Civil engineering cost data analysis method and system
By optimizing the fully connected layer through transfer learning strategy and gradient descent method, a civil construction cost data analysis model was constructed, which solved the problem of long training time in the existing technology and improved the efficiency of civil construction cost analysis.
Patent Information
- Application Number
- CN202510882876.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-28
- Publication Date
- 2025-10-10
AI Technical Summary
In the existing technology, training deep learning network models for construction cost analysis is time-consuming and inefficient, which affects the progress of civil engineering projects.
A transfer learning strategy is adopted to replace the fully connected layer of the engineering data recognition model. The target cost training samples are used for transfer learning training to construct a cost data analysis model, including feature extraction and mapping processing of the fully connected layer, and the gradient descent method is used to adjust the parameters.
It greatly saves the training time cost of the cost data analysis model and improves the efficiency of obtaining cost analysis results.
Smart Images
Figure CN120765286A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application relates to the technical field of civil construction cost data analysis, and in particular to a civil construction cost data analysis method and system. BACKGROUND
[0002] Civil engineering is a building activity based on civil engineering technology, which covers building structure design, construction management and whole life cycle maintenance. With the development of science and technology, a deep learning network model can be used for data analysis in the field of civil engineering, such as construction cost analysis, project construction duration analysis, construction facility stability analysis, geological analysis of civil engineering areas, and environmental impact analysis of civil engineering areas. However, the deep learning network model applied to different data analysis needs to be trained for a long time, which will seriously affect the progress of civil engineering. Therefore, in the prior art, there is a technical defect of long training time and low training efficiency in the process of training a deep learning network model for construction cost analysis. SUMMARY
[0003] Therefore, the purpose of the present application is to provide a civil construction cost data analysis method and system which can overcome the shortcomings of the prior art.
[0004] In order to achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows:
[0005] A civil construction cost data analysis method comprises the following steps:
[0006] Obtaining the construction project type and project cost data of a target land;
[0007] According to the preset cost analysis target and the construction project type, a plurality of civil engineering training samples are screened to obtain a plurality of target cost training samples for cost analysis of the same construction project type; the target cost training sample comprises historical project cost data and historical cost analysis results;
[0008] Based on a transfer learning strategy, a cost data analysis model is obtained by performing transfer learning training according to the plurality of target cost training samples and a pre-trained engineering data recognition model; the engineering data recognition model is a deep learning model trained according to the plurality of civil engineering training samples;
[0009] The project cost data is input into the cost data analysis model to obtain the cost analysis result of the target land.
[0010] Compared with the prior art, the present application has the following advantages:
[0011] The civil construction cost data analysis method of the present application can filter target cost training samples for cost analysis of the same construction project type from a plurality of civil construction training samples according to the construction project type of the target land and the preset cost analysis target, and then perform transfer learning training based on the plurality of target cost training samples and a pre-trained engineering data recognition model to obtain a cost data analysis model, which can greatly save the training time cost of the cost data analysis model. The project cost data is input into the cost data analysis model to obtain the cost analysis result of the target land, which can reduce the time cost of obtaining the cost analysis result of the target land and improve the efficiency of obtaining the cost analysis result.
[0012] As an implementation form, the step of performing transfer learning training based on the plurality of target cost training samples and the pre-trained engineering data recognition model includes:
[0013] replacing the fully connected layer of the engineering data recognition model to obtain an initial analysis model;
[0014] performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model.
[0015] In the present embodiment, by replacing the fully connected layer of the engineering data recognition model and then performing transfer learning training according to the target cost training samples, the model parameters that need to be learned and trained can be reduced, and the training time cost of the cost data analysis model can be saved.
[0016] As an implementation form, the initial analysis model includes a feature extraction layer and a fully connected layer; wherein the feature extraction is the feature extraction layer of the engineering data recognition model;
[0017] The step of performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model includes:
[0018] inputting the plurality of target cost training samples into the feature extraction layer for feature extraction processing to obtain cost features of the plurality of samples;
[0019] inputting the cost features of the plurality of samples into the fully connected layer for mapping processing to obtain a plurality of prediction analysis results;
[0020] training the parameters of the fully connected layer according to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results to obtain the cost data analysis model.
[0021] In the embodiment, the full connection layer is trained according to historical cost analysis results of a plurality of target cost training samples and a plurality of prediction analysis results, and the cost data analysis model can be quickly obtained.
[0022] As an implementation form, the step of training parameters of the full connection layer according to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results to obtain the cost data analysis model comprises:
[0023] A result loss function is constructed according to the historical cost analysis results of the plurality of target cost training samples and the corresponding plurality of prediction analysis results.
[0024] The value of the loss function is minimized by using a gradient descent method, and the parameters of the full connection layer are adjusted.
[0025] The cost data analysis model is obtained according to the feature extraction layer and the full connection layer with the adjusted parameters.
[0026] In the embodiment, the value of the loss function is minimized by using the gradient descent method to adjust the parameters of the full connection layer, and the difference between the prediction analysis result and the historical cost analysis result can be reduced.
[0027] As an implementation form, the project cost data comprises construction project content, construction project building materials and construction project cost; and the historical project cost data comprises historical project content, historical project building materials, historical project cost and historical project cost analysis result.
[0028] A civil engineering cost data analysis system comprises:
[0029] A project information acquisition module is configured to acquire a construction project type and project cost data of a target land.
[0030] A training sample screening module is configured to screen a plurality of civil engineering training samples according to a preset cost analysis target and the construction project type to obtain a plurality of target cost training samples of the same construction project type for cost analysis; the target cost training sample comprises historical project cost data and historical cost analysis result.
[0031] A transfer learning training module is configured to perform transfer learning training based on a transfer learning strategy according to the plurality of target cost training samples and a pre-trained engineering data recognition model to obtain a cost data analysis model; the engineering data recognition model is a deep learning model trained according to the plurality of civil engineering training samples.
[0032] A cost analysis result acquisition module is configured to input the project cost data into the cost data analysis model to obtain a cost analysis result of the target land.
[0033] Compared with the prior art, the application has the following advantages:
[0034] The civil construction cost data analysis system can filter target cost training samples for cost analysis of the same construction project type from a plurality of civil construction training samples according to the construction project type of the target land and a preset cost analysis target, and then perform transfer learning training based on a pre-trained engineering data recognition model according to the plurality of target cost training samples and the pre-trained engineering data recognition model to obtain a cost data analysis model, which can greatly save the training time cost of the cost data analysis model. The project cost data is input into the cost data analysis model to obtain the cost analysis result of the target land, which can reduce the time cost of obtaining the cost analysis result of the target land and improve the efficiency of obtaining the cost analysis result.
[0035] As an implementation form, the transfer learning training module is configured to perform the following steps:
[0036] Replace the fully connected layer of the engineering data recognition model to obtain an initial analysis model;
[0037] Perform transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model.
[0038] In this embodiment, by replacing the fully connected layer of the engineering data recognition model and then performing transfer learning training according to the target cost training samples, the model parameters that need to be learned and trained can be reduced, and the training time cost of the cost data analysis model can be saved.
[0039] As an implementation form, the initial analysis model includes a feature extraction layer and a fully connected layer; wherein the feature extraction layer is the feature extraction layer of the engineering data recognition model.
[0040] The step of performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model includes:
[0041] Input the plurality of target cost training samples into the feature extraction layer for feature extraction processing to obtain cost features of the plurality of samples;
[0042] Input the cost features of the plurality of samples into the fully connected layer for mapping processing to obtain a plurality of prediction analysis results;
[0043] According to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results, train the parameters of the fully connected layer to obtain the cost data analysis model.
[0044] In the embodiment, the full connection layer is trained according to the historical cost analysis results of the several target cost training samples and the several prediction analysis results of the full connection layer output, and the cost data analysis model can be quickly obtained.
[0045] As an implementation form, the step of training the parameters of the full connection layer according to the historical cost analysis results of the several target cost training samples and the several prediction analysis results to obtain the cost data analysis model comprises:
[0046] According to the historical cost analysis results of the several target cost training samples and the corresponding several prediction analysis results, a result loss function is constructed.
[0047] The value of the loss function is minimized by using the gradient descent method, and the parameters of the full connection layer are adjusted.
[0048] According to the feature extraction layer and the full connection layer with the adjusted parameters, the cost data analysis model is obtained.
[0049] In the embodiment, the value of the loss function is minimized by using the gradient descent method, and the parameters of the full connection layer are adjusted, so that the difference between the prediction analysis result and the historical cost analysis result can be reduced.
[0050] As an implementation form, the project cost data comprises construction project content, construction project building materials and construction project cost; and the historical project cost data comprises historical project content, historical project building materials, historical project cost and historical project cost analysis result.
[0051] In order to better understand and implement, the present application is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS
[0052] Figure 1 Flow chart of the civil engineering cost data analysis method of one embodiment of the present application;
[0053] Figure 2 Flow chart of step S3 of the civil engineering cost data analysis method of one embodiment of the present application;
[0054] Figure 3 Module connection schematic diagram of the civil engineering cost data analysis system of one embodiment of the present application;
[0055] 10, project information acquisition module; 20, training sample screening module; 30, transfer learning training module; 40, cost analysis result acquisition module. DETAILED DESCRIPTION
[0056] In order to make the purposes, technical solutions and advantages of the present application clearer, the following further describes the embodiments of the present application with reference to the accompanying drawings.
[0057] It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.
[0058] When the following description refers to the accompanying drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and do not necessarily describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances. The singular forms "a", "an" and "the" used in the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. The word "if" used herein can be interpreted as "when" or "when" or "in response to determining".
[0059] In addition, in the description of the present application, "multiple" means two or more, unless otherwise specified. The association between the objects described by "and / or" can represent three possible relationships, for example, A and / or B can represent the three cases of A alone, A and B together, and B alone. The character " / " generally represents an "or" relationship between the associated objects before and after it.
[0060] Please refer to Figure 1 which is a flowchart of the civil construction cost data analysis method of the first embodiment of the present application, the method comprising:
[0061] S1: obtaining the construction project type and project cost data of the target land.
[0062] Among them, the target land refers to the land corresponding to the to-be-constructed project, wherein the construction project type includes residential engineering, commercial land engineering, road engineering, water engineering, canal engineering and flood control engineering, etc. The project cost data refers to the construction project content and project cost data recorded in a plurality of construction planning documents.
[0063] S2: filtering a plurality of civil engineering training samples according to a preset cost analysis target and the construction project type to obtain a plurality of target cost training samples for cost analysis of the same construction project type; the target cost training sample includes historical project cost data and historical cost analysis results.
[0064] The plurality of civil engineering training samples refer to civil engineering training samples corresponding to different construction project types and different analysis targets, and the analysis targets of the civil engineering training samples include construction cost, project construction duration, construction facility stability, geology of the civil engineering area, environmental impact of the civil engineering area, and the like.
[0065] S3: Based on the transfer learning strategy, the transfer learning training is performed according to the plurality of target cost training samples and a pre-trained engineering data recognition model to obtain a cost data analysis model; the engineering data recognition model is a deep learning model trained according to the plurality of civil engineering training samples.
[0066] The engineering data recognition model is a deep learning model trained according to the plurality of civil engineering training samples and used for engineering data feature extraction. Since the plurality of civil engineering training samples involve a large amount of data, the pre-trained engineering data recognition model is a deep learning model that needs to be trained for a long time and can accurately extract engineering data features. However, the engineering data recognition model is only used to extract engineering data features, and therefore does not need to limit the construction project types and analysis targets of the civil engineering training samples. That is, the engineering data recognition model can be trained in advance without determining the construction project types of the target land and the preset analysis targets.
[0067] S4: The project cost data is input into the cost data analysis model to obtain a cost analysis result of the target land.
[0068] The cost analysis result refers to an analysis result obtained by analyzing the project cost data of each construction plan corresponding to the target land, and includes the advantages and disadvantages of the project cost data and a score.
[0069] Compared with the prior art, the civil engineering cost data analysis method can select target cost training samples for cost analysis of the same construction project type from the plurality of civil engineering training samples according to the construction project type of the target land and the preset cost analysis target, and then perform transfer learning training according to the plurality of target cost training samples and a pre-trained engineering data recognition model based on the transfer learning to obtain a cost data analysis model, which can greatly save the training time cost of the cost data analysis model. Then, the project cost data is input into the cost data analysis model to obtain a cost analysis result of the target land, which can reduce the time cost of obtaining the cost analysis result of the target land and improve the efficiency of obtaining the cost analysis result.
[0070] Please refer to Figure 2In an example, the S3 comprises:
[0071] S31: replacing the fully connected layer of the engineering data recognition model to obtain an initial analysis model.
[0072] S32: performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model.
[0073] In the example, by replacing the fully connected layer of the engineering data recognition model and then performing transfer learning training according to the target cost training samples, the model parameters that need to be learned and trained can be reduced, and the training time cost of the cost data analysis model can be saved.
[0074] In an example, the initial analysis model comprises a feature extraction layer and a fully connected layer; and the feature extraction layer is the feature extraction layer of the engineering data recognition model.
[0075] S32: performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model.
[0076] S321: inputting the plurality of target cost training samples into the feature extraction layer for feature extraction processing to obtain cost features of the plurality of samples.
[0077] S322: inputting the cost features of the plurality of samples into the fully connected layer for mapping processing to obtain a plurality of prediction analysis results.
[0078] S323: training parameters of the fully connected layer according to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results to obtain the cost data analysis model.
[0079] In the example, by training the fully connected layer according to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results output by the fully connected layer, the cost data analysis model can be quickly obtained.
[0080] In an example, the S323 comprises:
[0081] S3231: constructing a result loss function according to historical cost analysis results of the several target cost training samples and corresponding prediction analysis results.
[0082] As an implementation, the result loss function can adopt the following formula:
[0083]
[0084] wherein L is a function result of the result loss function, N is a total number of target cost training samples, y i is a historical cost analysis result of the i-th target cost training sample, F i (c, f(x)) is a prediction analysis result of the i-th target cost training sample, c is a feature extraction layer, f(x) is a full connection layer, and x is a parameter of the full connection layer.
[0085] S3232: adjusting the parameter of the full connection layer by using a gradient descent method to minimize a value of the loss function.
[0086] S3233: obtaining the cost data analysis model according to the feature extraction layer and the full connection layer with the adjusted parameter.
[0087] In the embodiment, the parameter of the full connection layer is adjusted by using the gradient descent method to minimize the value of the loss function, so as to reduce the difference between the prediction analysis result and the historical cost analysis result.
[0088] In a feasible embodiment, the project cost data includes construction project content, construction project building material and construction project cost; and the historical project cost data includes historical project content, historical project building material, historical project cost and historical project cost analysis result.
[0089] The second embodiment of the present application provides a civil engineering cost data analysis system, which comprises:
[0090] A project information acquisition module 10 is configured to acquire a construction project type and project cost data of a target land.
[0091] A training sample screening module 20 is configured to screen a plurality of civil engineering training samples according to a preset cost analysis target and the construction project type, to obtain several target cost training samples of the same construction project type for cost analysis; the target cost training sample comprises historical project cost data and historical cost analysis result.
[0092] The migration learning training module 30 is configured to perform migration learning training on the basis of a migration learning strategy, according to the plurality of target cost training samples and a pre-trained engineering data recognition model, to obtain a cost data analysis model.
[0093] The cost analysis result acquisition module 40 is configured to input the project cost data into the cost data analysis model to obtain a cost analysis result of the target land.
[0094] Compared with the prior art, the civil engineering cost data analysis system can filter target cost training samples for cost analysis of the same construction project type from a plurality of civil engineering training samples according to the construction project type of the target land and a preset cost analysis target, and then perform migration learning training on the basis of migration learning, according to the plurality of target cost training samples and a pre-trained engineering data recognition model, to obtain a cost data analysis model, thereby greatly saving the training time cost of the cost data analysis model. Furthermore, the project cost data is input into the cost data analysis model to obtain a cost analysis result of the target land, thereby reducing the time cost of obtaining the cost analysis result of the target land and improving the efficiency of obtaining the cost analysis result.
[0095] In one possible implementation, the migration learning training module 30 is configured to perform the following steps:
[0096] The full connection layer of the engineering data recognition model is replaced to obtain an initial analysis model;
[0097] The initial analysis model is subjected to migration learning training according to the plurality of target cost training samples to obtain the cost data analysis model.
[0098] In this embodiment, by replacing the full connection layer of the engineering data recognition model and then performing migration learning training according to the target cost training samples, the model parameters that need to be learned and trained can be reduced, thereby saving the training time cost of the cost data analysis model.
[0099] In one possible implementation, the initial analysis model includes a feature extraction layer and a full connection layer; wherein the feature extraction layer is the feature extraction layer of the engineering data recognition model.
[0100] The step of subjecting the initial analysis model to migration learning training according to the plurality of target cost training samples to obtain the cost data analysis model includes:
[0101] The plurality of target cost training samples are input into the feature extraction layer for feature extraction processing to obtain cost features of the plurality of samples.
[0102] The cost features of the plurality of samples are input to the full connection layer for mapping processing to obtain a plurality of prediction analysis results.
[0103] According to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results, the parameters of the full connection layer are trained to obtain the cost data analysis model.
[0104] In this embodiment, the full connection layer is trained according to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results output by the full connection layer, and the cost data analysis model can be quickly obtained.
[0105] In one possible implementation, the step of training the parameters of the full connection layer according to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results to obtain the cost data analysis model comprises:
[0106] A result loss function is constructed according to the historical cost analysis results of the plurality of target cost training samples and the corresponding plurality of prediction analysis results.
[0107] The value of the loss function is minimized by using a gradient descent method to adjust the parameters of the full connection layer.
[0108] The cost data analysis model is obtained according to the feature extraction layer and the full connection layer with the adjusted parameters.
[0109] In this embodiment, the value of the loss function is minimized by using the gradient descent method to adjust the parameters of the full connection layer, and the difference between the prediction analysis results and the historical cost analysis results can be reduced.
[0110] In one possible implementation, the project cost data comprises construction project content, construction project building materials, and construction project cost; and the historical project cost data comprises historical project content, historical project building materials, historical project cost, and historical project cost analysis results.
[0111] The device embodiments described above are merely illustrative, and the components illustrated as separate components can or can not be physically separated, and the components illustrated as units can or can not be physical units, i.e., can be located at one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purposes of the present application according to actual needs. Those skilled in the art can understand and implement without creative labor.
[0112] Those skilled in the art will appreciate that embodiments of the application can be readily used as software, hardware, or a combination of software and hardware. In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.
[0113] The present application is described in reference to the flowchart illustrations and / or block diagrams according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing system, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions of a flow or multiple flows and / or blocks Figure 1 one or more functions of a flow or multiple flows and / or blocks Figure 1 one or more functions of a flow or multiple flows and / or blocks Figure 1 one or more functions of a flow or multiple flows and / or blocks
[0114] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions of a flow or multiple flows and / or blocks Figure 1 one or more functions of a flow or multiple flows and / or blocks
[0115] In one typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.
[0116] The memory can include non-persistent memory and / or volatile memory, such as a random access memory (RAM) including a cache area for the temporary storage of data. The memory can also include non-volatile memory, such as read only memory (ROM) for storing structural information and / or instruction code. Both can be within one or more memory devices 1225. Alternatively, some memory devices can provide a combination of one or more of the foregoing types of memories. Since memory is a computer-readable medium, it can also include a combination of one or more non- volatile memory and / or volatile memory, such as a cache area formed of RAM and a non-volatile storage area formed of ROM. The memory typically includes a cache area for the temporary storage of data, a
[0117] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.
[0118] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusions, so that a process, method, article or apparatus that includes a list of elements does not only include those elements, but also includes other elements not explicitly listed, or further includes elements inherent in such a process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.
[0119] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of claims of the present application.
Claims
1. A method for analyzing civil construction cost data, characterized in that: include: Obtain the construction project type and project cost data for the target land; Filtering a plurality of civil engineering training samples according to a preset cost analysis target and the type of the construction project to obtain a plurality of target cost training samples for cost analysis of the same construction project type; The target cost training samples include historical project cost data and historical cost analysis results; Based on the transfer learning strategy, transfer learning training is performed based on the plurality of target cost training samples and the pre-trained engineering data recognition model to obtain a cost data analysis model; the engineering data recognition model is a deep learning model trained based on the plurality of civil engineering training samples; The project cost data is input into the cost data analysis model to obtain the cost analysis result of the target land.
2. The civil construction cost data analysis method according to claim 1, characterized in that: The step of performing transfer learning training based on the plurality of target cost training samples and the pre-trained engineering data recognition model based on the transfer learning strategy includes: Replacing the fully connected layer of the engineering data recognition model to obtain an initial analysis model; The initial analysis model is subjected to transfer learning training according to the plurality of target cost training samples to obtain the cost data analysis model.
3. The civil construction cost data analysis method according to claim 2, characterized in that: The initial analysis model includes a feature extraction layer and a fully connected layer; wherein the feature extraction is the feature extraction layer of the engineering data recognition model; The step of performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model includes: Inputting the target cost training samples into the feature extraction layer for feature extraction processing to obtain cost features of the samples; Inputting the cost features of the plurality of samples into the fully connected layer for mapping processing to obtain a plurality of prediction analysis results; According to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results, the parameters of the fully connected layer are trained to obtain the cost data analysis model.
4. The civil construction cost data analysis method according to claim 3, characterized in that: The step of training the parameters of the fully connected layer based on the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results to obtain the cost data analysis model includes: Constructing a result loss function based on the historical cost analysis results of the plurality of target cost training samples and the corresponding plurality of prediction analysis results; Minimize the value of the loss function using the gradient descent method and adjust the parameters of the fully connected layer; The cost data analysis model is obtained based on the feature extraction layer and the fully connected layer after parameter adjustment.
5. The civil construction cost data analysis method according to any one of claims 1 to 4, characterized in that: The project cost data includes construction project content, construction project building materials and construction project cost; the historical project cost data includes historical project content, historical project building materials, historical project cost and historical project cost analysis results.
6. A civil construction cost data analysis system, characterized in that: include: Project information acquisition module, used to obtain the construction project type and project cost data of the target land; A training sample screening module is used to screen multiple civil engineering training samples according to a preset cost analysis target and the type of the construction project, and obtain a number of target cost training samples for cost analysis of the same construction project type; the target cost training samples include historical project cost data and historical cost analysis results; A transfer learning training module is configured to perform transfer learning training based on the plurality of target cost training samples and a pre-trained engineering data recognition model based on a transfer learning strategy to obtain a cost data analysis model; the engineering data recognition model is a deep learning model trained based on the plurality of civil engineering training samples; The cost analysis result acquisition module is used to input the project cost data into the cost data analysis model to obtain the cost analysis result of the target land.
7. The civil construction cost data analysis system according to claim 6, characterized in that: The transfer learning training module is used to perform the following steps: Replacing the fully connected layer of the engineering data recognition model to obtain an initial analysis model; The initial analysis model is subjected to transfer learning training according to the plurality of target cost training samples to obtain the cost data analysis model.
8. The civil construction cost data analysis system according to claim 7, characterized in that: The initial analysis model includes a feature extraction layer and a fully connected layer; wherein the feature extraction is the feature extraction layer of the engineering data recognition model; The step of performing transfer learning training on the initial analysis model according to the plurality of target cost training samples to obtain the cost data analysis model includes: Inputting the target cost training samples into the feature extraction layer for feature extraction processing to obtain cost features of the samples; Inputting the cost features of the plurality of samples into the fully connected layer for mapping processing to obtain a plurality of prediction analysis results; According to the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results, the parameters of the fully connected layer are trained to obtain the cost data analysis model.
9. The civil construction cost data analysis system according to claim 8, characterized in that: The step of training the parameters of the fully connected layer based on the historical cost analysis results of the plurality of target cost training samples and the plurality of prediction analysis results to obtain the cost data analysis model includes: Constructing a result loss function based on the historical cost analysis results of the plurality of target cost training samples and the corresponding plurality of prediction analysis results; Minimize the value of the loss function using the gradient descent method and adjust the parameters of the fully connected layer; The cost data analysis model is obtained based on the feature extraction layer and the fully connected layer after parameter adjustment.
10. The civil construction cost data analysis system according to any one of claims 6 to 9, characterized in that: The project cost data includes construction project content, construction project building materials and construction project cost; the historical project cost data includes historical project content, historical project building materials, historical project cost and historical project cost analysis results.