Cost index data platform

Abstract

The invention discloses a cost index data platform, and relates to the technical field of engineering data management. The platform comprises a principal component analysis module, a function model fitting module, a current index acquisition module, a unit price prediction module and a cost calculation module which are in communication connection. The function model fitting module obtains a historical principal component index vector corresponding to each historical project data set by calling the principal component analysis module, and constructs a principal component cost function model corresponding to each historical principal component index vector; and the cost calculation module obtains a cost calculation result according to the current principal component index vector matching model principal component index vector and according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index. According to the invention, complexity of cost data processing is reduced, accuracy and reliability of cost calculation results are ensured, and effective management and control of project cost are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering data management, and in particular to a cost index data platform. Background Art

[0002] In traditional construction projects, project cost management usually relies on rules of thumb and single indicator calculations. This method estimates project costs through simple analogies of project managers' personal experience and historical data. However, as projects grow in size and complexity, this traditional estimation method has gradually shown its limitations. Specifically, it manifests itself in insufficient data analysis, insufficient cost estimation accuracy, and difficulty in coping with changing market conditions and project requirements. These problems not only affect the cost control and decision-making quality of the project, but also increase the possibility of project cost overruns.

[0003] In order to address the limitations of traditional methods, cost management has gradually introduced some advanced technical means in recent years. The most commonly used are statistical analysis methods based on historical data. These methods use regression analysis and other technical means to identify factors and patterns that affect cost from data of past completed projects. Although these technologies provide improvements based on traditional methods, inaccurate cost forecasts and overly complex data processing may still occur in the complex and changing modern project environment. Summary of the invention

[0004] In order to solve the problems existing in the above background technology, the present invention adopts the following technical solutions:

[0005] A cost index data platform includes a communication-connected principal component analysis module, a function model fitting module, a current index acquisition module, a unit price prediction module and a cost calculation module;

[0006] A principal component analysis module is used to obtain a standardized indicator matrix according to a project data group, and obtain a principal component indicator vector according to the standardized indicator matrix;

[0007] The function model fitting module is configured to obtain a number of historical project data groups, obtain the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module, construct a principal component cost function model corresponding to each historical principal component index vector, and store the principal component cost function model and its corresponding model principal component index vector;

[0008] The current index acquisition module is configured to organize the current project data into a current project data group, and obtain the current principal component index vector corresponding to the current project data group by calling the principal component analysis module;

[0009] A unit price prediction module, configured to obtain a predicted unit price corresponding to each principal component indicator in the current principal component indicator vector;

[0010] The cost calculation module is configured to match the model principal component index vector according to the current principal component index vector, and obtain the cost calculation result according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index.

[0011] As a preferred solution, the principal component analysis module includes a preprocessing module, a standardization conversion module, a normalization conversion module, a covariance calculation module and a principal component acquisition module;

[0012] The preprocessing module is used to perform data cleaning and data screening on the project data group to obtain the first project data matrix; the standardization conversion module is used to perform standardization processing on the first project data matrix to obtain the second project data matrix; the normalization conversion module is used to perform normalization processing on the second project data matrix to obtain the third project data matrix; the covariance calculation module is used to obtain the covariance matrix of the third project data matrix; the principal component acquisition module is used to obtain the eigenvectors of the covariance matrix and their corresponding eigenvalues, and select several eigenvectors according to the eigenvalues ​​to construct the principal component indicator vector.

[0013] As a preferred solution, the principal component index vector is expressed as:

[0014] P = [P1, P2, ..., P k ],

[0015]

[0016] Among them, P represents the principal component index vector, P1, P2, ..., P k represents the first to kth principal component index, Z represents the second project data matrix, v j represents the eigenvector of the covariance matrix, λ j represents the eigenvalue corresponding to the eigenvector, m represents the number of eigenvectors of the covariance matrix; k represents the number of eigenvectors of the characteristic matrix; the characteristic matrix is ​​constructed according to the eigenvectors corresponding to the first k largest eigenvalues ​​of the covariance matrix.

[0017] As a preferred solution, the function model fitting module includes a historical data classification module, a historical principal component acquisition module and a genetic fitting module;

[0018] The historical data classification module is used to obtain historical project data and divide the historical project data into several historical project data groups according to the types of attributes;

[0019] The historical principal component acquisition module is used to acquire the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module;

[0020] The genetic fitting module is used to construct a principal component cost function model corresponding to each historical principal component index vector according to the historical principal component index vector and the genetic programming algorithm, and store the principal component cost function model and its corresponding model principal component index vector.

[0021] As a preferred solution, the method of constructing a principal component cost function model corresponding to each historical principal component index vector includes the following steps:

[0022] S1. Obtain the historical principal component index vector corresponding to each historical project data group;

[0023] S2, randomly generate an initial population; each population represents a principal component cost function model;

[0024] S3, evaluating the fitness function value of each population for the historical data set according to the fitness function;

[0025] S4, determining a candidate population for genetic operation according to the fitness function value and the selection algorithm; performing replication, mutation, transposition and / or recombination operations on the candidate population;

[0026] S5. Repeat steps S3 and S4 until the termination condition is met, and store the best performing principal component cost function model and its corresponding historical principal component index vector.

[0027] As a preferred solution, the fitness function is expressed as:

[0028]

[0029] Where E represents the preset range of fitness function, Ai is the actual cost value of the population for the i-th sample of the historical project data set, P i is the cost prediction value of the population for the i-th sample of the historical project data set; n is the number of samples in the historical project data set.

[0030] As a preferred solution, the principal component cost function model adopts a tree structure, and the nodes in the tree structure represent mathematical operators or attribute values ​​corresponding to the principal component indicators; the principal component indicators are obtained through historical principal component indicator vectors, and the attribute values ​​corresponding to the principal component indicators are obtained through historical project data groups.

[0031] As a preferred solution, the unit price prediction module includes a prediction index determination module and a price prediction module;

[0032] The prediction index determination module is used to determine the market pricing index in the current principal component index vector;

[0033] The price prediction module is used to obtain the predicted unit price of each market pricing indicator according to a price prediction model; the price prediction model is based on a machine learning model or a time series analysis model.

[0034] As a preferred solution, the cost calculation module includes a principal component matching module, an attribute value compensation module, a first calculation module and a second calculation module;

[0035] The principal component matching module is used to obtain the model principal component index vector with the highest matching degree according to the current principal component index vector;

[0036] The attribute value compensation module is used to obtain the first attribute value corresponding to each principal component index in the current principal component index vector according to the current project data group;

[0037] The first calculation module is used to compensate the first attribute value according to the predicted unit price corresponding to each principal component index to obtain the second attribute value;

[0038] The second calculation module is used to input the second attribute value corresponding to each principal component index into the principal component cost function model corresponding to the model principal component index vector to obtain the cost calculation result.

[0039] As a preferred solution, the project data group stores a number of project numbers and their corresponding attributes and standard unit prices corresponding to the attributes through a tree structure; the project data group is obtained through a natural language processing model.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] The present invention standardizes and reduces the dimension of data through a principal component analysis module, thereby improving the efficiency of calculation and the accuracy of the result; through a function model fitting module, historical data is grouped and analyzed and a principal component cost function model is constructed, thereby effectively capturing complex data patterns; through a prediction index acquisition module, the predicted unit price corresponding to each principal component index in the current principal component index vector is obtained, thereby providing a necessary prediction basis for cost calculation; through a cost calculation module, the model principal component index vector is matched according to the current principal component index vector, and then the cost calculation result is obtained according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index, thereby reducing the complexity of cost data processing, ensuring the accuracy and reliability of the cost calculation result, and further realizing effective management and control of the project cost.

[0042] The embodiment of the present invention utilizes the evolution mechanism of the genetic algorithm and performs replication, mutation, transposition and recombination operations based on fitness evaluation, thereby ensuring that the final model performs well in terms of error minimization and generating a principal component cost function model that conforms to the actual project.

[0043] The embodiment of the present invention introduces the dynamic changes of the market pricing model, inputs the compensated second attribute value into the principal component cost function model to dynamically compensate the output of the model without modifying the original principal component cost function model, so that the cost calculation result output by the principal component cost function model can dynamically respond to changes in market prices and be more in line with the predicted market situation, thereby improving the robustness and applicability of the model in different market environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0046] Figure 1 A schematic diagram of the structure of a cost index data platform provided by an embodiment of the present invention;

[0047] Figure 2 A schematic diagram of the structure of a function model fitting module provided in an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of a process for constructing a principal component cost function model corresponding to each historical principal component index vector provided by an embodiment of the present invention;

[0049] Figure 4 A schematic diagram of the structure of an electronic device provided in this embodiment. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not 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.

[0051] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0052] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0053] In traditional construction projects, project cost management usually relies on rules of thumb and single indicator calculations. This method estimates project costs through simple analogies of project managers' personal experience and historical data. However, as projects grow in size and complexity, this traditional estimation method has gradually shown its limitations. Specifically, it manifests itself in insufficient data analysis, insufficient cost estimation accuracy, and difficulty in coping with changing market conditions and project requirements. These problems not only affect the cost control and decision-making quality of the project, but also increase the possibility of project cost overruns.

[0054] In order to address the limitations of traditional methods, cost management has gradually introduced some advanced technical means in recent years. The most commonly used are statistical analysis methods based on historical data. These methods use regression analysis and other technical means to identify factors and patterns that affect cost from data of past completed projects. Although these technologies provide improvements on the basis of traditional methods, they may still cause problems such as inaccurate cost forecasts and overly complex data processing in the complex and changing modern project environment.

[0055] Example 1

[0056] like Figure 1 As shown, a cost index data platform includes a principal component analysis module, a function model fitting module, a current index acquisition module, a unit price prediction module and a cost calculation module connected in communication;

[0057] A principal component analysis module is used to obtain a standardized indicator matrix according to a project data group, and obtain a principal component indicator vector according to the standardized indicator matrix;

[0058] The function model fitting module is configured to obtain a number of historical project data groups, obtain the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module, construct a principal component cost function model corresponding to each historical principal component index vector, and store the principal component cost function model and its corresponding model principal component index vector;

[0059] The current index acquisition module is configured to organize the current project data into a current project data group, and obtain the current principal component index vector corresponding to the current project data group by calling the principal component analysis module;

[0060] A unit price prediction module, configured to obtain a predicted unit price corresponding to each principal component indicator in the current principal component indicator vector;

[0061] The cost calculation module is configured to match the model principal component index vector according to the current principal component index vector, and obtain the cost calculation result according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index.

[0062] The present invention standardizes and reduces the dimension of data through a principal component analysis module, thereby improving the efficiency of calculation and the accuracy of the result; through a function model fitting module, historical data are grouped and analyzed and a principal component cost function model is constructed, thereby effectively capturing complex data patterns; through a current indicator acquisition module, the current project data calls the principal component analysis module to obtain the current principal component indicator vector corresponding to the current project data group, thereby ensuring that the current project data and historical data have a consistent analysis process, thereby enhancing the standardization of data processing; through a prediction indicator acquisition module, the predicted unit price corresponding to each principal component indicator in the current principal component indicator vector is obtained, thereby providing necessary prediction basis for cost calculation; through a cost calculation module, the model principal component indicator vector is matched according to the current principal component indicator vector, and then the cost calculation result is obtained according to the principal component cost function model corresponding to the model principal component indicator vector and the predicted unit price corresponding to each principal component indicator, thereby reducing the complexity of cost data processing, ensuring the accuracy and reliability of the cost calculation result, and realizing effective management and control of project cost.

[0063] Specifically, a cost index data platform of the present invention is described in detail through the following contents:

[0064] A cost index data platform includes a communication-connected principal component analysis module, a function model fitting module, a current index acquisition module, a unit price prediction module and a cost calculation module;

[0065] A principal component analysis module is used to obtain a standardized indicator matrix according to a project data group, and obtain a principal component indicator vector according to the standardized indicator matrix;

[0066] The function model fitting module is configured to obtain a number of historical project data groups, obtain the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module, construct a principal component cost function model corresponding to each historical principal component index vector, and store the principal component cost function model and its corresponding model principal component index vector;

[0067] The current index acquisition module is configured to organize the current project data into a current project data group, and obtain the current principal component index vector corresponding to the current project data group by calling the principal component analysis module;

[0068] A unit price prediction module, configured to obtain a predicted unit price corresponding to each principal component indicator in the current principal component indicator vector;

[0069] The cost calculation module is configured to match the model principal component index vector according to the current principal component index vector, and obtain the cost calculation result according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index.

[0070] In the cost index data platform of the present invention, the principal component analysis module is used to process the received project data group into a principal component index vector by applying a standardization method and a principal component analysis method, so that attributes of different scales are comparable, and to reduce the dimension of the data in the high-dimensional space and extract the main principal components, thereby obtaining a principal component index vector representing the project data. The function model fitting module classifies the historical project data according to the attribute type to form several historical project data groups; uses the principal component analysis module to analyze each data group and extracts its historical principal component index vector; then, according to the historical principal component index vector, fits the principal component cost function model that can map the principal component index vector to the cost, and finally stores the fitted principal component cost function model and its corresponding historical principal component index vector for subsequent prediction; the current index acquisition module collects the relevant data of the current project and organizes it into a current project data group that meets the requirements, and then extracts the current principal component index vector of the current project data group by calling the principal component analysis module; the unit price prediction module obtains the price prediction of each principal component index in the current project principal component index vector through corresponding market analysis, historical data or external data sources, which is used for subsequent cost calculation; the cost calculation module matches the model principal component index vector according to the current principal component index vector, and then obtains the cost calculation result according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index.

[0071] Furthermore, the project data group stores a number of project numbers and their corresponding attributes and standard unit prices corresponding to the attributes through a tree structure. When using a tree structure to store the project data group, the project numbers can be connected through a parent-child node relationship. Specifically, in the project data group, the nodes connected through the parent-child node relationship have different hierarchical relationships. A large project (parent node) can contain multiple sub-projects (child nodes). These sub-projects exist as independent work packages, but depend on the overall goal of the large project.

[0072] Each project number can correspond to multiple attributes, such as construction period, environmental indicators, materials, personnel, equipment and other data. The attributes can also be connected through parent-child node relationships. For example, under the "Materials" category, there can be names, quantities, standard unit prices and total costs of different materials.

[0073] In one embodiment, the project data group is obtained through a natural language processing model. Based on natural language processing (NLP), this embodiment can parse unstructured text data such as project documents, contracts, and bids, and store key data in a structured form. In addition, by using semantic analysis and knowledge graph technology, unstructured project data is associated with existing indicators to automatically generate context-related tags and annotations for each project, thereby realizing the storage and management of project data groups through a tree structure.

[0074] In this embodiment, the tree structure can help understand and manage the relationship between complex projects, clarify the affiliation between projects, and facilitate clarifying the hierarchical relationship and attribute distribution of each project in data analysis, thereby optimizing resource allocation and progress tracking.

[0075] Furthermore, the principal component analysis module includes a preprocessing module, a standardization conversion module, a normalization conversion module, a covariance calculation module and a principal component acquisition module;

[0076] The preprocessing module is used to perform data cleaning and data screening on the project data group to obtain the first project data matrix; wherein data cleaning is used to remove noise and outliers in the data through discreteness analysis; data screening is used to select the attributes corresponding to the project number and remove irrelevant or redundant data according to the set standards. For example, the attributes of a project data group include land area, and the preprocessing module will remove outliers (such as obviously wrong area data) and noise (such as temporarily input wrong price) therein, and filter out relevant data (such as removing duplicate or unnecessary data columns) according to the set standards.

[0077] The standardization conversion module is used to standardize the first project data matrix to obtain the second project data matrix; in the project data group, the dimensions of each attribute may be different, and the standardization conversion module converts these attributes into a form with a mean of 0 and a standard deviation of 1, so that each attribute can be compared and analyzed on the same scale. Among them, the elements of the second project data matrix are represented as:

[0078]

[0079] Among them, z ij represents the attribute value corresponding to the jth attribute of the i-th sample in the second project data matrix, x ijrepresents the attribute value corresponding to the jth attribute of the i-th sample in the first project data matrix, μ j represents the attribute average value of the jth attribute in the first item data matrix; σ j Represents the attribute standard deviation value of the jth attribute in the first item data matrix.

[0080] The normalization conversion module is used to normalize the second project data matrix to obtain the third project data matrix; the normalized data matrix is ​​normalized to convert the data values ​​into a unified range of [0,1]. The data after normalization may still have different magnitudes. The normalization conversion module maps these data to the range of 0 to 1, making the subsequent analysis more stable and accurate. Among them, the elements of the third project data matrix are expressed as:

[0081]

[0082] Among them, z ij represents the attribute value corresponding to the jth attribute of the i-th sample in the second project data matrix, z max,j represents the maximum value of the attribute corresponding to the jth attribute in the second item data matrix, z min,j Represents the minimum attribute value corresponding to the j-th attribute in the second item data matrix.

[0083] The covariance calculation module is used to obtain the covariance matrix of the third project data matrix; based on the third project data matrix, a covariance matrix composed of covariances between attributes is obtained to reveal the correlation between the attributes.

[0084] Wherein, the covariance matrix is ​​expressed as:

[0085]

[0086] Among them, C represents the covariance matrix, m represents the number of samples, R represents the third project data matrix, and T represents the transposed matrix.

[0087] The principal component acquisition module is used to obtain the eigenvectors of the covariance matrix and their corresponding eigenvalues, and select several eigenvectors according to the eigenvalues ​​to construct the principal component index vector. The principal component acquisition module allows us to select the first few principal components (eigenvectors corresponding to the largest eigenvalues) by analyzing the eigenvalues. The principal component can represent the main influencing factors in the project data group, reduce the dimensions without losing too much data information, and thus simplify the analysis model. Among them, the eigenvectors of the covariance matrix are obtained using the formula:

[0088] Cv j =λ j v j ,

[0089] Where C represents the covariance matrix, v j represents the eigenvector of the covariance matrix, λ j Represents the eigenvalue corresponding to the eigenvector, j represents the characteristic subscript of the covariance matrix, j=1,2,...,m, m represents the number of eigenvectors of the covariance matrix.

[0090] Based on the eigenvectors and their corresponding eigenvalues ​​obtained in the above steps, the eigenvectors are sorted from large to small according to the eigenvalues, and the characteristic matrix V is constructed according to the eigenvectors corresponding to the first k largest eigenvalues. k , and then transform the third project data matrix into the space of the feature matrix to obtain the principal component indicator vector.

[0091] Among them, the feature matrix V k The characteristic matrix V is composed of the eigenvectors corresponding to the first k largest eigenvalues. k The elements in represent the original features. The principal component index vector is expressed as:

[0092] P=[P1,P2,...,P k ],

[0093]

[0094] Among them, P represents the principal component index vector, P1, P2, ..., P k represents the first to kth principal component index, Z represents the second project data matrix, v j represents the eigenvector of the covariance matrix, λ j represents the eigenvalue corresponding to the eigenvector, m represents the number of eigenvectors of the covariance matrix; k represents the number of eigenvectors of the characteristic matrix; the characteristic matrix is ​​constructed according to the eigenvectors corresponding to the first k largest eigenvalues ​​of the covariance matrix.

[0095] The embodiment of the present invention is based on the principal component analysis module, which converts high-dimensional raw data into low-dimensional principal component index vectors, retains key features, and discards redundant information, which helps to improve the efficiency and accuracy of subsequent function model fitting and cost prediction. In the subsequent model fitting process, the function model fitting module uses the low-dimensional principal component index vectors to construct cost function models, which can more accurately reflect the hidden patterns and trends in historical data. The current indicator acquisition module standardizes and projects the data of the current project into the principal component space to generate the principal component index vector of the current project. Finally, the unit price prediction module and the cost calculation module use the principal component index vector to calculate the cost of the new project through the existing cost function model.

[0096] Furthermore, if Figure 2As shown, the function model fitting module includes a historical data classification module, a historical principal component acquisition module and a genetic fitting module;

[0097] The historical data classification module is used to obtain historical project data and divide the historical project data into several historical project data groups according to the type of attributes. For example, for the historical project data of a series of construction projects, the following key attributes are included: project number, project type (such as residential, commercial office, industry), geographical location (city or region), building area, material type (such as reinforced concrete, steel structure, wooden structure), start year, completion year, total cost, etc.; for the above historical project data, according to the above key attributes, the historical project data is divided into several historical project data groups to ensure that the projects in each historical project data group have similarities in these key attributes. Through the historical data classification module, a large amount of complex historical project data can be effectively classified into project data groups with their own similarities. This makes the subsequent principal component analysis more accurate and efficient, reduces the interference of noise, and improves the accuracy of model fitting.

[0098] The historical principal component acquisition module is used to obtain the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module. By calling the principal component analysis module, each historical project data group is subjected to principal component analysis processing, and the principal component index vector of each historical project data group is generated, and the principal component index vector is transmitted to the genetic fitting module for further analysis and model fitting. Through the historical principal component acquisition module, the scheme performs principal component analysis on each project data group and extracts the main features. This not only reduces the dimension of the data and improves the computational efficiency, but also improves the interpretability of the data, so that key influencing factors can be identified and quantified.

[0099] The genetic fitting module is used to construct a principal component cost function model corresponding to each historical principal component index vector according to the historical principal component index vector and the genetic programming algorithm, and store the principal component cost function model and its corresponding model principal component index vector. Among them, for the convenience of distinction, the historical principal component index vector used to construct the principal component cost function model is recorded as the model principal component index vector in the present invention. The genetic fitting module of this embodiment automatically generates a principal component cost function model that meets the requirements through a genetic programming algorithm, and can find the optimal or approximately optimal solution in a wide search space, reduce human intervention and prejudice, and improve the accuracy and reliability of the cost estimation model.

[0100] Among them, genetic programming (GP) is an optimization technology based on genetic algorithms, whose goal is to automatically generate computer programs or function models that meet specific requirements. By simulating the process of natural selection and biological evolution, genetic programming searches for optimal or near-optimal solutions in a huge problem space.

[0101] Furthermore, if Figure 3 As shown, in the cost index data platform of the present invention, the construction of the principal component cost function model corresponding to each historical principal component index vector includes the following steps:

[0102] S1. Obtain the historical principal component index vector corresponding to each historical project data group;

[0103] S2. Randomly generate an initial population; each population represents a principal component cost function model; the principal component cost function model adopts a tree structure, and the nodes in the tree structure represent mathematical operators or attribute values ​​corresponding to the principal component indicators; the principal component indicators are obtained through historical principal component indicator vectors, and the attribute values ​​corresponding to the principal component indicators are obtained through historical project data groups; mathematical operators may include addition, multiplication, exponentiation, etc.

[0104] S3. Evaluate the fitness function value of each population for the historical data set according to the fitness function; the fitness function is used to evaluate the error between the actual cost value and the predicted cost value, that is, the fitness function value. Specifically, the fitness function is expressed as:

[0105]

[0106] Among them, E represents the preset range of fitness function values, A i is the actual cost value of the population for the i-th sample of the historical project data set, P i is the cost prediction value of the population for the i-th sample of the historical project data set; n is the number of samples in the historical project data set.

[0107] In this embodiment, limiting the fitness value can prevent the problem of numerical instability caused by abnormally large or small fitness values, and in the fitness-based selection process (such as roulette selection), limiting the range ensures that each individual has a relatively balanced chance to enter the next generation, avoiding abnormally high fitness individuals dominating the selection process.

[0108] S4. Determine the candidate population for genetic operation according to the fitness function value and the selection algorithm; perform replication, mutation, transposition and / or recombination operations on the candidate population; wherein the selection algorithm adopts roulette selection or tournament selection. Replication is to copy the individuals in the candidate population directly to the next generation, that is, without any modification to them, but to retain the current individuals; the replication operation ensures that the high-quality individuals are retained so that the excellent characteristics can be continued in the next generation. In the specific implementation, the replication is performed according to the fitness selection, and the probability of individuals with high fitness being replicated is relatively high. Mutation is a random modification of the gene sequence of a single individual, which can introduce new genotypes, increase the diversity of the population, and prevent premature convergence to the local optimal solution. Recombination, also known as recombination, is the exchange of gene fragments between two or more individuals to generate new individuals; recombination includes single-point recombination, two-point recombination and gene recombination; single-point recombination is to select a recombination point and exchange the genes after this point; two-point recombination is to select two recombination points and exchange the gene sequences between the two points; gene recombination is to select each gene position from both parents with a certain probability.

[0109] S5, repeating steps S3 and S4 until the termination condition is met, and storing the best performing principal component cost function model and its corresponding historical principal component index vector. Accordingly, each principal component cost function model can accurately simulate the relationship between the historical project data set and the cost, and can be used for project cost prediction.

[0110] This embodiment utilizes the evolution mechanism of the genetic algorithm to perform replication, mutation, transposition and recombination operations based on fitness evaluation, ensuring that the final model performs well in terms of error minimization and generates a principal component cost function model that conforms to the actual project.

[0111] Further, the unit price prediction module includes a prediction index determination module and a price prediction module;

[0112] The prediction index determination module is used to determine the market pricing index in the current principal component index vector; in this embodiment, the principal component index can be set to a quota pricing or market pricing mode. Quota pricing is a method of calculating the cost based on pre-established quota standards, unit labor hours, unit material consumption, etc., combined with the basic price level in a certain period; market pricing is a method of calculating the project cost in real time based on the actual material prices, labor costs, machinery rental fees, etc. in the current market. The prediction index determination module selects indicators that use the market pricing mode and are sensitive to market price changes from the current principal component index vector.

[0113] The price prediction module is used to obtain the predicted unit price of each market pricing indicator according to the price prediction model; wherein the price prediction model can be based on machine learning or time series analysis model, input the historical unit price time series corresponding to the market pricing indicator into the price prediction model, output the corresponding predicted unit price curve, and obtain the predicted unit price corresponding to the market pricing time point set in the predicted unit price curve. The predicted unit price of each market pricing indicator and the standard unit price corresponding to the quota pricing indicator can be combined to obtain the complete predicted unit price corresponding to each principal component indicator.

[0114] Furthermore, the cost calculation module includes a principal component matching module, an attribute value compensation module, a first calculation module and a second calculation module;

[0115] The principal component matching module is used to obtain the model principal component index vector with the highest matching degree according to the current principal component index vector; the principal component matching module is responsible for finding the principal component index vector of the historical model that best matches the current principal component index vector, which is specifically achieved by calculating cosine similarity and the like. The goal is to find the historical model that is closest to the characteristics of the current project so that further calculations can be performed using its corresponding cost function.

[0116] The attribute value compensation module is used to obtain the first attribute value corresponding to each principal component indicator in the current principal component indicator vector according to the current project data group; the first attribute value can be understood as the initial measurement of the current project on each principal component indicator, and these attribute values ​​are used for further calculation and compensation.

[0117] The first calculation module is used to compensate the first attribute value according to the predicted unit price corresponding to each principal component index to obtain the second attribute value; specifically, the first calculation module is used to calculate the error ratio between the predicted unit price and the standard unit price of each principal component index, and the error ratio represents the change of the predicted value relative to the standard value. Among them, the standard unit price is determined when the function model fitting module constructs the principal component cost function model, which represents the uniform price of a certain material or service in the historical project data group within a set time period. Then, the error ratio is multiplied by the first attribute value to obtain the second attribute value corresponding to each principal component index. The second attribute value is the compensated value.

[0118] The second calculation module is used to input the second attribute value corresponding to each principal component index into the principal component cost function model corresponding to the model principal component index vector to obtain the cost calculation result. The second attribute value corresponding to each principal component index is input into the principal component cost function model of the matching model principal component index vector to calculate the final cost calculation result. The cost calculation result obtained based on the second attribute value is a more accurate cost estimate after market dynamic adjustment.

[0119] In this embodiment, the principal component cost function model predicts the cost calculation result according to the attribute value corresponding to the principal component index. Since the input of the principal component cost function model is the attribute value, it is impossible to compensate the error of the model output by inputting the predicted unit price and the standard unit price. Therefore, the embodiment of the present invention obtains the first attribute value through the current project data group, and then calculates the error ratio of the predicted unit price and the standard unit price and the product of the first attribute value as the second attribute value, so as to dynamically compensate the output of the model without modifying the original model.

[0120] This embodiment introduces the dynamic changes of the market pricing model through the attribute value compensation module and the first calculation module, and inputs the compensated second attribute value into the principal component cost function model to dynamically compensate the output of the model without modifying the original model, so that the cost calculation result output by the principal component cost function model can dynamically respond to changes in market prices and be more in line with the predicted market situation, thereby improving the robustness and applicability of the model in different market environments.

[0121] It should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, each functional module can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0122] Example 2

[0123] An electronic device 2, such as Figure 4 As shown, a processor 21 and a memory 22, the memory 22 is used to store computer program codes, the computer program codes include computer instructions, when the processor 21 executes the computer instructions, the electronic device executes the above-mentioned cost index data platform.

[0124] The electronic device 2 includes a processor 21, a memory 22, an output device 23, and an input device 24. The processor 21, the memory 22, the input device 24, and the output device 23 are coupled via a connector, and the connector includes various interfaces, transmission lines, or buses, etc., which are not limited in the embodiments of the present invention. It should be understood that in various embodiments of the present invention, coupling refers to mutual connection in a specific manner, including direct connection or indirect connection through other devices, for example, through various interfaces, transmission lines, buses, etc.

[0125] The processor 21 may be one or more graphics processing units (GPUs). When the processor 21 is a GPU, the GPU may be a single-core GPU or a multi-core GPU. Optionally, the processor 21 may be a processor group consisting of multiple GPUs, and the multiple processors are coupled to each other via one or more buses. Optionally, the processor may also be other types of processors, etc., which are not limited in the embodiments of the present invention.

[0126] The memory 22 can be used to store computer program instructions and various computer program codes including program codes for executing the scheme of the present invention. Optionally, the memory includes but is not limited to random access memory (RAM), read-only memory (ROM), erasable programmable read only memory (EPROM), or portable read only memory (CD-ROM), which is used for related instructions and data.

[0127] The input device 24 is used to input data and / or signals, and the output device 23 is used to output data and / or signals. The output device 23 and the input device 24 can be independent devices or an integrated device.

[0128] The present application achieves preliminary risk assessment by fitting the generalized autoregressive conditional heteroskedasticity model corresponding to each digital asset, obtaining the conditional variance of each digital asset at several set time points according to the generalized autoregressive conditional heteroskedasticity model, and then calculating the value at risk; by constructing a first dependency matrix between the values ​​at risk of each digital asset, an error compensation model is constructed and trained according to risk influencing factors and an error analysis matrix, and the first dependency matrix is ​​corrected according to the error compensation model to obtain a second dependency matrix, so as to achieve error compensation for the dependency, thereby achieving accurate dependency prediction; by combining the second dependency matrix with the generalized autoregressive conditional heteroskedasticity model of each digital asset, accurate risk prediction of the digital assets can be achieved, thereby improving the overall risk prediction capability.

[0129] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A cost index data platform, characterized by: It includes a principal component analysis module, a function model fitting module, a current index acquisition module, a unit price prediction module and a cost calculation module connected by communication; The principal component analysis module is used to obtain a standardized indicator matrix according to the project data group, and obtain a principal component indicator vector according to the standardized indicator matrix; The function model fitting module is configured to obtain a number of historical project data groups, obtain the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module, construct a principal component cost function model corresponding to each historical principal component index vector, and store the principal component cost function model and its corresponding model principal component index vector; The current index acquisition module is configured to organize the current project data into a current project data group, and obtain the current principal component index vector corresponding to the current project data group by calling the principal component analysis module; The unit price prediction module is configured to obtain the predicted unit price corresponding to each principal component indicator in the current principal component indicator vector; The cost calculation module is configured to match the model principal component index vector according to the current principal component index vector, and obtain the cost calculation result according to the principal component cost function model corresponding to the model principal component index vector and the predicted unit price corresponding to each principal component index.

2. A cost index data platform according to claim 1, characterized in that: The principal component analysis module includes a preprocessing module, a standardization conversion module, a normalization conversion module, a covariance calculation module and a principal component acquisition module; The preprocessing module is used to perform data cleaning and data screening on the project data group to obtain the first project data matrix; the standardization conversion module is used to perform standardization processing on the first project data matrix to obtain the second project data matrix; the normalization conversion module is used to perform normalization processing on the second project data matrix to obtain the third project data matrix; the covariance calculation module is used to obtain the covariance matrix of the third project data matrix; the principal component acquisition module is used to obtain the eigenvectors of the covariance matrix and their corresponding eigenvalues, and select several eigenvectors according to the eigenvalues ​​to construct the principal component indicator vector.

3. A cost index data platform according to claim 2, characterized in that: The principal component index vector is expressed as: P=[P1,P2,...,P k ], Among them, P represents the principal component index vector, P1, P2, ..., P k represents the first to kth principal component index, Z represents the second project data matrix, v j represents the eigenvector of the covariance matrix, λ j represents the eigenvalue corresponding to the eigenvector, m represents the number of eigenvectors of the covariance matrix; k represents the number of eigenvectors of the characteristic matrix; the characteristic matrix is ​​constructed according to the eigenvectors corresponding to the first k largest eigenvalues ​​of the covariance matrix.

4. A cost index data platform according to claim 1, characterized in that: The function model fitting module includes a historical data classification module, a historical principal component acquisition module and a genetic fitting module; The historical data classification module is used to obtain historical project data and divide the historical project data into several historical project data groups according to the type of attributes; the historical principal component acquisition module is used to obtain the historical principal component index vector corresponding to each historical project data group by calling the principal component analysis module; the genetic fitting module is used to construct a principal component cost function model corresponding to each historical principal component index vector based on the historical principal component index vector and the genetic programming algorithm, and store the principal component cost function model and its corresponding model principal component index vector.

5. A cost index data platform according to claim 4, characterized in that: The method of constructing a principal component cost function model corresponding to each historical principal component index vector comprises the following steps: S1. Obtain the historical principal component index vector corresponding to each historical project data group; S2, randomly generate an initial population; each population represents a principal component cost function model; S3, evaluating the fitness function value of each population for the historical data set according to the fitness function; S4, determining a candidate population for genetic operation according to the fitness function value and the selection algorithm; performing replication, mutation, transposition and / or recombination operations on the candidate population; S5. Repeat steps S3 and S4 until the termination condition is met, and store the best performing principal component cost function model and its corresponding historical principal component index vector.

6. A cost index data platform according to claim 5, characterized in that: The fitness function is expressed as: Among them, E represents the preset fitness function value range, A i is the actual cost value of the population for the i-th sample of the historical project data set, P i is the cost prediction value of the population for the i-th sample of the historical project data set; n is the number of samples in the historical project data set.

7. A cost index data platform according to claim 5, characterized in that: The principal component cost function model adopts a tree structure, and the nodes in the tree structure represent mathematical operators or attribute values ​​corresponding to the principal component indicators; the principal component indicators are obtained through historical principal component indicator vectors, and the attribute values ​​corresponding to the principal component indicators are obtained through historical project data groups.

8. The cost index data platform according to claim 1, characterized in that: The unit price prediction module includes a prediction index determination module and a price prediction module; The prediction index determination module is used to determine the market pricing index in the current principal component index vector; The price prediction module is used to obtain the predicted unit price of each market pricing indicator according to a price prediction model; the price prediction model is based on a machine learning model or a time series analysis model.

9. A cost index data platform according to claim 1, characterized in that: The cost calculation module includes a principal component matching module, an attribute value compensation module, a first calculation module and a second calculation module; The principal component matching module is used to obtain the model principal component index vector with the highest matching degree according to the current principal component index vector; The attribute value compensation module is used to obtain the first attribute value corresponding to each principal component index in the current principal component index vector according to the current project data group; The first calculation module is used to compensate the first attribute value according to the predicted unit price corresponding to each principal component index to obtain the second attribute value; The second calculation module is used to input the second attribute value corresponding to each principal component index into the principal component cost function model corresponding to the model principal component index vector to obtain the cost calculation result.

10. A cost index data platform according to claim 1, characterized in that: The project data group stores a number of project numbers and their corresponding attributes and standard unit prices corresponding to the attributes through a tree structure; the project data group is obtained through a natural language processing model.

Images