Construction task priority determination method and system and storage medium

Abstract

Embodiments of the invention provide a construction task priority determination method and system, and a storage medium. The method comprises the steps of obtaining task data of a plurality of construction tasks; and inputting the task data into the priority prediction model to obtain task priorities of the plurality of construction tasks. The priority prediction model is obtained through first training and second training based on historical task data. The first data set is used for training in the first training, and the second data set is used for training in the second training. The fluctuation of the second data set relative to the first data set is greater than a preset threshold. According to the method, the priority of the construction task can be quickly and accurately determined, comprehensive and accurate identification and timely processing of the high-priority task are ensured, the progress and efficiency of the whole construction project are ensured, and meanwhile, the resource allocation efficiency is improved.

Description

Technical Field

[0001] This specification relates to the field of engineering construction management, and particularly to a method, system, and storage medium for determining the priority of construction tasks. Background Art

[0002] In engineering construction management, the priority of construction tasks directly affects construction efficiency and progress. Currently, the priority of construction tasks is usually determined manually by construction managers (e.g., team leaders), which takes a lot of time and reduces the accuracy of the determined construction task priority due to human errors and other reasons. In addition, during actual construction, the construction site often changes rapidly, and construction tasks are uncertain. When a certain task among multiple construction tasks needs to be adjusted, construction managers need to comprehensively re-determine the priority of construction tasks, lacking flexibility and adaptability.

[0003] Therefore, it is desirable to provide a method, system, and storage medium for determining the priority of construction tasks to improve the efficiency, accuracy, and adaptability of determining the priority of construction tasks. Summary of the Invention

[0004] One embodiment of this specification provides a method for determining the priority of construction tasks. The method includes: obtaining task data of multiple construction tasks; inputting the task data into a priority prediction model to obtain the task priorities of the multiple construction tasks, where the priority prediction model is obtained through a first training and a second training based on historical task data, the first training is performed using a first data set, the second training is performed using a second data set, and the fluctuation of the second data set relative to the first data set is greater than a preset threshold.

[0005] In some embodiments, the priority prediction model includes multiple sub-prediction models. The task data of the multiple construction tasks can be respectively input into each sub-prediction model of the multiple sub-prediction models to obtain the priority scores of the multiple construction tasks output by each sub-prediction model; for each construction task among the multiple construction tasks, the priority scores of the construction task output by each sub-prediction model are weighted and combined to obtain the task priority score of the construction task; based on the task priority scores of all construction tasks among the multiple construction tasks, the priority ranking of the multiple construction tasks is obtained.

[0006] In some embodiments, in the weighted combination, the weight corresponding to each sub-prediction model can be related to the sum of the priority scores output by the sub-prediction model processing the first data set.

[0007] In some embodiments, the priority score of each construction task among the multiple construction tasks can be calculated based on the accuracy rate and recall rate of the construction task.

[0008] In some embodiments, the second data set satisfies a preset constraint condition, and the preset constraint condition is related to the amount of the construction task and the deadline.

[0009] In some embodiments, the training process of the priority prediction model may include: inputting the first data set into an initial prediction model for training to obtain an intermediate prediction model; dividing the first data set into at least two subsets, and iteratively training the intermediate prediction model based on the at least two subsets to obtain the accuracy rate and recall rate of at least some of the at least two subsets; calculating the standard deviation of the first data set based on the accuracy rate and recall rate of at least some of the at least two subsets; inputting a third data set that satisfies the preset constraint condition into the intermediate prediction model to obtain a prediction result of the third data set; screening out the second data set from the third data set based on the prediction result of the third data set and the standard deviation of the first data set; and retraining the intermediate prediction model with the second data set to obtain the trained priority prediction model.

[0010] In some embodiments, in each iteration of the iterative training, one of the at least two subsets may be used as a test set, and the remaining subsets may be used as a training set, and the intermediate prediction model is trained based on the training set to obtain the accuracy rate and recall rate corresponding to the test set.

[0011] In some embodiments, the prediction result of the third data set includes the accuracy rate and recall rate of the third data set. The fluctuation score of the third data set relative to the first data set may be calculated based on the accuracy rate and recall rate of the third data set and the standard deviation of the first data set; and the set of construction task data in the third data set whose fluctuation score meets the preset threshold requirement is used as the second data set.

[0012] One embodiment of this specification provides a construction task priority determination system, including a data acquisition module, a priority prediction module, and a model training module. The data acquisition module is configured to acquire task data of multiple construction tasks; the priority prediction module is configured to input the task data into a priority prediction model to obtain the task priorities of the multiple construction tasks; and the model training module is configured to obtain the priority prediction model through a first training and a second training based on historical task data. The first training is performed using a first data set, and the second training is performed using a second data set, and the fluctuation of the second data set relative to the first data set is greater than a preset threshold.

[0013] One embodiment of this specification provides a device for determining the priority of construction tasks, including a processor, which is used to execute the method for determining the priority of construction tasks.

[0014] One embodiment of this specification provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the method for determining the priority of construction tasks.

[0015] In some embodiments of this specification, by using a priority prediction model trained multiple times based on historical task data to determine the task priorities of multiple construction tasks, it is possible to quickly and accurately determine the priorities of construction tasks, ensure the comprehensive and accurate identification and timely processing of high-priority tasks, guarantee the progress and efficiency of the entire construction project, and the accurate division of construction task priorities also improves the allocation efficiency of resources; when adjusting construction tasks, the model can also quickly determine the new priorities of construction tasks, improving the flexibility and adaptability in determining the priorities of construction tasks and enhancing the quality of construction task management; by weighted-combining the output results of multiple sub-models, the advantages of multiple models are integrated, improving the overall performance of the priority prediction model, enhancing the accuracy, comprehensiveness, and stability of the entire priority prediction model, thereby further improving the accuracy of determining the priorities of construction tasks; through the priority prediction model, intelligent and dynamic task management can be realized, making the overall operation of the enterprise more efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] This specification will be further described by way of exemplary embodiments, which will be described in detail through the accompanying drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where:

[0017] Figure 1 is a schematic diagram of the application scenario of an exemplary construction task priority system shown according to some embodiments of this specification;

[0018] Figure 2 is a block diagram of an exemplary construction task priority system shown according to some embodiments of this specification;

[0019] Figure 3 is a flowchart of an exemplary construction task priority method shown according to some embodiments of this specification;

[0020] Figure 4 is a flowchart of an exemplary process of obtaining the construction task priority using a priority prediction model shown according to some embodiments of this specification;

[0021] Figure 5 is a schematic diagram of an exemplary process of obtaining the construction task priority using a priority prediction model shown according to some embodiments of this specification;

[0022] Figure 6 is a flowchart for training an exemplary priority prediction model shown in some embodiments of this specification. Detailed implementation manners

[0023] To more clearly illustrate the technical solutions of the embodiments of this specification, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the figures represent the same structure or operation.

[0024] It should be understood that the "system", "device", "unit" and / or "module" used herein is a way to distinguish different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the said words can be replaced by other expressions.

[0025] As shown in this specification and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0026] Flowcharts are used in this specification to illustrate the operations performed by the systems according to the embodiments of this specification. It should be understood that the operations before or after do not necessarily need to be executed precisely in sequence. On the contrary, they can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several operations can be removed from these processes.

[0027] Figure 1 is a schematic diagram of the application scenario of a construction task priority determination system shown in some embodiments of this specification. As Figure 1 shown, in some embodiments, the construction task priority determination system 100 may include a first computing device 110, a second computing device 120, a user terminal 130, a storage device 140, and a network 150.

[0028] The first computing device 110 and the second computing device 120 are systems with computing and processing capabilities, which may include various computers, such as servers and personal computers, or may also be computing platforms composed of multiple computers connected in various structures. In some embodiments, the first computing device 110 and the second computing device 120 may be the same device or different devices.

[0029] One or more sub-processing devices (e.g., single-core processing devices or multi-core and multi-chip processing devices) may be included in the first computing device 110 and the second computing device 120, and the processing devices may execute program instructions. By way of example only, the processing devices may include various common general-purpose central processing units (CPUs), graphics processing units (GPUs), microprocessors, application-specific integrated circuits (ASICs), or other types of integrated circuits.

[0030] The first computing device 110 may process information and data. In some embodiments, the first computing device 110 may execute the construction task priority determination method as shown in some embodiments of this specification to obtain the task priorities of multiple construction tasks, for example, task priority scores, sorted results of priorities, etc. In some embodiments, the first computing device 110 may include a machine learning model, and the first computing device 110 may obtain the priority scores of construction tasks through the machine learning model. In some embodiments, the first computing device 110 may obtain a trained machine learning model from the second computing device 120. In some embodiments, the first computing device 110 may exchange information and data through the network 150 and / or other components in the system 100 (e.g., medical imaging device 110, second computing device 120, user terminal 130, storage device 140). In some embodiments, the first computing device 110 may be directly connected to the second computing device 120 and exchange information and / or data.

[0031] The second computing device 120 may be used for model training. In some embodiments, the second computing device 120 may execute the training method of the priority prediction model as shown in some embodiments of this specification to obtain a trained machine learning model. In some embodiments, the second computing device 120 may train a machine learning model based on historical task data. In some embodiments, the first computing device 110 and the second computing device 120 may also be the same computing device.

[0032] The user terminal 130 can be used to interact with a user (e.g., a construction management personnel such as a team leader). In some embodiments, the user terminal 130 can receive the task priority determination results of multiple construction tasks from the first computing device 110 and guide the construction plan based on the determination results. In some embodiments, the user terminal 130 can instruct the first computing device 110 to execute the construction task priority determination method as shown in some embodiments of this specification. In some embodiments, the user terminal 130 can be one or any combination of a mobile device 130-1, a tablet computer 130-2, a laptop computer 130-3, a desktop computer, and other devices with input and / or output functions.

[0033] The storage device 140 can store data or information generated by other devices. In some embodiments, the storage device 140 can store the data and / or information processed by the first computing device 110 and / or the second computing device 120, such as a machine learning model, the priority score of a construction task, the priority ranking of a construction task, etc. The storage device 140 can include one or more storage components, and each storage component can be an independent device or a part of other devices. The storage device can be local or implemented through the cloud.

[0034] The network 150 can connect the components of the system and / or connect the system to the external resource part. The network 150 enables communication between the components and between the system and other parts outside the system, facilitating the exchange of data and / or information. In some embodiments, one or more components in the system 100 (e.g., the first computing device 110, the second computing device 120, the user terminal 130, the storage device 140) can send data and / or information to other components through the network 150. In some embodiments, the network 150 can be any one or more of a wired network or a wireless network.

[0035] It should be noted that the above description is provided for illustrative purposes only and is not intended to limit the scope of this specification. For those of ordinary skill in the art, various changes and modifications can be made under the guidance of the content of this specification. The features, structures, methods, and other features of the exemplary embodiments described in this specification can be combined in various ways to obtain additional and / or alternative exemplary embodiments. For example, the first computing device 110 and / or the second computing device 120 can be based on a cloud computing platform, such as a public cloud, a private cloud, a community cloud, and a hybrid cloud, etc. However, these changes and modifications will not deviate from the scope of this specification.

[0036] Figure 2 is a block diagram of an exemplary construction task priority system shown in some embodiments of this specification. As Figure 2As shown, the construction task priority system 200 includes a data acquisition module 210 and a priority prediction module 220. In some embodiments, the data acquisition module 210 and the priority prediction module 220 may be implemented by the first computing device 110.

[0037] The data acquisition module 210 is configured to acquire task data of multiple construction tasks. For more details on how to acquire task data, refer to step S310.

[0038] The priority prediction module 220 is configured to input the task data into a priority prediction model to obtain the task priorities of multiple construction tasks. For more details on how to obtain the task priorities, refer to step S320 Figure 4 and Figure 5 。

[0039] In some embodiments, the construction task priority system 200 may further include a model training module 230, which may be implemented by the second computing device 120. The model training module 230 is configured to obtain a priority prediction model based on historical task data through a first training and a second training. The first training is performed using a first data set, and the second training is performed using a second data set. The fluctuation of the second data set relative to the first data set is greater than a preset threshold. For more details on how to train the priority prediction model, refer to Figure 6 。

[0040] Figure 3 is a flowchart of an exemplary construction task priority method according to some embodiments of the present specification. As Figure 3 shown, the process 300 includes the following steps. In some embodiments, the process 300 may be executed by the first computing device 110.

[0041] Step S310, acquiring task data of multiple construction tasks. In some embodiments, step S310 may be executed by the data acquisition module 210.

[0042] Task data is data related to construction tasks. For example, task description, task urgency, task importance, task amount, time to task deadline, task type, etc. The task description includes basic information of the task, such as task content, etc. The task urgency is the description and classification of the task urgency, such as not urgent, moderately urgent, urgent, etc. The task importance is the description and classification of the task importance, such as unimportant, moderately important, important, very important, etc. The amount is the budget amount of the task. The time to task deadline is the time from the current time to the task deadline. The task type is the classification of the task, such as sudden task, daily task, urgent task, task arranged by superiors, etc. The first computing device 110 can obtain the task data of multiple construction tasks in the engineering project in various ways. For example, obtaining from the construction plan, etc. This specification does not limit the specific way of obtaining task data.

[0043] Step S320, input the task data into the priority prediction model to obtain the task priorities of multiple construction tasks. In some embodiments, step S320 can be executed by the priority prediction module 220.

[0044] The task priority is information indicating the priority of the task. The task priority can include various forms, such as task priority score, task priority level, task priority ranking, etc. The task priority score is a numerical value representing the priority of a certain construction task. For example, the larger the task priority score of a certain construction task, the higher the priority of the task. The task priority level is the priority level of the construction task. For example, the task priority level can include 5 levels from 1 to 5. The larger the number, the higher the task priority and the more important the task. The task priority ranking is the ranking of the priorities of multiple construction tasks, which can be obtained based on the task priority scores of each task among these construction tasks. For example, sort multiple construction tasks in descending order of the task priority scores, and the obtained sorting result represents the ranking of the priorities of these construction tasks from high to low. The first computing device 110 can obtain the task priorities of multiple construction tasks through the priority prediction model.

[0045] In some embodiments, the priority prediction model includes various machine learning models. For example, decision tree, random forest, AdaBoost, neural network model, etc. The input of the priority prediction model includes the task data of multiple construction tasks, and the output includes at least one of the task priority scores, task priority levels, task priority rankings, etc. of these construction tasks. In some embodiments, the output of the priority prediction model can include the priority ranking of these construction tasks.

[0046] In some embodiments, the priority prediction model includes multiple sub - prediction models, and these sub - prediction models are not the same. For example, the priority prediction model may include three sub - prediction models, and the types of these three sub - models are decision tree, random forest, and AdaBoost respectively. The first computing device 110 may input the task data of multiple construction tasks into each of the multiple sub - prediction models of the priority prediction model respectively, and obtain the priority scores of the multiple construction tasks output by each sub - prediction model. For each construction task among the multiple construction tasks, the first computing device 110 may perform a weighted combination of the priority scores of the construction task output by each sub - prediction model to obtain the task priority score of this construction task. The first computing device 110 may obtain the priority ranking of the multiple construction tasks based on the task priority scores of the construction tasks among the multiple construction tasks. For more content on how to obtain the task priorities of multiple construction tasks through the multiple sub - prediction models of the priority prediction model, reference can be made to Figure 4 .

[0047] The historical task data is the construction task data in the completed engineering projects, including priority markers, task descriptions, task urgency, task importance, task amounts, task deadlines, task types, etc. The priority marker is the identifier of the priority of the construction task. For example, the priority markers may include P0, P1, P2, and from P0 to P2, the priority gradually decreases.

[0048] In some embodiments, the priority prediction model is obtained through a first training and a second training based on the historical task data. Among them, the first training is performed using the first data set, the second training is performed using the second data set, and the fluctuation of the second data set relative to the first data set is greater than a preset threshold.

[0049] The first data set is obtained from historical task data. For example, the first data set can be part or all of the historical task data. The second data set is obtained based on the first data set. The second data set is a subset of the first data set, or the second data set and the first data set are partially or completely non-overlapping. For example, at least part of the second data set can be obtained from the first data set. In some embodiments, the second data set meets a preset constraint condition, which is related to the contract amount and deadline of the construction task. For example, the preset constraint condition includes one of the following: the contract amount is greater than the first contract threshold (e.g., 5000000, 10000000, etc.), the contract amount is less than the second contract threshold (e.g., 1000, 2000, etc.), the distance from the deadline is less than the first time threshold (e.g., 1 day, 2 days, etc.), the distance from the deadline is greater than the second time threshold (e.g., 30 days, 40 days, etc.), etc. By using the contract amount and deadline of the construction task to filter the data and obtain the data set for the second training of the priority prediction model, the model performance can be further improved, the performance of the model in specific scenarios or extreme scenarios can be improved, the interference of special scenarios can be excluded, and the model can perform better when processing data in real scenarios.

[0050] In some embodiments, the second computing device 120 can obtain a trained priority prediction model by performing the following steps: inputting the first data set into the initial prediction model for training to obtain an intermediate prediction model; dividing the first data set into at least two subsets (each subset includes at least one task data), and iteratively training the intermediate prediction model based on the at least two subsets to obtain the accuracy rate and recall rate of each of the at least two subsets; calculating the standard deviation of the first data set based on the accuracy rates and recall rates of all subsets in the at least two subsets; inputting the third data set that meets the preset constraint condition into the intermediate prediction model to obtain the prediction result of the third data set; screening out the second data set from the third data set based on the prediction result of the third data set and the standard deviation of the first data set; and using the second data set to retrain the intermediate prediction model to obtain a trained priority prediction model. For more content on how to train the priority prediction model, see Figure 5 。

[0051] In some embodiments of this specification, by using a machine learning model to determine the task priorities of multiple construction tasks, the task priorities of construction tasks can be determined quickly and accurately, ensuring the comprehensive and accurate identification and timely processing of high-priority tasks, guaranteeing the progress and efficiency of the entire construction project, and the accurate division of construction task priorities also improves the allocation efficiency of resources; when the construction tasks are adjusted, the new construction task priorities can also be quickly determined through the model, improving the flexibility and adaptability of determining construction task priorities and improving the quality of construction task management.

[0052] Figure 4It is a flowchart for obtaining the construction task priority according to the exemplary usage priority prediction model shown in some embodiments of this specification. As Figure 4 shown, process 400 includes the following steps. In some embodiments, the first computing device 110 or the priority prediction module 220 may implement step S320 by executing at least part of process 400, that is, inputting task data into the priority prediction model to obtain the task priorities of multiple construction tasks.

[0053] Step S410, input the task data of multiple construction tasks into each of the multiple sub-prediction models respectively to obtain the priority scores of multiple construction tasks output by each sub-prediction model.

[0054] Figure 5 It is a schematic diagram for obtaining the construction task priority according to the exemplary usage priority prediction model shown in some embodiments of this specification. As Figure 5 shown, the construction task data 510 includes the task data of m (m>1) construction tasks (task 1, task 2,..., task m), that is, task data 1, task data 2,..., task data m; the priority prediction model 520 includes n (n>1) sub-models, that is, sub-model 1, sub-model 2,..., sub-model n. The first computing device 110 inputs the construction task data 510 into each sub-model, thereby obtaining n sets of priority scores corresponding to the n sub-models, that is, priority score sets 530-1, 530-2,..., 530-n. The priority score set output by each sub-model includes the priority scores corresponding to task 1, task 2,..., task m respectively, that is, priority score 1, priority score 2,..., priority score m. It can be understood that since each sub-model is different, the priority scores corresponding to the same construction task output by different sub-models may be different. For example, both sub-model 1 and sub-model 2 output the priority score 1 corresponding to task 1, but these two priority scores 1 may be the same or different.

[0055] Step S420, for each construction task among the multiple construction tasks, perform weighted combination on the priority scores of the construction task output by each sub-prediction model to obtain the task priority score of the construction task.

[0056] Different types of sub-prediction models have their own advantages. For example, decision tree models are highly interpretable. In construction task management, a decision tree can clearly show the decision-making process of task priorities, helping managers understand why certain tasks are given priority. This is crucial for team communication and decision-making transparency. Decision tree models can handle complex decisions. Construction tasks usually involve multiple factors (e.g., task urgency, resource availability, dependencies, etc.), and decision trees can effectively handle these complex decision logics through their branch structures. Decision tree models can respond quickly. The construction and prediction speed of decision trees is relatively fast, making them suitable for construction sites where task priorities need to be adjusted in real time. Another example is that random forest models have high accuracy. In the prediction of construction task priorities, random forests can improve the prediction accuracy by integrating multiple decision trees, reducing the bias that may be brought by a single model. Random forests have the ability to resist overfitting. Construction task data may contain noise and uncertainties. The randomness and integration characteristics of random forests make them perform well in the face of complex and variable data, reducing the risk of overfitting. Random forests can conduct feature importance assessment. Random forests can evaluate the impact of each feature (such as task type, resource requirements, time limits, etc.) on task priorities, helping managers optimize resource allocation and task arrangements. Another example is that the AdaBoost model can improve the performance of weak classifiers. In the scenario of construction task priorities, there may be some tasks that are difficult to classify (e.g., tasks that are urgent but lack resources). The AdaBoost model can improve the recognition ability of these complex tasks by weighted combination of multiple weak classifiers. The AdaBoost model has strong adaptability. The AdaBoost model can dynamically adjust the weights of classifiers according to the actual situation of tasks, which is very important for the rapid changes and uncertainties on construction sites. The AdaBoost model can handle imbalanced data. In construction tasks, some tasks may be more urgent or important than others. The AdaBoost model can effectively handle this imbalance to ensure that important tasks are given priority.

[0057] In some embodiments, for each construction task among multiple construction tasks, the first computing device 110 may obtain the task priority score of the construction task based on the priority scores of the construction task output by each sub-prediction model. The first computing device 110 may calculate the weighted sum of the priority scores of the construction task output by all sub-prediction models according to the weights corresponding to each sub-prediction model, and use this weighted sum as the task priority score of the construction task. By combining the priority scores of the construction task output by each sub-model to obtain the task priority score of the construction task, the advantages of different types of sub-models can be aggregated, making the obtained construction task priority more accurate, comprehensive, and more adaptable to various construction scenarios.

[0058] Such as Figure 5As shown, the first computing device 110 performs a weighted combination of the priority scores in the set of priority scores 530-1, 530-2, ……, 530-n, that is, calculates the weighted sum of the priority scores corresponding to each task in these sets of priority scores to obtain the task priority score corresponding to the task. The task priority scores of all tasks form the set of task priority scores 540. Taking task 1 as an example, the first computing device 110 extracts the priority score 1 from each of the sets of priority scores 530-1, 530-2, ……, 530-n, and then calculates the weighted sum of these priority scores 1, and takes this weighted sum as the task priority score 1 corresponding to task 1. The task priority score 1 represents the priority of task 1.

[0059] In some embodiments, when performing a weighted combination of the priority scores, the weight corresponding to each sub-prediction model is related to the score of the sub-prediction model in processing the first data set. The first computing device 110 can obtain the score of the sub-prediction model according to the result of the sub-prediction model in processing the first data set, and this score is used to characterize the performance of the model in processing the first data set. It can be understood that for different data sets, the performance of the model may be different. Therefore, when comparing the performance of different models in processing data, it is compared on the basis of the same data set. By associating the weight corresponding to the sub-prediction model with the score of the sub-prediction model in processing the data set, the weight of the model can accurately reflect the performance of the model in processing data, so that when obtaining the priority score by calculating the weighted sum, the different importance levels of different sub-models can be reflected, and the obtained priority score is more accurate and comprehensive.

[0060] For example only, assume that the priority prediction model includes three sub-prediction models: Model M, Model N, and Model O. The scores of Model M, Model N, and Model O for processing dataset A (e.g., the first dataset) are MscoreA, NscoreA, and OscoreA respectively, and the total score of the priority prediction model for processing the first dataset is MscoreA + NscoreA + OscoreA. Then the weight corresponding to Model M is MscoreA / (MscoreA + NscoreA + OscoreA), the weight corresponding to Model N is NscoreA / (MscoreA + NscoreA + OscoreA), and the weight corresponding to Model O is OscoreA / (MscoreA + NscoreA + OscoreA). Assume that for a certain construction task, the priority scores of this construction task output by Model M, Model N, and Model O are Mscore, Nscore, and Oscore respectively. Then the task priority score of this construction task can be expressed as Mscore · MscoreA / (MscoreA + NscoreA + OscoreA) + Nscore · NscoreA / (MscoreA + NscoreA + OscoreA) + Oscore · OscoreA / (MscoreA + NscoreA + OscoreA).

[0061] In some embodiments, the accuracy of the model prediction result can be measured by precision and recall. Precision is the ratio of the number of correctly predicted samples output by the model to the total number of input samples. In this specification, a sample refers to a construction task, a correct prediction means that the task priority output by the model is accurate, and an incorrect prediction means that the task priority output by the model is inaccurate; precision is the ratio of the number of correctly predicted construction tasks output by the priority prediction (sub) model to the total number of input construction tasks. In some embodiments, precision can be as shown in formula (1) below: Accuracy=(TP+TN) / (TP+TN+FP+FN) (1), Among them, Accuracy represents the accuracy rate; TP represents the number of samples correctly predicted as the positive class; TN represents the number of samples correctly predicted as the negative class; FP represents the number of negative class samples wrongly predicted as the positive class; FN represents the number of positive class samples wrongly predicted as the negative class. The positive class and the negative class refer to the types of output samples. For example, for an input construction task, if the model outputs the priority level of the construction task, and the levels are from 1 to 6, the larger the number, the higher the priority. Then the positive class can include the priority of the output construction task being 4 - 6, and the negative class can include the priority of the output construction task being 1 - 3. Assume that the actual priority of construction task 1 is level 5, and the actual priority of construction task 2 is level 2. Then correctly predicting as the positive class can be that the priority of the output construction task 1 is 5, and correctly predicting as the negative class can be that the priority of the output construction task 2 is 2; wrongly predicting as the positive class of the negative class can be any one of the priorities of the output construction task 2 being 4 - 6, and wrongly predicting as the negative class of the positive class can be any one of the priorities of the output construction task 1 being 1 - 3.

[0062] In some embodiments, the accuracy rate of the model can be obtained during the model training process. When calculating the accuracy rate of the model, the second computing device 120 can use a dataset divided into different test sets and training sets, train the model through multiple iterations, and calculate the average accuracy rate after multiple iterations as the accuracy rate of the finally obtained trained model. For example, the second computing device 120 can randomly divide the first dataset into 10 subsets Z1 - Z10; perform 10 iterations, each time selecting one subset as the test set and the remaining 9 subsets as the training set; in each iteration, use the training set to train the model; use the trained model to make predictions on the test set and record the prediction results; obtain the accuracy rate of the model after each training according to the prediction results; calculate the average accuracy rate after multiple iterations. In some embodiments, the average accuracy rate can be as shown in the following formula (2): Among them, α represents the average accuracy rate; Accuracy i represents the accuracy rate of the trained model after the i - th iteration; n is the number of training iterations.

[0063] The recall rate is the ratio of the number of correctly predicted positive class samples to the number of actual positive class samples. In some embodiments, the recall rate can be as shown in the following formula (3): Recall = TP / (TP + FN) (3), Among them, Recall represents the recall rate; the meanings of TP and FN are the same as those in formula (1). It can be understood that the number of actual positive class samples is equal to the sum of the number of samples correctly predicted as the positive class TP and the number of positive class samples wrongly predicted as the negative class FN.

[0064] In some embodiments, similar to the accuracy rate, the second computing device 120 may calculate the average recall rate after multiple iterations as the recall rate of the finally obtained trained model. In some embodiments, the average recall rate may be as shown in formula (4) below: where β represents the average recall rate; Recall i represents the recall rate of the trained model after the i-th iteration; n is the number of training iterations.

[0065] In some embodiments, the score of the sub-prediction model for processing the first data set is calculated based on the accuracy rate and recall rate of the sub-prediction model for processing the first data set. The higher the score, the more accurate the prediction. Since the accuracy of the model prediction result can be measured by the accuracy rate and recall rate, by representing the prediction accuracy as a score, a comprehensive and accurate quantitative evaluation of the prediction accuracy can be made using one metric, making the prediction accuracy more intuitive and precise.

[0066] For construction operations, the recall rate is usually prioritized over the accuracy rate. In construction management, high-priority tasks are usually closely related to the progress, safety, and cost control of the project. If the model fails to identify some high-priority tasks (i.e., positive class samples mispredicted as negative classes), it may cause delays in these tasks, thus affecting the progress and efficiency of the entire construction project. The situation at the construction site may change rapidly, and timely identification and handling of high-priority tasks are crucial for ensuring the smooth progress of construction. Therefore, ensuring that as many high-priority tasks as possible are identified (increasing the recall rate) is the key. In the construction scenario, some tasks may involve safety hazards, and missing these tasks may lead to serious consequences. Therefore, it is very important to ensure that all potential high-priority tasks are identified. By increasing the recall rate, managers can better allocate resources, ensure that high-priority tasks are processed in a timely manner, and thus optimize construction efficiency and resource utilization.

[0067] However, the accuracy rate is still important. A high accuracy rate can reduce incorrect priority judgments, avoid resource waste and unnecessary interference. Therefore, in the early stage of model application (when the recall rate is low), the recall rate is extremely important, while the accuracy rate is not that important and is only used to eliminate interference (for example, screening out low-priority tasks).

[0068] In some embodiments, taking the model M as an example, the score of a single model for processing a certain data set can be represented by formula (5) as follows: MscoreX = δ·M Accuracy +(1 - δ)·M Recall (5), where MscoreX represents the score of the model M for processing a certain data set X; M AccuracyDenotes the accuracy score of model M, obtained based on the accuracy of model M; M Recall Denotes the recall score of model M, obtained based on the recall of model M; δ denotes the weight of the accuracy score; 1 - δ denotes the weight of the recall score. In some embodiments, the accuracy score of model M may be equal to the accuracy of model M, and the recall score of model M may be equal to the recall of model M. In some embodiments, the accuracy of model M may be the average accuracy, and the recall of model M may be the average recall.

[0069] Step S430, obtaining the priority ranking of multiple construction tasks based on the task priority scores of all construction tasks among the multiple construction tasks.

[0070] The first computing device 110 sorts the task priority scores of all construction tasks among the multiple construction tasks in a preset order (for example, from largest to smallest score, from smallest to largest score), to obtain the priority ranking of these construction tasks (for example, from highest to lowest priority, from lowest to highest priority). As Figure 5 shown, the first computing device 110 sorts the task priority scores 1, task priority score 1,..., task priority score m in the task priority score set 540 corresponding to task 1, task 2,..., task m respectively. For example, it can be sorted in descending order to obtain the task priority ranking 550, where the task priorities are arranged in descending order.

[0071] In some embodiments, after obtaining the priority ranking of multiple construction tasks, the first computing device 110 may display this ranking to the user (such as the team leader) through a visual interface or the like. In some embodiments, after the user receives a new construction task, the first computing device 110 may obtain a new priority ranking based on the new construction task data and display it to the user.

[0072] In some embodiments of this specification, by performing weighted combination on the priority scores of multiple sub - prediction models in the priority prediction model, thereby obtaining the task priority scores of construction tasks and finally obtaining the priority ranking of multiple construction tasks, it is possible to combine the different advantages of multiple models, improve the overall performance of the priority prediction model, make the prediction more stable, and the priority prediction result more accurate and comprehensive.

[0073] Figure 6 Is a flowchart of training an exemplary priority prediction model shown in some embodiments of this specification. As Figure 6 shown, process 600 includes the following steps. In some embodiments, the second computing device 120 or the model training module 230 may train the priority prediction model in Figure 3 and Figure 4 by executing at least a part of process 600.

[0074] Step S610: Input the first data set into the initial prediction model for training to obtain an intermediate prediction model.

[0075] The initial prediction model is an original priority prediction model that has not been trained. The second computing device 120 inputs at least part of the first data set as training samples into the initial prediction model for training to obtain an intermediate prediction model. For example, when the model is a decision tree model, the initial prediction model may only include a root node, and the intermediate prediction model is a tree structure, where the leaf nodes represent task priorities.

[0076] In some embodiments, the second computing device 120 may divide the first data set into a training set and a test set, where the training set is larger than the test set. For example, 70% of the first data set is used as the training set and 30% as the test set. When dividing the training set and the test set, the second computing device 120 may divide based on preset features (such as information gain, Gini index, etc.). The second computing device 120 may divide the first data set into different subsets according to different preset features (different preset features correspond to different training sets and test sets), use the training set in the divided subsets to train the initial prediction model, and the steps of dividing the subsets and training may be performed recursively until a preset stop condition is met, and the initial prediction model at this time is used as the intermediate prediction model. For example, at least one of the model reaching a preset depth, the number of node samples being less than a sample number threshold, etc. In some embodiments, the second computing device 120 may use the test set to test the trained initial prediction model. The test may be performed after each training or only on the intermediate prediction model.

[0077] After obtaining the intermediate prediction model, the second computing device 120 may perform performance evaluation and optimization on the intermediate prediction model by performing the operations of steps S620 - S660, so as to obtain a trained priority prediction model.

[0078] Step S620: Divide the first data set into at least two subsets, and perform iterative training on the intermediate prediction model based on the at least two subsets to obtain the accuracy rate and recall rate of at least part of the at least two subsets.

[0079] In some embodiments, in each iteration of iterative training, the second computing device 120 uses one of at least two subsets as the test set and the remaining subsets as the training set, and trains the intermediate prediction model based on the training set to obtain the accuracy rate and recall rate corresponding to the test set. Specifically, the second computing device 120 may randomly divide the first data set into multiple subsets and perform multiple iterative trainings on the intermediate prediction model. In each iteration, one of the subsets is used as the test set and the other subsets are used as the training set, that is, the number of training sets is greater than the number of test sets.

[0080] By way of example only, the second computing device 120 may divide the first data set into 10 subsets Z1 to Z10 and perform 10 iterations. In each iteration, one subset is selected as the test set and the remaining 9 subsets are used as the training set. First iteration: Select the first subset Z1 as the test set and the rest as the training set. Second iteration: Select the second subset Z2 as the test set and the rest as the training set. And so on until the 10th iteration: Select the 10th subset Z10 as the test set and the rest as the training set.

[0081] In each iteration, the second computing device 120 uses the training set to train the model and uses the trained model to make predictions on the test set, that is, uses the trained model to process the test set and records the prediction results, that is, the task priorities of the construction tasks in the test set. The second computing device 120 may obtain the accuracy rate and recall rate of the model corresponding to the test set after this training based on the prediction results of each iteration, that is, the accuracy rate and recall rate of the model processing the test set. In some embodiments, the second computing device 120 may obtain the average accuracy rate and average recall rate based on the accuracy rates and recall rates of the models after multiple iterative trainings.

[0082] It can be understood that the accuracy rate and recall rate of the model are for specific subsets. The second computing device 120 may calculate the accuracy rates and recall rates corresponding to all subsets in the first data set, or only calculate the accuracy rates and recall rates corresponding to some of the subsets. For example, only calculate the accuracy rate and recall rate corresponding to the subsets in the test set. The accuracy rate and recall rate corresponding to all subsets in the test set are the accuracy rate and recall rate of the model processing the test set. The accuracy rate and recall rate of the model processing the first data set may be equal to the accuracy rate and recall rate of the model processing the test set respectively. For more information about the accuracy rate and recall rate of the model, see the relevant content of step S430.

[0083] Step S630, calculate the standard deviation of the first data set based on the accuracy rates and recall rates of at least some of the at least two subsets.

[0084] The standard deviation is a statistic that describes the degree to which the observed values in a dataset deviate from the mean. It is the square root of the variance and can provide information about the dispersion (i.e., volatility) of the data. The standard deviation can be used to represent the fluctuations in accuracy and recall. In some embodiments, the second computing device 120 can calculate the standard deviation of the first dataset based on the accuracy and recall of the test set in the first dataset. The standard deviation of the samples in the dataset can be calculated using the following formula (6): where s represents the standard deviation of the dataset; n represents the number of samples in the dataset; x i represents the data corresponding to the i-th sample, for example, accuracy, recall, etc.; represents the mean of the sample data. Assuming that the dataset A represents the first dataset, the standard deviation of the accuracy of the dataset A can be represented by s αA and the standard deviation of the recall of the dataset A can be represented by s βA where, since the accuracy and recall are relative to a single test set, the samples in formula (6) can be the test sets in the first dataset.

[0085] Step S640: Input the third dataset that meets the preset constraint conditions into the intermediate prediction model to obtain the prediction result of the third dataset.

[0086] To ensure that the prediction result of the model is more in line with the actual situation, it is necessary to evaluate the performance of the model in specific scenarios or extreme scenarios, identify the challenges that the model may encounter in actual applications, and optimize the training for scenarios where the accuracy and recall change too much based on this. In some embodiments, the specific scenarios that conform to the current construction business can be represented by preset constraint conditions, and the preset constraint conditions are related to the target amount and deadline of the construction task. The second computing device 120 can use the third dataset to evaluate and optimize the training of the model.

[0087] The second computing device 120 can obtain the third dataset based on the first dataset. In some embodiments, the second computing device 120 can use the construction task data in the first dataset that meets the preset constraint conditions as the data in the third dataset. In some embodiments, when there is no data in the first dataset that meets the preset constraint conditions, or the number of data that meets the preset constraint conditions is less than the preset number, the second computing device 120 generates the data in the third dataset according to the simulated business scenario, so that the number of data in the third dataset is greater than or equal to the preset number. The preset number is the number of data that meets the model optimization requirements and can be determined based on experience or through the historical data of model training.

[0088] Step S650: Screen out the second data set from the third data set based on the prediction result of the third data set and the standard deviation of the first data set.

[0089] The second computing device 120 can obtain the prediction result of the third data set. The prediction result of the third data set includes the task priority of the construction task obtained by the intermediate prediction model processing the third data set. In some embodiments, the prediction result of the third data set may further include the accuracy rate and recall rate of the third data set, that is, the accuracy rate and recall rate of the intermediate prediction model processing the third data set. The second computing device 120 can calculate the accuracy rate and recall rate of the intermediate prediction model processing the third data set based on the task priority of the construction task obtained by the intermediate prediction model processing the third data set. The calculation method is similar to that of the accuracy rate and recall rate of the first data set and will not be elaborated here.

[0090] The second computing device 120 can calculate the fluctuation score of the third data set relative to the first data set (for example, the fluctuation score of the accuracy rate, recall rate, etc. of the third data set relative to the first data set) based on the accuracy rate and recall rate of the third data set and the standard deviation of the first data set, which can be denoted as Z-score. In some embodiments, assuming there are two data sets A and B, the fluctuation score Z of data set B relative to data set A can be shown as the following formula (7): Where, X represents the value of a certain index d (for example, accuracy rate, recall rate) of the data in data set B; μ represents the average value of index d in data set A (for example, average accuracy rate, average recall rate); σ represents the standard deviation of index d in data set A (for example, accuracy rate standard deviation, recall rate standard deviation).

[0091] According to formula (7), for the accuracy rate, the fluctuation score Z of data set B relative to data set A α is shown as the following formula (8): Where, α B represents the accuracy rate of data set B in a certain iteration. As mentioned above, the accuracy rate of data set B can be equal to the accuracy rate of some subsets in data set B; α A represents the average accuracy rate of data set A after multiple iterations; s αA represents the standard deviation of the accuracy rate of data set A after multiple iterations.

[0092] According to formula (7), for the recall rate, the fluctuation score Z of data set B relative to data set A β is shown as the following formula (9): Where, β Bdenotes the recall rate of dataset B in a certain iteration. As mentioned before, the recall rate of dataset B can be equal to the recall rate of some subsets in dataset B; β A denotes the average recall rate of dataset A after multiple iterations; s βA denotes the standard deviation of the recall rate of dataset A after multiple iterations.

[0093] The second processing device 120 can use the set of construction task data in the third dataset whose fluctuation scores meet the preset threshold requirements as the second dataset. Specifically, the second processing device 120 can use the construction task data in all subsets of the third dataset whose fluctuation scores meet the preset threshold requirements as the data of the second dataset. The fluctuation score can be set according to experience, the precision requirements of the data, etc., and this specification does not limit it. In some embodiments, the preset threshold can be a numerical range. For example, the fluctuation threshold can be [-2, 2]. The preset threshold requirement can be that the fluctuation score is within or outside the preset threshold range. In some embodiments, the second processing device 120 can use all subsets in the first dataset whose fluctuation scores are outside the preset threshold range as the second dataset. For example, in formula (8), assume α B is the accuracy rate of the intermediate prediction model processing the test set Z 1 (subset of the third dataset) in the first iteration, and the fluctuation score Z α is not within the preset threshold range, then the second processing device 120 uses the construction task data in the test set Z 1 as the data in the second dataset.

[0094] Step S660, use the second dataset to retrain the intermediate prediction model to obtain a trained priority prediction model.

[0095] The second processing device 120 can use the construction task data in the second dataset as training samples to retrain the intermediate prediction model. Among them, the training can be performed iteratively until the fluctuation scores of the accuracy rate and recall rate of the second dataset relative to the first dataset are both within the preset threshold range. The second processing device 120 can use the intermediate prediction model at this time as the trained priority prediction model.

[0096] In some embodiments, for each sub-prediction model in the priority prediction model, the second processing device 120 can obtain a trained sub-prediction model by performing the operations in steps S610 - S660 respectively. When all sub-prediction models are trained, the training of the entire priority prediction model is completed.

[0097] In some embodiments, the second processing device 120 may further optimize the priority prediction model based on the feedback from the user's use of the priority prediction model. Specifically, the second processing device 120 may regularly collect the feedback from the user's use of the priority prediction model at a preset period (e.g., weekly, monthly, etc.), and extract data from the construction task data involved in the feedback as the fourth data set. The second processing device 120 may use the fourth data set to train the priority prediction model again, and adjust the model parameters according to the user feedback to optimize the model. For example, if the user feedback shows that the task priority of a certain type of construction task data output by the model is inaccurate, the relevant model features are adjusted for this type of construction task data.

[0098] In some embodiments of this specification, by training the priority model twice based on historical construction task data, the model is fully optimized on the basis of the intermediate prediction model obtained from the first training. The model can exclude the interference of special scenarios, be closer to the actual usage scenario, and improve the accuracy and adaptability of the model.

[0099] It should be noted that the above descriptions of processes 300, 400, and 600 are only for illustration and explanation, and do not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to processes 300, 400, and 600 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification. For example, in process 600, the training of different sub-prediction models can be carried out sequentially or in parallel. Also, for example, steps S620 - S630 and step S640 can be carried out in parallel.

[0100] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only an example and does not constitute a limitation to this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are proposed in this specification, so such modifications, improvements, and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

[0101] At the same time, this specification uses specific terms to describe the embodiments of this specification. Such as "one embodiment", "an embodiment", and / or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that the "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more at different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0102] In addition, unless otherwise specified in the claims, the order of the processing elements and sequences, the use of numerical and alphabetical characters, or the use of other names described in this specification are not used to limit the order of the processes and methods in this specification. Although some currently useful embodiments of the invention are discussed through various examples in the above disclosure, it should be understood that such details are for illustrative purposes only. The appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that conform to the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only through software solutions, such as installing the described system on existing servers or mobile devices.

[0103] Similarly, it should be noted that, in order to simplify the presentation of the disclosure in this specification and thus help the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of this specification, multiple features are sometimes grouped into one embodiment, drawing, or description thereof. However, this method of disclosure does not mean that the features required by the subject matter of this specification are more than those mentioned in the claims. In fact, the features of the embodiments are less than all the features of the individual embodiments disclosed above.

[0104] In some embodiments, numbers are used to describe components and the quantity of attributes. It should be understood that such numbers used to describe the embodiments are modified by the modifiers "about", "approximate" or "substantially" in some examples. Unless otherwise stated, "about", "approximate" or "substantially" indicate that the stated number allows a variation of ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may vary according to the characteristics required by individual embodiments. In some embodiments, the numerical parameters should consider the specified significant digits and adopt the method of retaining the general number of digits. Although the numerical ranges and parameters used in some embodiments of this specification to confirm the breadth of their scope are approximate values, in specific embodiments, such numerical settings are as precise as possible within the feasible range.

[0105] For each patent, patent application, patent application publication, and other materials cited in this specification, such as articles, books, specifications, publications, documents, etc., their entire contents are hereby incorporated into this specification by reference. This excludes the application history documents that are inconsistent with or conflict with the content of this specification, as well as the documents that limit the broadest scope of the claims of this specification (currently or subsequently appended to this specification). It should be noted that if there are inconsistencies or conflicts between the descriptions, definitions, and / or the use of terms in the supplementary materials of this specification and the content described in this specification, the descriptions, definitions, and / or the use of terms in this specification shall prevail.

[0106] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be regarded as consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly presented and described in this specification.

Claims

1. A method for determining construction task priority, characterized in that: include: Obtain task data for multiple construction tasks; The task data is input into a priority prediction model to obtain task priorities of the multiple construction tasks, wherein the priority prediction model is obtained based on historical task data after a first training and a second training, the first training is performed using a first data set, and the second training is performed using a second data set, and a fluctuation of the second data set relative to the first data set is greater than a preset threshold.

2. The method according to claim 1, characterized in that The priority prediction model includes a plurality of sub-prediction models, and the task data is input into the priority prediction model to obtain the task priorities of the plurality of construction tasks, including: Inputting the task data of the multiple construction tasks into each of the multiple sub-prediction models respectively, and obtaining the priority scores of the multiple construction tasks output by each sub-prediction model; For each construction task among the multiple construction tasks, weighted combination is performed on the priority scores of the construction tasks output by each sub-prediction model to obtain a task priority score of the construction task; The priority ranking of the multiple construction tasks is obtained based on the task priority scores of all the construction tasks in the multiple construction tasks.

3. The method according to claim 2, characterized in that In the weighted combination, the weight corresponding to each sub-prediction model is related to the score of the sub-prediction model processing the first data set.

4. The method according to claim 3, characterized in that The score of the sub-prediction model processing the first data set is calculated based on the accuracy and recall of the sub-prediction model processing the first data set.

5. The method according to claim 1, characterized in that The second data set satisfies preset constraints, and the preset constraints are related to the target amount and deadline of the construction task.

6. The method according to claim 5, characterized in that The training process of the priority prediction model includes: Inputting the first data set into an initial prediction model for training to obtain an intermediate prediction model; Dividing the first data set into at least two subsets, iteratively training the intermediate prediction model based on the at least two subsets, and obtaining the accuracy and recall of at least some of the at least two subsets; Calculating a standard deviation of the first data set based on the precision and recall of at least some of the at least two subsets; Inputting the third data set that satisfies the preset constraint condition into the intermediate prediction model to obtain a prediction result of the third data set; Filtering the second data set from the third data set based on the prediction result of the third data set and the standard deviation of the first data set; The intermediate prediction model is trained again using the second data set to obtain the trained priority prediction model.

7. The method according to claim 6, characterized in that The iterative training of the intermediate prediction model based on the at least two subsets to obtain the accuracy and recall of at least some of the at least two subsets includes: In each iteration of the iterative training, one of the at least two subsets is used as a test set, and the other subsets are used as training sets. The intermediate prediction model is trained based on the training set to obtain the accuracy and recall rate corresponding to the test set.

8. The method according to claim 6, characterized in that The prediction result of the third data set includes the accuracy and recall rate of the third data set, and the filtering out the second data set from the third data set based on the prediction result of the third data set and the standard deviation of the first data set includes: Calculating a fluctuation score of the third data set relative to the first data set based on the precision and recall of the third data set and the standard deviation of the first data set; The set of construction task data in the third data set whose fluctuation scores meet the preset threshold requirement is used as the second data set.

9. A construction task priority determination system, characterized in that: Includes data acquisition module, priority prediction module and model training module; The data acquisition module is used to acquire task data of multiple construction tasks; The priority prediction module is used to input the task data into the priority prediction model to obtain the task priorities of the multiple construction tasks; The model training module is used to obtain the priority prediction model based on historical task data through a first training and a second training, the first training is performed using a first data set, the second training is performed using a second data set, and the fluctuation of the second data set relative to the first data set is greater than a preset threshold.

10. A computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the method according to any one of claims 1 to 8.

Images