Project cost prediction method and device, electronic equipment and storage medium

Abstract

The invention provides a project cost prediction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining multi-modal data of a target project, processing the multi-modal data, and obtaining a plurality of cost prediction factors of the target project; inputting the plurality of cost prediction factors of the target project into a pre-constructed multi-element collision model to obtain an optimal cost prediction factor of the target project output by the multi-element collision model; inputting the optimal cost prediction factor of the target project into a pre-constructed project cost prediction model to obtain a cost prediction result of the target project output by the project cost prediction model; the multi-element collision model is obtained by training a plurality of sample cost prediction factors and an optimal cost prediction factor label of a sample project; the project cost prediction model is obtained by training based on the optimal sample cost prediction factor and the cost prediction result label of the sample project. According to the invention, the accuracy of project cost prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of project cost prediction, and in particular to a project cost prediction method, device, electronic equipment and storage medium. Background Art

[0002] With the rapid development of communication technology, there are more and more communication network construction projects. In the process of project management, it is usually necessary to predict the project cost to ensure the quality of the project and control the progress and cost of the project. At present, the existing technology only considers a single factor in the process of predicting the project cost, which affects the accuracy of the cost prediction. Summary of the invention

[0003] The present invention provides a method, device, electronic device and storage medium for predicting construction cost, which are used to solve the defect in the prior art that construction cost is predicted based on only a single factor, thus affecting the accuracy of cost prediction.

[0004] In a first aspect, the present invention provides a method for predicting construction cost, comprising: Acquiring multimodal data of a target project, and processing the multimodal data to obtain a plurality of cost prediction factors of the target project; Inputting a plurality of cost prediction factors of the target project into a pre-built multi-factor collision model to obtain an optimal cost prediction factor of the target project output by the multi-factor collision model; Inputting the optimal cost prediction factor of the target project into a pre-built project cost prediction model to obtain a cost prediction result of the target project output by the project cost prediction model; The multi-factor collision model is obtained by training based on multiple sample cost prediction factors of sample projects and the optimal cost prediction factor labels of the sample projects; The project cost prediction model is obtained by training based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

[0005] In some embodiments, the obtaining of multimodal data of the target project includes: Establishing a mapping relationship between the real scene and the virtual scene of the target project, and constructing a digital twin model of the target project based on the mapping relationship; Using an artificial intelligence Internet of Things (AIOT) device to load the digital twin model to form a touchable three-dimensional image; Using the AIOT device to collect the user's operation instructions or voice instructions, and in response to the operation instructions or voice instructions, acquiring multimodal data of the target project in real time; The AIOT device is communicatively connected to a data acquisition sensor of the target project, and the data acquisition sensor is used to collect multimodal data of the target project.

[0006] In some embodiments, the processing of the multimodal data to obtain a plurality of cost prediction factors of the target project includes: Determine whether the multimodal data is structured data; if so, obtain multiple cost prediction factors of the target project based on the multimodal data; otherwise, perform feature recognition and feature extraction on the multimodal data to obtain feature information of the multimodal data; and obtain multiple cost prediction factors of the target project based on the feature information of the multimodal data; The multiple cost prediction factors of the target project are stored in a database.

[0007] In some embodiments, after obtaining the cost prediction result of the target project output by the project cost prediction model, the method further includes: Constructing a building information model (BIM) of the target project, decoupling the BIM, and obtaining a plurality of components of the target project; Associating the plurality of components with the project cost prediction model; Presenting a cost prediction result of the target project using the multiple components; Obtaining an emergency plan for the target project according to a cost prediction result of the target project, and presenting the emergency plan for the target project using the multiple components; The emergency plan of the target project is sent to the user end.

[0008] In some embodiments, the cost prediction result of the target project includes a cost prediction value of each component; The step of obtaining an emergency plan for the target project based on the cost forecast result of the target project includes: Compare the cost prediction value of each component with the cost planning value of each component to determine whether there is a deviation in the cost prediction value of each component. If so, determine the emergency plan for each component based on the pre-constructed knowledge graph; The construction process of the knowledge graph includes: Obtain historical emergency plans of historical projects, perform natural language processing on the historical emergency plans, and extract key information of the historical emergency plans; The knowledge graph is constructed based on the key information of the historical emergency plans.

[0009] In some embodiments, the inputting of the plurality of cost prediction factors of the target project into a pre-built multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model includes: Based on the multi-factor collision model, the cost impact weights of the multiple cost prediction factors of the target project are calculated; based on the cost impact weights of the multiple cost prediction factors, the multiple cost prediction factors are screened to obtain the screened multiple cost prediction factors; the screened multiple cost prediction factors are sorted to determine the optimal cost prediction factor of the target project.

[0010] In some embodiments, the training process of the multi-factor collision model includes: Acquiring sample multimodal data of the sample project, processing the sample multimodal data, and obtaining a plurality of sample cost prediction factors of the sample project; Determining the optimal cost prediction factor label for the sample project; Taking multiple sample cost prediction factors of the sample project as training samples and the optimal cost prediction factor label of the sample project as sample label, an initial multi-factor collision model is trained, and the parameters of the initial multi-factor collision model are iteratively optimized to obtain the multi-factor collision model.

[0011] In some embodiments, the training process of the construction cost prediction model includes: Acquiring sample multimodal data of the sample project, processing the sample multimodal data, and obtaining a plurality of sample cost prediction factors of the sample project; Inputting a plurality of sample cost prediction factors of the sample project into the multi-factor collision model to obtain the optimal sample cost prediction factor of the sample project output by the multi-factor collision model; Determine a cost prediction result label of the sample project; Inputting the optimal sample cost prediction factor of the sample project into a pre-built initial project cost prediction model to obtain a cost prediction result of the sample project output by the initial project cost prediction model; Calculating a loss function value based on the cost prediction result of the sample project and the cost prediction result label of the sample project; Based on the loss function value, the parameters of the initial project cost prediction model are iteratively optimized to obtain the project cost prediction model.

[0012] In a second aspect, the present invention further provides a construction cost prediction device, comprising: An acquisition unit, used for acquiring multimodal data of a target project, processing the multimodal data, and obtaining a plurality of cost prediction factors of the target project; An optimization unit, used for inputting a plurality of cost prediction factors of the target project into a pre-built multi-factor collision model to obtain an optimal cost prediction factor of the target project output by the multi-factor collision model; A prediction unit, used to input the optimal cost prediction factor of the target project into a pre-built project cost prediction model to obtain a cost prediction result of the target project output by the project cost prediction model; The multi-factor collision model is obtained by training based on multiple sample cost prediction factors of sample projects and the optimal cost prediction factor labels of the sample projects; The project cost prediction model is obtained by training based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

[0013] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting engineering cost as described above is implemented.

[0014] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for predicting construction costs.

[0015] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned methods for predicting construction costs.

[0016] The project cost prediction method, device, electronic device and storage medium provided by the present invention obtain multimodal data of a target project, process the multimodal data, obtain multiple cost prediction factors of the target project, use a multi-factor collision model to determine the optimal cost prediction factor from the multiple cost prediction factors, adopt a project cost prediction model, and obtain a cost prediction result of the target project based on the optimal cost prediction factor, thereby improving the accuracy of project cost prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is an architecture diagram of a construction cost prediction system provided by an embodiment of the present invention; Figure 2 It is a flowchart of a method for predicting construction cost provided by an embodiment of the present invention; Figure 3 is a flow chart of a training process of a multi-factor collision model provided by an embodiment of the present invention; Figure 4 It is a flowchart of the training process of the engineering cost prediction model provided by an embodiment of the present invention; Figure 5 It is a structural schematic diagram of a project cost prediction device provided by an embodiment of the present invention; Figure 6 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0020] The terms "first", "second", etc. in the present invention are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present invention can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more.

[0021] Figure 1 This is a diagram of the architecture of the project cost prediction system provided by the embodiment of the present invention. Figure 1 As shown, a construction cost prediction system is provided, which includes a data layer, a capability layer, an application layer and a user layer; the data layer is used to obtain multimodal data of the target project, such as table data, audio and video data and document data; the capability layer includes a construction cost prediction model, a multi-factor collision model, smart devices and a digital twin model; the application layer is used to realize smart device data collection, accurate construction cost prediction, emergency plan push and presentation based on Building Information Modeling (BIM); the user layer is used to obtain users' data collection needs, cost prediction needs, solution needs and visualization presentation needs.

[0022] Figure 2 The following is a flow chart of a method for predicting construction cost provided by an embodiment of the present invention. Figure 2 As shown, a method for predicting construction cost is provided, which can be implemented based on the above system, and includes the following steps: step 210, step 220, and step 230. The steps of the method flow are only a possible implementation of the present invention.

[0023] Step 210: Acquire multimodal data of the target project, process the multimodal data, and obtain multiple cost prediction factors of the target project.

[0024] Optionally, the multimodal data includes at least table data, audio and video data, and document data.

[0025] Optionally, the multimodal data is processed by data cleaning, feature recognition, feature extraction, etc.

[0026] Optionally, the multiple cost prediction factors of the target project include at least: project type, project scale, material price, labor cost, equipment cost, and project period.

[0027] In some embodiments, obtaining multimodal data of the target project in step 210 includes: Step 211: Establish a mapping relationship between the real scene and the virtual scene of the target project, and build a digital twin model of the target project based on the mapping relationship; Step 212: Use an artificial intelligence Internet of Things (AIOT) device to load the digital twin model to form a touchable three-dimensional image; Step 213: using the AIOT device to collect the user's operation instructions or voice instructions, and in response to the operation instructions or voice instructions, obtaining multimodal data of the target project in real time; Among them, the AIOT device is communicated with the data acquisition sensor of the target project, and the data acquisition sensor is used to collect multimodal data of the target project.

[0028] Among them, AIOT devices combine artificial intelligence (AI) and Internet of Things (IoT) technologies, aiming to process data collected by IoT devices in real time through intelligent algorithms, improve the intelligence and automation level of IoT devices, and thus achieve autonomous decision-making and optimized management.

[0029] Optionally, the AIOT device includes a projection device, a high-precision sensor and a data processing device, and the AIOT device is a wearable smart device.

[0030] Among them, the projection device adopts advanced air projection technology, which forms a stable image in the air through a high-speed rotating mirror or a special optical system, and can accurately present the digital twin model. The AIOT device has an adjustment function that can adjust the projection size, angle, and brightness, and can adapt to different scenes and environments.

[0031] Among them, high-precision sensors can locate and capture user actions in real time. When the user clicks on the digital twin model, the sensor can quickly obtain the specific location of the user's operation and pass the operation information to the data unit. The sensing device of the AIOT device has the characteristics of high sensitivity and fast response, and can capture user operation information in real time and accurately.

[0032] The data processing device can support the import of model files, including but not limited to RVT and 3DS file formats. The data processing device is used to receive the interactive information transmitted by the sensing device and obtain the attribute data of the corresponding component in the model according to the interactive information. The data processing device has a built-in model attribute database, which stores the attribute data of each component in the digital twin model. When receiving the user's operation information, the data processing device retrieves the corresponding component attribute data from the database according to the operation position and presents it to the user through the projection device.

[0033] Optionally, the AIOT device has voice recognition function, data storage and transmission function.

[0034] In some embodiments, the multimodal data is processed to obtain multiple cost prediction factors of the target project, including: Determine whether the multimodal data is structured data. If so, obtain multiple cost prediction factors of the target project based on the multimodal data. Otherwise, perform feature recognition and feature extraction on the multimodal data to obtain feature information of the multimodal data. Based on the feature information of the multimodal data, obtain multiple cost prediction factors of the target project. Store multiple cost prediction factors of the target project into the database.

[0035] Optionally, a pre-established neural network model is used to perform feature recognition and feature extraction on unstructured multimodal data.

[0036] Optionally, establish AIOT device data docking function.

[0037] It should be noted that AIOT equipment data docking mainly establishes data relationships with the IoT platform based on the equipment model and code at the project site, uses artificial intelligence technology to complete data processing, and saves the final results to the application database to provide a data source for subsequent use. The algorithm is as follows: “begin

[0038] {

[0039] 1)platformEntity = CreateAiotPlatform(ProjName,ProjNo)

[0040] 2)data = platformEntity.GetProjData()

[0041] 3)if(data.type == StructuredData)

[0042] 4)SaveData(data)

[0043] 5)else

[0044] 6){

[0045] 7)LoadTrainData(trainData)

[0046] 8)model = CNNModel()

[0047] 9)model.add_convolution_layer

[0048] 10)model.add_pooling_layer

[0049] 11)model.add_fully_connected_layer

[0050] 12)model.complieModel

[0051] 13)result =model.train(trainData,parameter)

[0052] 14)if(result>95%)

[0053] 15)processedresult=model.recognition(data)

[0054] 16)SaveData(processedresult)

[0055] 17)esle

[0056] 18)model.train(trainData,parameter)

[0057] 19)}

[0058] }

[0059] end".

[0060] Among them, CreateAiotPlatform means to establish an IoT platform according to the project name and number. The IoT platform is used to collect data and process artificial intelligence recognition results; platformEntity.GetProjData means to create a data transmission channel. Data.type is used to determine whether the multimodal data is structured data. If it is structured data, it can be directly stored in the database. Otherwise, a neural network model is established to perform feature recognition and extraction on unstructured data. The historical data of the model is loaded through LoadTrainData, and the initial neural network model is established based on CNNModel. The basic parameters of the network model are set through model.add. complieModel means to compile the model, model.train means to train the model with historical data, model.recognition(data) means to recognize historical data, and processedresult means the standardized result returned after recognition. When the credibility of the training result is greater than 95%, the model is considered to be usable. If the credibility of the training result is less than 95%, it is necessary to increase the training data and adjust the parameters to retrain the model.

[0061] Step 220: Input multiple cost prediction factors of the target project into a pre-built multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model.

[0062] Among them, the multi-factor collision model is trained based on multiple sample cost prediction factors of sample projects and the optimal cost prediction factor labels of sample projects.

[0063] Optionally, a data transmission channel between the AIOT device and the multi-factor collision model is established. The algorithm is as follows: “begin

[0064] {

[0065] 1) pipe = CreatePipeLine(deviceNo,projectNo)

[0066] 2) projectCostData = pipe.GetProjectData()

[0067] 3) ClassificationData(projectCostData,type)

[0068] 4) Validate(projectCostData)

[0069] 5)Save(projectCosatData)

[0070] }

[0071] end".

[0072] Among them, CreatePipeLine means to establish a data transmission channel between smart devices and multi-factor collision models. GetProjectData means to extract data from the channel; ClassificationData means to classify data and classify and manage structured data and unstructured data; Validate means to preliminarily verify the validity of data and preliminarily clean up data through multiple dimensions such as project name, factor judgment, and engineering data; Save means to store data related to engineering cost, laying the foundation for subsequent cost forecasting and data asset accumulation.

[0073] In some embodiments, multiple cost prediction factors of the target project are input into a pre-built multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model, including: Based on the multi-factor collision model, the cost impact weights of multiple cost prediction factors of the target project are calculated, and based on the cost impact weights of the multiple cost prediction factors, the multiple cost prediction factors are screened to obtain the screened multiple cost prediction factors, and the screened multiple cost prediction factors are sorted to determine the optimal cost prediction factor of the target project.

[0074] Optionally, a multi-factor collision algorithm is used to optimize the cost prediction factor of the target project. The specific algorithm is as follows: “begin

[0075] {

[0076] 1) mdata=GetDataFromDataBase(url)

[0077] 2) datacollection = DivideFactor(mdata)

[0078] 3) datacollection = {y1 y2 y3 .....yn}

[0079] 4) compareCollection = [x1 x2 x3 .....x_l ]

[0080] 5) relevance = R * len

[0081] 6)if(Judge(MAX(relevance_col>Col_all))

[0082] {

[0083] 7)advantageColFactor =relevance_col

[0084] }

[0085] 8)if(Judege(MAX(relevance_row>ROW_all)))

[0086] {

[0087] 9)advantageRowFactor =relevance_row

[0088] }

[0089] 10)outdata = MuliFactorMode(datacollection,comparecollection,

[0090] advantageColFactor,advantageRowFactor )

[0091] }

[0092] end”。

[0093] Among them, GetDataFromDataBase means to retrieve the engineering cost collision factor from the database; DivideFactor means to divide the engineering cost factor into N sets, each set has a multi-factor collision factor, and the N sets are collided by loop traversal to screen out the main influencing factors, and finally the main influencing factors that meet the target project are introduced; datacollection means the current collision factor, compareCollection means the comparison factor; relevance=R*len means the weight set of the current collision factor, where R represents the correlation between the factors, and len represents the length of the set; advantageColFactor =relevance_col means that when the length of a column in the collection is greater than that of other sequences, advantageColFactor is set as the advantage factor. advantageRowFactor =relevance_row means that when the elements of a row in the collection are greater than the elements of other rows, advantageRowFactor is set as the excellent parent factor; MuliFactorMode means that the collision factor, comparison factor, advantage factor and parent factor are input into the multi-factor collision model, and the multi-factor collision model introduces the factor with the largest weight of engineering cost impact.

[0094] Step 230: Input the optimal cost prediction factor of the target project into the pre-built project cost prediction model to obtain the cost prediction result of the target project output by the project cost prediction model.

[0095] Among them, the engineering cost prediction model is trained based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

[0096] It can be understood that by using the optimal cost prediction factors of the target project, the project cost prediction model can better capture the key factors affecting the project cost, thereby improving the accuracy of the prediction.

[0097] In some embodiments, after obtaining the cost prediction result of the target project output by the project cost prediction model, the method further includes: Build a Building Information Model (BIM) of the target project, decouple the BIM, and obtain multiple components of the target project; Associating multiple components with the project cost prediction model; Use multiple components to present the cost forecast results of the target project; According to the cost prediction result of the target project, the emergency plan of the target project is obtained, and the emergency plan of the target project is presented by using multiple components; Send the emergency plan of the target project to the user end.

[0098] Among them, BIM not only shows the appearance and structure of the building, but also integrates important data such as materials, costs, schedules, energy efficiency, etc.

[0099] It is understandable that by decoupling BIM, we can obtain multiple components of the target project, and by associating the multiple components with the engineering cost prediction model, we can more accurately estimate the cost of each component and reduce the error of the overall budget. By using multiple components for visual display, it can help relevant personnel understand the project cost and emergency plan more intuitively.

[0100] In some embodiments, the cost prediction result of the target project includes a cost prediction value of each component; According to the cost forecast results of the target project, the emergency plan of the target project is obtained, including: Compare the cost forecast value of each component with the cost planning value of each component to determine whether there is a deviation in the cost forecast value of each component. If so, determine the emergency plan for each component based on the pre-built knowledge graph; The construction process of the knowledge graph includes: Obtain historical emergency plans of historical projects, perform natural language processing on the historical emergency plans, and extract key information of the historical emergency plans; Build a knowledge graph based on the key information of historical emergency plans.

[0101] Optionally, based on the engineering cost prediction model, the cost prediction value of each component of BIM is obtained, and based on the cost prediction value of each component, a corresponding emergency plan is obtained. The algorithm involved is as follows: “begin

[0102] {

[0103] 1) result = GetPredictionResult(predictionModel)

[0104] 2) m_BimModel = DecouplingComponent(bimModel)

[0105] 3) Relevance(m_BimModel,result)

[0106] 4) if(result <planValue )

[0107] {

[0108] 5)SetBimColor(Green)

[0109] 6)SendCostPlanToManager(reason)

[0110] }else if (result = planValue)

[0111] {

[0112] 7)SetBimColor(Blue)

[0113] 8)SendCostPlanToManager(reason)

[0114] }

[0115] else

[0116] {

[0117] 9)SetBimColor(Red)

[0118] 10)SendCostPlanToManager(reason)

[0119] }

[0120] 11)SummaryCostReoprt()

[0121] }

[0122] end”。

[0123] GetPredictionResult indicates that the cost prediction model of the project is optimized by multi-factor collision to complete the cost prediction of each component in the BIM model; DecouplingComponent indicates that BIM is decoupled to make each component of BIM an independent individual; Relevance indicates associating the component with the project cost prediction model; result indicates the predicted value, and planValue indicates the planned value. Compare the predicted value with the planned value. When the predicted value is less than the planned value, the BIM component is set to green through SetBimColor, indicating that there is a surplus of expenses according to the current cost expenditure. Pay attention to the control of quality, progress, safety and other aspects of each link of the project, and send the control plan to the user through SendCostPlanToManager; when the predicted value is greater than the planned value, the BIM component is red, indicating that the current predicted value is greater than the planned value, reminding the management personnel that there is a possibility of cost overruns, and push overrun plans to users to control costs and improve overall work efficiency; when the predicted value is equal to the planned value, the color of the BIM component is blue, reminding the user that the current project is progressing smoothly and the project can be successfully completed according to the current situation; SummaryCostReoprt indicates that an overall analysis of the project is conducted based on the plan push situation and cost analysis situation to form a project cost report.

[0124] In an embodiment of the present invention, multimodal data of a target project is acquired and processed to obtain a plurality of cost prediction factors of the target project. An optimal cost prediction factor is determined from the plurality of cost prediction factors using a multi-factor collision model. An engineering cost prediction model is used to obtain a cost prediction result of the target project based on the optimal cost prediction factor, thereby improving the accuracy of engineering cost prediction.

[0125] Figure 3 FIG. 1 is a flow chart of the training process of the multi-factor collision model provided by an embodiment of the present invention. Figure 3 As shown, in some embodiments, the process of determining the multi-element collision model includes: Step 310: Acquire sample multimodal data of the sample project, process the sample multimodal data, and obtain multiple sample cost prediction factors of the sample project; Step 320, determining the optimal cost prediction factor label of the sample project; Step 330: Using multiple sample cost prediction factors of sample projects as training samples and the optimal cost prediction factor labels of sample projects as sample labels, an initial multi-factor collision model is trained, and the parameters of the initial multi-factor collision model are iteratively optimized to obtain a multi-factor collision model.

[0126] The sample multimodal data includes at least image, text, voice and other data.

[0127] Optionally, based on the initial multi-factor collision model, multiple sample cost prediction factors of the sample project are collided, the collision results are output in the form of weights, and finally the primary cost prediction factor of the sample project is output according to the weight sorting.

[0128] Figure 4 The following is a flow chart of the training process of the project cost prediction model provided by the embodiment of the present invention. Figure 4 As shown, in some embodiments, the training process of the project cost prediction model includes: Step 410: Acquire sample multimodal data of the sample project, process the sample multimodal data, and obtain multiple sample cost prediction factors of the sample project; Step 420: input multiple sample cost prediction factors of the sample project into the multi-factor collision model to obtain the optimal sample cost prediction factor of the sample project output by the multi-factor collision model; Step 430, determining the cost prediction result label of the sample project; Step 440: input the optimal sample cost prediction factor of the sample project into the pre-built initial project cost prediction model to obtain the cost prediction result of the sample project output by the initial project cost prediction model; Step 450: Calculate the loss function value based on the cost prediction result of the sample project and the cost prediction result label of the sample project; Step 460: Based on the loss function value, iteratively optimize the parameters of the initial project cost prediction model to obtain the project cost prediction model.

[0129] Optionally, the sample multimodal data is processed by data cleaning, feature recognition, feature extraction, etc. to obtain multiple sample cost prediction factors of the sample project.

[0130] It can be understood that by using the multi-factor collision model to obtain the optimal sample cost prediction factor of the sample project, training samples can be obtained quickly and accurately, which helps to improve the training efficiency of the engineering cost prediction model. The optimal sample cost prediction factor of the sample project is used as the training sample, and the cost prediction result label of the sample project is used as the sample label to train the initial engineering cost prediction model, which improves the accuracy and robustness of the prediction of the engineering cost prediction model.

[0131] The following is a description of a construction cost prediction device provided in an embodiment of the present invention. The construction cost prediction device described below and the construction cost prediction method described above can be referenced to each other.

[0132] Figure 5 A schematic diagram of the structure of a project cost prediction device provided by an embodiment of the present invention is shown in FIG. Figure 5 As shown, the construction cost prediction device 500 includes: An acquisition unit 510 is used to acquire multimodal data of a target project, process the multimodal data, and obtain multiple cost prediction factors of the target project; An optimization unit 520 is used to input multiple cost prediction factors of the target project into a pre-built multi-factor collision model to obtain an optimal cost prediction factor of the target project output by the multi-factor collision model; The prediction unit 530 is used to input the optimal cost prediction factor of the target project into a pre-built project cost prediction model to obtain a cost prediction result of the target project output by the project cost prediction model; Among them, the multi-factor collision model is trained based on multiple sample cost prediction factors of sample projects and the optimal cost prediction factor labels of sample projects; Among them, the engineering cost prediction model is trained based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

[0133] Optionally, obtain multimodal data of the target project, including: Establish a mapping relationship between the real scene and the virtual scene of the target project, and build a digital twin model of the target project based on the mapping relationship; Use AIOT devices to load digital twin models to form touchable three-dimensional images; Use AIOT devices to collect user operation instructions or voice instructions, respond to the operation instructions or voice instructions, and obtain multimodal data of the target project in real time; Among them, the AIOT device is communicated with the data acquisition sensor of the target project, and the data acquisition sensor is used to collect multimodal data of the target project.

[0134] Optionally, the multimodal data is processed to obtain multiple cost prediction factors of the target project, including: Determine whether the multimodal data is structured data. If so, obtain multiple cost prediction factors of the target project based on the multimodal data. Otherwise, perform feature recognition and feature extraction on the multimodal data to obtain feature information of the multimodal data. Based on the feature information of the multimodal data, obtain multiple cost prediction factors of the target project. Store multiple cost prediction factors of the target project into the database.

[0135] Optionally, the construction cost prediction device 500 further includes: A construction unit is used to construct a building information model (BIM) of a target project, decouple the BIM, and obtain multiple components of the target project; An association unit, used to associate multiple components with the project cost prediction model; A first display unit, used to present the cost prediction result of the target project using multiple components; A second display unit is used to obtain an emergency plan for the target project based on the cost prediction result of the target project, and present the emergency plan for the target project using multiple components; The sending unit is used to send the emergency plan of the target project to the user end.

[0136] Optionally, the cost prediction result of the target project includes a cost prediction value of each component; According to the cost forecast results of the target project, the emergency plan of the target project is obtained, including: Compare the cost forecast value of each component with the cost planning value of each component to determine whether there is a deviation in the cost forecast value of each component. If so, determine the emergency plan for each component based on the pre-built knowledge graph; The construction process of the knowledge graph includes: Obtain historical emergency plans of historical projects, perform natural language processing on the historical emergency plans, and extract key information of the historical emergency plans; Build a knowledge graph based on the key information of historical emergency plans.

[0137] Optionally, multiple cost prediction factors of the target project are input into a pre-built multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model, including: Based on the multi-factor collision model, the cost impact weights of multiple cost prediction factors of the target project are calculated, and based on the cost impact weights of the multiple cost prediction factors, the multiple cost prediction factors are screened to obtain the screened multiple cost prediction factors, and the screened multiple cost prediction factors are sorted to determine the optimal cost prediction factor of the target project.

[0138] Optionally, the training process of the multi-factor collision model includes: Acquire sample multimodal data of the sample project, process the sample multimodal data, and obtain multiple sample cost prediction factors of the sample project; Determine the optimal cost prediction factor labels for sample projects; Taking multiple sample cost prediction factors of sample projects as training samples and the optimal cost prediction factor labels of sample projects as sample labels, an initial multi-factor collision model is trained, and the parameters of the initial multi-factor collision model are iteratively optimized to obtain a multi-factor collision model.

[0139] Optionally, the training process of the project cost prediction model includes: Acquire sample multimodal data of the sample project, process the sample multimodal data, and obtain multiple sample cost prediction factors of the sample project; Inputting multiple sample cost prediction factors of the sample project into the multi-factor collision model to obtain the optimal sample cost prediction factor of the sample project output by the multi-factor collision model; Determine the cost forecast result label of the sample project; The optimal sample cost prediction factor of the sample project is input into the pre-built initial project cost prediction model to obtain the cost prediction result of the sample project output by the initial project cost prediction model; Based on the cost prediction results of the sample project and the cost prediction result labels of the sample project, the loss function value is calculated; Based on the loss function value, the parameters of the initial engineering cost prediction model are iteratively optimized to obtain the engineering cost prediction model.

[0140] It should be noted here that the engineering cost prediction device provided in the embodiment of the present invention can implement all the method steps implemented in the above-mentioned engineering cost prediction method embodiment, and can achieve the same technical effect. The parts and beneficial effects of this embodiment that are the same as the method embodiment will not be described in detail here.

[0141] Figure 6 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 6 As shown, the electronic device may include: a processor (processor) 610 , a communication interface (Communications Interface) 620 , a memory (memory) 630 and a communication bus 640 , wherein the processor 610 , the communication interface 620 , and the memory 630 communicate with each other through the communication bus 640 . The processor 610 can call the logic instructions in the memory 630 to execute the project cost prediction method, which includes: obtaining multimodal data of the target project, processing the multimodal data, and obtaining multiple cost prediction factors of the target project; inputting the multiple cost prediction factors of the target project into a pre-built multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model; inputting the optimal cost prediction factor of the target project into a pre-built project cost prediction model to obtain the cost prediction result of the target project output by the project cost prediction model; wherein the multi-factor collision model is trained based on multiple sample cost prediction factors of the sample project and the optimal cost prediction factor label of the sample project; wherein the project cost prediction model is trained based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

[0142] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0143] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the engineering cost prediction method provided by the above-mentioned methods, which includes: acquiring multimodal data of a target project, processing the multimodal data, and obtaining multiple cost prediction factors of the target project; inputting the multiple cost prediction factors of the target project into a pre-constructed multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model; inputting the optimal cost prediction factor of the target project into a pre-constructed engineering cost prediction model to obtain the cost prediction result of the target project output by the engineering cost prediction model; wherein the multi-factor collision model is obtained by training based on multiple sample cost prediction factors of a sample project and the optimal cost prediction factor label of the sample project; wherein the engineering cost prediction model is obtained by training based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

[0144] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the engineering cost prediction method provided by the above-mentioned methods, the method comprising: acquiring multimodal data of a target project, processing the multimodal data, and obtaining a plurality of cost prediction factors of the target project; inputting the plurality of cost prediction factors of the target project into a pre-constructed multi-factor collision model, and obtaining an optimal cost prediction factor of the target project output by the multi-factor collision model; inputting the optimal cost prediction factor of the target project into a pre-constructed engineering cost prediction model, and obtaining a cost prediction result of the target project output by the engineering cost prediction model; wherein the multi-factor collision model is trained based on a plurality of sample cost prediction factors of a sample project, and an optimal cost prediction factor label of a sample project; wherein the engineering cost prediction model is trained based on an optimal sample cost prediction factor of a sample project, and a cost prediction result label of a sample project.

[0145] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0146] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting construction cost, characterized in that: include: Acquiring multimodal data of a target project, and processing the multimodal data to obtain a plurality of cost prediction factors of the target project; Inputting a plurality of cost prediction factors of the target project into a pre-built multi-factor collision model to obtain an optimal cost prediction factor of the target project output by the multi-factor collision model; Inputting the optimal cost prediction factor of the target project into a pre-built project cost prediction model to obtain a cost prediction result of the target project output by the project cost prediction model; The multi-factor collision model is obtained by training based on multiple sample cost prediction factors of sample projects and the optimal cost prediction factor labels of the sample projects; The project cost prediction model is obtained by training based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

2. The method for predicting construction cost according to claim 1, characterized in that: The obtaining of multimodal data of the target project includes: Establishing a mapping relationship between the real scene and the virtual scene of the target project, and constructing a digital twin model of the target project based on the mapping relationship; Using an artificial intelligence Internet of Things (AIOT) device to load the digital twin model to form a touchable three-dimensional image; Using the AIOT device to collect the user's operation instructions or voice instructions, and in response to the operation instructions or voice instructions, acquiring multimodal data of the target project in real time; The AIOT device is communicatively connected to a data acquisition sensor of the target project, and the data acquisition sensor is used to collect multimodal data of the target project.

3. The method for predicting construction cost according to claim 1, characterized in that: The multimodal data is processed to obtain a plurality of cost prediction factors of the target project, including: Determine whether the multimodal data is structured data; if so, obtain multiple cost prediction factors of the target project based on the multimodal data; otherwise, perform feature recognition and feature extraction on the multimodal data to obtain feature information of the multimodal data; and obtain multiple cost prediction factors of the target project based on the feature information of the multimodal data; The multiple cost prediction factors of the target project are stored in a database.

4. The method for predicting construction cost according to claim 1, characterized in that: After obtaining the cost prediction result of the target project output by the project cost prediction model, the method further includes: Constructing a building information model (BIM) of the target project, decoupling the BIM, and obtaining a plurality of components of the target project; Associating the plurality of components with the project cost prediction model; Presenting a cost prediction result of the target project using the multiple components; Obtaining an emergency plan for the target project according to a cost prediction result of the target project, and presenting the emergency plan for the target project using the multiple components; The emergency plan of the target project is sent to the user end.

5. The method for predicting construction cost according to claim 4, characterized in that: The cost forecast result of the target project includes the cost forecast value of each component; The step of obtaining an emergency plan for the target project based on the cost forecast result of the target project includes: Compare the cost prediction value of each component with the cost planning value of each component to determine whether there is a deviation in the cost prediction value of each component. If so, determine the emergency plan for each component based on the pre-constructed knowledge graph; The construction process of the knowledge graph includes: Obtain historical emergency plans of historical projects, perform natural language processing on the historical emergency plans, and extract key information of the historical emergency plans; The knowledge graph is constructed based on the key information of the historical emergency plans.

6. The method for predicting construction cost according to claim 1, characterized in that: The step of inputting the multiple cost prediction factors of the target project into a pre-built multi-factor collision model to obtain the optimal cost prediction factor of the target project output by the multi-factor collision model includes: Based on the multi-factor collision model, the cost impact weights of the multiple cost prediction factors of the target project are calculated; based on the cost impact weights of the multiple cost prediction factors, the multiple cost prediction factors are screened to obtain the screened multiple cost prediction factors; the screened multiple cost prediction factors are sorted to determine the optimal cost prediction factor of the target project.

7. The method for predicting construction cost according to claim 1, characterized in that: The training process of the multi-factor collision model includes: Acquiring sample multimodal data of the sample project, processing the sample multimodal data, and obtaining a plurality of sample cost prediction factors of the sample project; Determining the optimal cost prediction factor label for the sample project; Taking multiple sample cost prediction factors of the sample project as training samples and the optimal cost prediction factor label of the sample project as sample label, an initial multi-factor collision model is trained, and the parameters of the initial multi-factor collision model are iteratively optimized to obtain the multi-factor collision model.

8. The method for predicting construction cost according to claim 1, characterized in that: The training process of the engineering cost prediction model includes: Acquiring sample multimodal data of the sample project, processing the sample multimodal data, and obtaining a plurality of sample cost prediction factors of the sample project; Inputting a plurality of sample cost prediction factors of the sample project into the multi-factor collision model to obtain the optimal sample cost prediction factor of the sample project output by the multi-factor collision model; Determine a cost prediction result label of the sample project; Inputting the optimal sample cost prediction factor of the sample project into a pre-built initial project cost prediction model to obtain a cost prediction result of the sample project output by the initial project cost prediction model; Calculating a loss function value based on the cost prediction result of the sample project and the cost prediction result label of the sample project; Based on the loss function value, the parameters of the initial project cost prediction model are iteratively optimized to obtain the project cost prediction model.

9. A construction cost prediction device, characterized in that: include: An acquisition unit, used to acquire multimodal data of a target project, process the multimodal data, and obtain a plurality of cost prediction factors of the target project; An optimization unit, used for inputting a plurality of cost prediction factors of the target project into a pre-built multi-factor collision model to obtain an optimal cost prediction factor of the target project output by the multi-factor collision model; A prediction unit, used to input the optimal cost prediction factor of the target project into a pre-built project cost prediction model to obtain a cost prediction result of the target project output by the project cost prediction model; The multi-factor collision model is obtained by training based on multiple sample cost prediction factors of sample projects and the optimal cost prediction factor labels of the sample projects; The project cost prediction model is obtained by training based on the optimal sample cost prediction factor of the sample project and the cost prediction result label of the sample project.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the construction cost prediction method according to any one of claims 1 to 8 is implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the construction cost prediction method according to any one of claims 1 to 8 is implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the construction cost prediction method according to any one of claims 1 to 8 is implemented.

Images