background Art

[0002] To achieve scientific management and investment decision-making, establishing an efficient, quantitative, and objective project evaluation mechanism has become a practical necessity. Traditional evaluation methods, primarily based on economic indicators or single functional indicators, are no longer sufficient to meet the comprehensive assessment needs of emerging urban construction projects. Therefore, there is an urgent need to establish an evaluation system and methodology that integrates multi-dimensional indicators, supports multi-project comparison, and supports standardized management.

[0003] Currently, evaluation methods for new urban integration projects often rely on qualitative expert scoring, subjective weight assignments, or simple weighted summation models, lacking rigorous mathematical fitting mechanisms and a unified cross-project standardization system. Existing methods lack standardized normalization between indicators, leading to incomparable scoring results across projects. This is particularly serious when indicators use different units and dimensions, or when they come from heterogeneous sources (such as ecological and investment indicators). The lack of scoring function template-based mechanisms for score conversion and parameter versioning makes it impossible to track and revisit historical project scoring standards, resulting in poor reusability, traceability, and model evolution capabilities. Existing methods also lack goodness-of-fit judgment and scoring function structure selection mechanisms, making them difficult to adapt to the varying responses of multiple indicator types to numerical trends. Furthermore, while some existing solutions incorporate data-driven models, they often employ fixed function structures or black-box models, failing to clearly explain the scoring basis and calculation process. Therefore, there is an urgent need for a comprehensive evaluation method that is oriented towards new urban construction integration scenarios, supports multi-dimensional indicator integration, multi-model scoring function fitting and template encapsulation, so as to achieve a project evaluation mechanism with unified cross-project standards, comparable results and traceable scoring process. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing new city construction integrated project evaluation methods have different indicators with inconsistent dimensions, resulting in incomparable results, the scoring function lacks a template and version control mechanism, and lacks an adaptive fitting mechanism for the form and parameters of the scoring function, as well as how to construct a multi-dimensional integrated project evaluation system that supports the unification of scoring standards and backtracking of the scoring process.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a multi-dimensional new urban construction integration project evaluation method, comprising collecting the original values ​​and scores of the same indicator in multiple new urban construction integration projects, and constructing an indicator scoring point set.

[0007] The score values ​​in the indicator scoring point set are normalized and the influence of outliers is eliminated.

[0008] Fit a multi-class scoring function model and calculate the goodness of fit, and determine the standard scoring function form based on the goodness of fit.

[0009] A scoring function template is established based on the standard scoring function form and function parameters, and stored in the scoring function template library.

[0010] The method of establishing a scoring function template based on the standard scoring function form and function parameters includes structured recording of the determined function model and parameters. The template content includes function type, parameter value set, normalized input range, original indicator name and unit, score output rule, function establishment time and applicable conditions, and binding a unique function ID to each template.

[0011] The standard scoring function shape determined based on the goodness of fit includes:

[0012] The coefficient of determination of the evaluation index of each scoring function is calculated. If the coefficient of determination of the model evaluation index is greater than or equal to 0.85, it is considered an acceptable fitting function.

[0013] If the coefficient of determination of all model evaluation indicators is less than 0.85, they are marked as unscorable indicators.

[0014] Standard scoring functions are documented as a model type label and a list of fitted parameters.

[0015] The scoring function template library includes:

[0016] All scoring function templates are stored in the scoring function template library as key-value pairs and indexed by indicator code, supporting the scoring engine to call the corresponding template according to indicator type and scoring time.

[0017] The scoring function template library supports a version management mechanism. A new version of the template is generated each time the parameters are updated, and the template change history is recorded.

[0018] When multiple projects reference the same indicator, the scoring function defined in the template library is automatically called to achieve unified scoring standards across projects.

[0019] As a preferred embodiment of the multi-dimensional new urban construction integration project evaluation method described in the present invention, the method of collecting the original values ​​and scores of the same indicator across multiple new urban construction integration projects includes extracting historical new urban construction integration project data based on the indicator's unique identifier and associating the original values ​​with corresponding score records one by one. The original values ​​are automatically and synchronously acquired based on on-site equipment collection and data platform upload. The score records are derived from system scoring records. Each pair of sample data is provided with a timestamp and scoring source identifier. Unstructured samples caused by changes in the scoring mechanism are eliminated from the data, and a sample point set data structure that meets structural standards is constructed.

[0020] As a preferred solution of the multi-dimensional new urban construction integration project evaluation method described in the present invention, the construction of the index scoring point set includes storage in the form of triples, which include the original value, score, and scoring source. The original value is unified in dimension and mapped to the standard value range, and the full score and scoring step size are limited for the score record. When the indicator is the coverage rate of the sensing device, the original value unit is a percentage. Under the condition that the number of data points is not less than the preset number, the fitting process is entered and the data points are grouped and archived according to the project label.

[0021] As a preferred solution of the multi-dimensional new urban construction integration project evaluation method described in the present invention, the normalization processing of the score values ​​in the indicator scoring point set includes identifying the maximum and minimum scores in the original scoring point set. If the maximum and minimum values ​​are equal, all scores are treated as the default median. Otherwise, all scores are standardized and the normalized values ​​are compressed to a preset interval based on the compression factor.

[0022] As a preferred embodiment of the multi-dimensional new urban integration project evaluation method described in the present invention, eliminating the influence of outliers includes establishing a distribution view of the scoring point set, calculating the mean and standard deviation of all scores, and identifying outliers by determining whether the data point score exceeds the mean by plus or minus three standard deviations. If an outlier is identified, the sample record is excluded from the function fitting calculation, and the original data is retained. For the boundary values ​​of the normalized score equal to 0 and 1, a score perturbation factor is set to perform a numerical pullback.

[0023] As a preferred embodiment of the multi-dimensional new urban integration project evaluation method described in the present invention, the method of fitting multiple scoring function models and calculating goodness of fit includes sequentially fitting a linear function, a logarithmic function, and an S-shaped function model to the normalized scoring point set, solving each model parameter using the least squares method, and retaining the sum of squares of the residuals of each function fitting and the parameter combination results. All function fitting parameters are calculated by minimizing the error term, and after the fitting is completed, the goodness of fit is ranked according to the determination coefficient of the evaluation index.

[0024] As a preferred embodiment of the multi-dimensional new urban construction integration project evaluation method described in the present invention, the method of establishing a scoring function template from the standard scoring function form and function parameters includes structurally encapsulating the selected standard scoring function type and the corresponding parameter results to form a scoring function template object. The scoring function template includes a model type field, a parameter set field, a normalized input range field, an original indicator field, a score output rule field, an establishment time field, and an applicable condition field. Each scoring function template is assigned a unique function template ID, which is generated as a fixed-length string and consists of a function type code, a generation timestamp, and a parameter summary fingerprint.

[0025] Another object of the present invention is to provide a new urban construction integration project evaluation system based on multiple dimensions, which can be used to fit multiple types of scoring function models and calculate the goodness of fit through a unified function fitting module, determine the standard scoring function form according to the goodness of fit, and generate a scoring template library module for establishing a scoring function template based on the standard scoring function form and function parameters, and store it in the scoring function template library, so as to solve the problems of the existing new urban construction integration project evaluation method, such as the incomparable results caused by the inconsistent dimensions of different indicators, the lack of template and version control mechanism for scoring functions, the lack of adaptive fitting mechanism for scoring function form and parameters, and how to build a multi-dimensional integration project evaluation system that supports the unification of scoring standards and backtracking of the scoring process.

[0026] As a preferred solution of the multi-dimensional new urban construction integration project evaluation system described in the present invention, it includes: building a scoring data set module, a normalized and cleaned scoring value module, a unified function shape fitting module, and a scoring template library generation module.

[0027] The scoring data set construction module is used to collect the original values ​​and scores of the same indicator in multiple new urban construction integration projects to construct an indicator scoring point set.

[0028] The normalized cleaning score value module is used to normalize the score values ​​in the indicator scoring point set and eliminate the influence of outliers.

[0029] The unified function fitting module is used to fit multiple scoring function models and calculate the goodness of fit, and determine the standard scoring function form according to the goodness of fit.

[0030] The scoring template library generation module is used to create a scoring function template based on the standard scoring function form and function parameters, and store it in the scoring function template library.

[0031] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of a new urban construction integration project evaluation method based on multiple dimensions.

[0032] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a multi-dimensional new urban construction integration project evaluation method.

[0033] The present invention provides a multidimensional evaluation method for integrated urban development projects. By constructing a multidimensional indicator matrix and performing normalization, the method achieves a comprehensive quantitative expression of integrated urban development projects across different indicator dimensions. This normalization process unifies data of varying dimensions into a comparable space, effectively avoiding evaluation bias caused by unit differences in the original data. It also provides a unified input basis for subsequent scoring function fitting, thereby improving data consistency and scientific comparison.

[0034] Constructing a scoring point set based on manually scored samples enables the transfer of scoring knowledge into data models. The introduction of an effectiveness screening mechanism ensures the uniformity and diversity of the scoring point distribution, preventing model overfitting to specific intervals or sample imbalance, thereby enabling the construction of a reliable scoring sample set. This effectively improves the training stability and generalization ability of the scoring function model, a prerequisite for high-quality modeling.

[0035] Fitting multiple scoring function models and calculating goodness of fit enables accurate modeling of scoring behaviors for different indicators. Fitting results are ranked by calculating the coefficient of determination, and an acceptance threshold is set to identify the optimal function form. This approach balances scoring function diversity with result accuracy, avoiding the inability of traditional single scoring models to adapt to complex indicator forms and improving the explanatory power and adaptability of the scoring function.

[0036] Storing scoring functions as templates in a function template library enables standardized management and cross-project reuse of scoring models. A version control mechanism ensures traceability of the evolution of scoring rules, and the automatic matching of scoring time call logic enhances the automation and intelligence of the system, achieving the beneficial effects of unified scoring logic and continuous maintainability across multiple projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 This is an overall flow chart of a multi-dimensional new urban construction integration project evaluation method provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0039] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0040] Example 1, with reference to Figure 1 , which is an embodiment of the present invention, provides a new urban construction integration project evaluation method based on multiple dimensions, including:

[0041] S1: Collect the original values ​​and scores of the same indicator in multiple new urban construction integration projects to construct an indicator scoring point set.

[0042] Historical data on new urban construction integration projects is extracted based on the unique identifier of the indicator, and the original value is linked to the corresponding score record. The original value is automatically and synchronously acquired based on on-site equipment collection and data platform upload. The score record is derived from the system scoring record. Each pair of sample data is timestamped and the scoring source is identified. Unstructured samples caused by changes in the scoring mechanism are eliminated from the data, and a sample point set data structure that meets the structural standards is constructed.

[0043] Furthermore, the original value sources include collection by automated sensing equipment deployed on-site, uploading of project data via API interfaces, and real-time synchronization with the construction management system. Score records are derived from automatically calculated scores in the project scoring system, manual scoring records from the expert review system, and weighted scoring results generated by a hybrid scoring mechanism.

[0044] During the point set construction process, data that should be eliminated includes records with inconsistent scoring structures due to adjustments to the scoring mechanism, resulting in changes in the scoring dimensions, weights, or rules of historical projects. Records with missing raw values ​​or any fields in the score value. Records with untraceable scoring or data anomalies.

[0045] Data is stored as a triplet consisting of the original value, score, and scoring source. The original values ​​are dimensionalized and mapped to a standard range. The maximum score and scoring step size are limited for each score. When the metric is sensor device coverage, the original value is expressed as a percentage. The fitting process begins when the number of data points is at least the preset number, and data is grouped and archived according to the project label.

[0046] Furthermore, the standard range of the original value is [0,100], the standard range of the score is 10 points, and the minimum scoring step is 0.5 points.

[0047] Furthermore, the original value refers to the observed value or measured value of the indicator during the actual operation of the project. The sources in the present invention include automated sensing equipment deployed on-site (such as environmental monitors, traffic flow sensors, etc.), data uploaded by the construction management platform, and data from third-party systems connected through API interfaces. The score refers to the evaluation level or quantitative score corresponding to the original value. The sources include the system's automatic scoring results (such as automatic scoring based on the original value through the rule engine), manual evaluation records of the expert scoring system, and the results of the hybrid scoring mechanism, that is, automatic scoring + weighted summary value after manual calibration; each score contains metadata such as the scoring timestamp and scoring method identifier.

[0048] Historical project data includes project ID (unique identifier), indicator code, original value (measured value, data uploaded value), score value (manual or system scoring result), timestamp (time when scoring occurs), scoring source identifier (manual, automatic, mixed), indicator version number (for easy identification of scoring rule updates), and project label (such as ecological type, traffic-oriented type, etc., used for group fitting).

[0049] It should be noted that the present invention S1 realizes the standardized collection and cleaning of evaluation data across multiple projects by constructing a unified structure of indicator scoring point sets. The design idea is to obtain the original values ​​by using device collection and platform synchronization, and associate the system scoring records to form triple samples. At the same time, data quality is ensured through dimensional unification, anomaly elimination and group archiving. It can eliminate the interference caused by differences in scoring mechanisms, improve the reliability and applicability of scoring points, provide high-quality input for subsequent scoring function fitting, and enhance the generalization ability of the model and the comparability of evaluation results.

[0050] S2: Normalize the score values ​​in the indicator scoring point set and eliminate the influence of outliers.

[0051] In the original scoring point set, identify the maximum and minimum scores. If the maximum and minimum scores are equal, all scores are treated as the default median. Otherwise, all scores are standardized and the normalized values ​​are compressed to the preset range based on the compression factor.

[0052] Furthermore, the default median is 0.5, and the normalization formula is expressed as:

[0053] ;

[0054] in, Represents standardized data, which is used to unify the comparison scale of scores of different items. Indicates the minimum score in the data. Indicates the maximum score in the data. Indicates the The score values ​​come from the scoring records of historical new urban construction integration projects.

[0055] After normalization, in order to prevent the scoring results from having boundary problems in subsequent function fitting, a compression factor is introduced , perform boundary compression on the normalized value, expressed as:

[0056] ;

[0057] in, Indicates the data after boundary compression processing and the score value after boundary compression processing. It represents the compression factor, which is set to 0.9 in the present invention.

[0058] Create a distribution view of the score set, calculate the mean and standard deviation of all scores, and identify outliers by determining whether the data point score exceeds the mean by plus or minus three standard deviations. If an outlier is identified, the sample record is excluded from the function fitting calculation, and the original data is retained. For the boundary values ​​of the normalized score equal to 0 and 1, a score perturbation factor is set to pull the value back.

[0059] Furthermore, for the case where the score in the normalized result is close to the boundary (i.e. the normalized value is 0 or 1), in order to prevent the scoring function model from having gradient explosion or fitting distortion problems at the boundary, a score perturbation factor is set to numerically pull back the boundary value. If the data after boundary compression processing If the value is 0, the score perturbation factor is set. If it is 1, the value is set to 1-score perturbation factor. The score perturbation factor range is 0~0.05.

[0060] It should be noted that the design concept of S2 is to normalize the original scoring data and remove outliers to ensure that the scoring data has a unified scale and statistical reliability before function fitting. The scores are uniformly processed through standardized formulas and compression factors to avoid fitting deviations caused by data boundaries. The mean ±3 times the standard deviation is introduced to identify anomalies and enhance data robustness. The disturbance factor is set to pull back the boundary values ​​of 0 and 1 to prevent the fitting function from becoming unstable in the extreme value range. The fitting accuracy and transferability of the scoring function are improved, providing a high-quality data foundation for the subsequent construction of the evaluation model.

[0061] S3: Fit a multi-class scoring function model and calculate the goodness of fit, and determine the standard scoring function form based on the goodness of fit.

[0062] The score point set after normalization and boundary correction is used as the input data set of the fitting function model, and the fitting process is performed as follows.

[0063] The normalized scoring point set is sequentially fitted with linear, logarithmic, and sigmoid function models. The least squares method is used to solve for each model parameter, retaining the sum of squares of the residuals for each function fit and the parameter combination results. All function fitting parameters are calculated by minimizing the error term. After the fitting is complete, the evaluation index is ranked according to its determination coefficient.

[0064] Furthermore, the linear scoring function fitting model fits the normalized original value and the score value into a linear function form and minimizes the residual sum of squares, which is expressed as:

[0065] ;

[0066] in, Indicates the The normalized original value of each scoring point is in the range of [0,1]. Indicates the The normalized score of each scoring point is in the range of [0,1]. Represents the slope coefficient of the linear scoring function. Represents the intercept coefficient of the linear scoring function. The total number of samples in the scoring point set.

[0067] The logarithmic scoring function fitting model fits the scoring points into a logarithmic function form, which is expressed as:

[0068] ;

[0069] in, Represents the scaling coefficient of the logarithmic scoring function. Represents the shift coefficient of the logarithmic scoring function. Indicates the number of valid samples in the scoring point set.

[0070] The S-type scoring function fitting model uses the S-type Sigmoid function for fitting, and uses the least squares method to fit the nonlinear function, which is expressed as:

[0071] ;

[0072] in, Represents the steepness control parameter of the Sigmoid function, which affects the slope of the curve. Indicates the inflection point of the function and the midpoint of the curve. Indicates the total number of samples in the scoring point set.

[0073] The coefficient of determination for each scoring function is calculated. If the coefficient of determination for each model's evaluation metric is greater than or equal to 0.85, the function is considered acceptable. If all model evaluation metric coefficients are less than 0.85, the function is marked as unscorable. Standard scoring functions are archived with a model type label and a list of fitting parameters.

[0074] Furthermore, the coefficient of determination is calculated for each function fitting result as a basis for evaluating the goodness of the function. The closer the coefficient of determination is to 1, the better the fitting effect is:

[0075] ;

[0076] in, Indicates the The coefficient of determination of the class scoring function ( =1 is linear, =2 is the logarithm, =3 for S type). Indicates the Samples in the function The predicted value in . Represents the average of all sample scores. Representation function The number of samples to fit.

[0077] Each scoring function is ranked based on its coefficient of determination (CDR) after fitting. Functions with a CDR ≥ 0.85 are considered valid, and their function type label (e.g., linear, logarithmic, sigmoid) and final fitted parameter set are saved. If no function meets this threshold, the metric is marked as unscorable. The final scoring function is stored in the scoring function template library as a structured entry, including the function type, parameter set, applicable metric ID, and construction timestamp, for subsequent use in new project evaluations.

[0078] Among them, the S3 step fits linear, logarithmic and S-shaped functions to the normalized scoring point set respectively, uses the least squares method to solve the optimal fitting parameters, and uses the determination coefficient as the evaluation index to automatically select the optimal scoring function form. It fully considers the nonlinear characteristics and scoring distribution differences of different indicators in actual projects to avoid subjective errors caused by artificially setting the function form. Unlike the existing technology that usually adopts a single scoring function or empirical scoring method, S3 realizes an automatic fitting and selection mechanism for the scoring function model, which can dynamically adapt to the actual data characteristics of multiple types of indicators and improve the scientificity, adaptability and reliability of the scoring. Especially when dealing with boundary values ​​or nonlinear distribution data, the fitting process is more stable and the scoring model is more explanatory and generalizable.

[0079] S4: Create a scoring function template based on the standard scoring function form and function parameters, and store it in a scoring function template library.

[0080] The selected standard scoring function type and the corresponding parameter results are structurally encapsulated to form a scoring function template object. The scoring function template includes a model type field, a parameter set field, a normalized input range field, a raw indicator field, a scoring output rule field, a creation time field, and an applicable condition field. Each scoring function template is assigned a unique function template ID. The function template ID is generated as a fixed-length string and consists of a function type code, a generation timestamp, and a parameter summary fingerprint.

[0081] Furthermore, the model type field indicates the scoring function type. The parameter set field includes the function parameter values ​​obtained by fitting (such as slope, intercept, steepness coefficient, etc.). The normalized input range field is uniformly set to [0,1] to limit the valid range of the original value after normalization. The original indicator field is used to identify the specific indicator code applicable to this scoring function. The score output rule field describes the output range and accuracy of the score (for example, 0 to 10 points, with a step size of 0.5 points). The creation time field records the timestamp of the template creation. The applicable condition field is an optional field used to identify the applicable scenario of the function, such as project type, scoring mechanism version, etc.

[0082] Furthermore, the template ID generation formula is expressed as:

[0083] ;

[0084] in, Indicates the scoring function template ID. Indicates the model type code, such as LIN (linear), LOG (logarithmic), SIG (S-type). Indicates the build timestamp (in the format of YYYYMMDDHHMMSS). Represents a parameter combination summary.

[0085] All scoring function templates are stored in a scoring function template library as key-value pairs and indexed by indicator code, allowing the scoring engine to call the corresponding template based on indicator type and scoring time. The scoring function template library supports a version management mechanism, generating a new version of the template with each parameter update and recording template change history. When multiple projects reference the same indicator, the scoring function defined in the template library is automatically called, achieving unified scoring standards across projects.

[0086] Furthermore, the scoring function template is stored in the scoring function template library as a key-value pair structure, where the key corresponds to the indicator code and the value is the content of the scoring function template object (including the seven fields mentioned above).

[0087] Supports the coexistence of multiple versions of the same indicator scoring function template, using the version number field to distinguish them.

[0088] To support automatic adaptation of scoring rules and unified calling across projects, the present invention provides the following calling logic:

[0089] Whenever the scoring engine initiates the scoring process for a new project, it first retrieves the list of indicators included in the project. For each indicator, it calls the corresponding key value in the template library. If multiple versions exist, the timestamp of the most recent version is matched based on the scoring time. After returning the scoring function template, the scoring engine executes the scoring function value calculation process based on the function and parameters specified in the template.

[0090] It should be noted that S4 achieves unified management and cross-project calling of scoring functions by structurally encapsulating the types and parameter results of scoring functions, constructing standardized scoring function templates, and using a unique ID coding mechanism to store them in the scoring function template library. The template structure clearly distinguishes fields such as model type, parameter set, scope of application, and version control, making the scoring function traceable, updatable, and shareable, and supporting dynamic matching and loading of the scoring engine on demand. Compared with the existing technology where scoring functions rely on manual settings within the project and lack standard interfaces and cross-project reuse mechanisms, S4 realizes automatic adaptation and version control of scoring standards, solves pain points such as inconsistent scoring, difficult parameter traceability, and inconsistent standard upgrades, and greatly improves the automation, standardization, and horizontal comparison capabilities of the scoring process.

[0091] Example 2 is an embodiment of the present invention, which provides a new urban construction integration project evaluation system based on multiple dimensions, including a scoring data set construction module, a normalized and cleaned scoring value module, a unified function shape fitting module, and a scoring template library generation module.

[0092] The scoring dataset construction module is used to collect the original values ​​and scores of the same indicator in multiple new urban construction integration projects and construct the indicator scoring point set.

[0093] The normalized cleaning score value module is used to normalize the score values ​​in the indicator scoring point set and eliminate the influence of outliers.

[0094] The unified function shape fitting module is used to fit multiple scoring function models and calculate the goodness of fit, and determine the standard scoring function shape based on the goodness of fit.

[0095] The scoring template library generation module is used to create a scoring function template based on the standard scoring function form and function parameters, and store it in the scoring function template library.

Claims

1. A multi-dimensional new urban construction integration project evaluation method, characterized by: include: Collect the original values ​​and scores of the same indicator in multiple new urban construction integration projects to construct an indicator scoring point set; Normalize the score values ​​in the indicator scoring point set and eliminate the influence of outliers; Fit multiple scoring function models and calculate the goodness of fit, and determine the standard scoring function form based on the goodness of fit; Create a scoring function template based on the standard scoring function form and function parameters, and store it in the scoring function template library; The step of establishing a scoring function template based on the standard scoring function form and function parameters includes: making a structured record of the determined function model and parameters; the template content includes the function type, parameter value set, normalized input range, original indicator name and unit, score output rule, function establishment time and applicable conditions; and binding a unique function ID to each template; The standard scoring function shape determined based on the goodness of fit includes: Calculate the coefficient of determination of the evaluation index of each scoring function. If the coefficient of determination of the model evaluation index is greater than or equal to 0.85, it is considered an acceptable fitting function; If the coefficient of determination of all model evaluation indicators is less than 0.85, it will be marked as an unscorable indicator; Standard scoring functions are archived as a model type label and a list of fitted parameters; The scoring function template library includes: All scoring function templates are stored in the scoring function template library as key-value pairs and indexed by indicator code, allowing the scoring engine to call the corresponding template according to indicator type and scoring time; The scoring function template library supports a version management mechanism. A new version of the template is generated each time the parameters are updated, and the template change history is recorded. When multiple projects reference the same indicator, the scoring function defined in the template library is automatically called to achieve unified scoring standards across projects.

2. The multi-dimensional new urban construction integration project evaluation method according to claim 1 is characterized by: The collection of original values ​​and scores of the same indicator in multiple new urban construction integration projects includes: Extract historical new urban construction integration project data based on the unique indicator identifier, and associate the original value with the corresponding score record one by one; The original values ​​are automatically and synchronously acquired based on on-site equipment collection and data platform upload. The score records are derived from the system scoring records. Each pair of sample data has a timestamp and scoring source identifier. Unstructured samples caused by changes in the scoring mechanism are eliminated from the data, and a sample point set data structure that meets the structural standards is constructed.

3. The multi-dimensional new urban construction integration project evaluation method according to claim 2 is characterized by: The construction index scoring point set includes: Stored in triples, the triples include the original value, score, and score source; The original values ​​are dimensionalized and mapped to the standard value range, and the full score and scoring step length are limited for the score records; When the indicator is the sensing device coverage rate, the original value unit is a percentage. When the number of data points is not less than the preset number, the fitting process is entered and the data points are grouped and archived according to the project label.

4. The multi-dimensional new urban construction integration project evaluation method according to claim 3 is characterized by: The normalization of the score values ​​in the indicator scoring point set includes: In the original scoring point set, identify the maximum and minimum scores. If the maximum and minimum scores are equal, all scores are treated as the default median. Otherwise, all scores are standardized and the normalized values ​​are compressed to the preset range based on the compression factor.

5. The multi-dimensional new urban construction integration project evaluation method according to claim 4 is characterized by: The elimination of outlier impacts includes: Create a distribution view of the scoring point set, calculate the mean and standard deviation of all score values, and identify outliers by determining whether the data point score exceeds the mean plus or minus 3 times the standard deviation; If an outlier is identified, the sample record will not be included in the function fitting calculation and the original data will be retained; for the boundary values ​​of the normalized score equal to 0 and 1, the score disturbance factor is set to perform numerical pullback.

6. The multi-dimensional new urban construction integration project evaluation method according to claim 5 is characterized by: The fitting of the multi-class scoring function model and calculating the goodness of fit include: Fit the normalized scoring point set with linear function, logarithmic function and S-shaped function models in turn, use the least squares method to solve the parameters of each model respectively, and retain the sum of squares of the residuals of each group of function fitting and the parameter combination results; During the fitting process of all functions, the parameters are calculated by minimizing the error term. After the fitting is completed, the quality is ranked according to the determination coefficient of the evaluation index.

7. The multi-dimensional new urban construction integration project evaluation method according to claim 6 is characterized by: The step of establishing a scoring function template by using the standard scoring function form and function parameters includes: The selected standard scoring function type and the corresponding parameter results are structurally encapsulated to form a scoring function template object. The scoring function template includes a model type field, a parameter set field, a normalized input range field, an original indicator field, a score output rule field, a creation time field, and an applicable condition field. A unique function template ID is assigned to each scoring function template. The function template ID is generated as a fixed-length string and consists of a combination of function type code, generation timestamp, and parameter summary fingerprint.

8. A multi-dimensional new urban construction integration project evaluation system, using the multi-dimensional new urban construction integration project evaluation method according to any one of claims 1 to 7, characterized in that: It includes the module for building a scoring data set, the module for normalizing and cleaning the scoring value, the module for fitting a unified function shape, and the module for generating a scoring template library; The scoring data set construction module is used to collect the original values ​​and scores of the same indicator in multiple new urban construction integration projects to construct an indicator scoring point set; The normalization and cleaning score value module is used to normalize the score values ​​in the indicator scoring point set and eliminate the influence of outliers; The unified function fitting module is used to fit multiple scoring function models and calculate the goodness of fit, and determine the standard scoring function form according to the goodness of fit; The scoring template library generation module is used to create a scoring function template based on the standard scoring function form and function parameters, and store it in the scoring function template library.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the multi-dimensional new urban construction integration project evaluation method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the multi-dimensional new urban construction integration project evaluation method described in any one of claims 1 to 7 are implemented.

Images