Task prediction processing method, device, product and medium

Abstract

The disclosed embodiment discloses a task prediction processing method, device, product and medium, for the task to be processed and any candidate task executor among multiple candidate task executors, static features and time series features are obtained; the static features include: task features, executor features and actor features of the task to be processed; the time series features include: the behavior features of the candidate task executors and the change features of the actor of the task to be processed within a preset recent time period; using a neural network model, based on the static features and time series features, the comprehensive probability of any candidate task executor completing multiple goals is predicted, so as to determine the task executor of the task to be processed from multiple candidate task executors based on the comprehensive probability; wherein, completing multiple goals includes: completing the task to be processed and completing the associated goals of the task to be processed. The disclosed embodiment can improve the comprehensive effect of task completion and the achievement of project goals after task completion.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to a task prediction processing method, device, product and medium. Background Art

[0002] Matching tasks with operators and assigning tasks to operators are common problems in the industry. For example, in the field of real estate transactions, in order to improve the work efficiency of real estate agents, the platform will issue tasks to agents, such as negotiating the price of a certain house, showing a certain house to customers, etc., to guide the work of the agent, thereby improving the efficiency of the agent and the platform.

[0003] Normally, a platform will have a large number of projects at the same time, and each project will involve tasks in multiple links. The difficulty level of tasks in the same link may vary in different projects, and the difficulty level of tasks in different links of the same project may also vary. Moreover, the difficulty level and completion status of tasks in each link may affect the achievement of the ultimate project goal.

[0004] In the prior art, task sorting and allocation are mainly based on the single-objective model sorting method: that is, a certain important optimization goal is used as the sorting standard, and the model is used to estimate the goal. For example, in the house task system, an estimation model is constructed to predict whether the task is completed or not, and the probability of whether the task is completed or not output by the estimation model is used as the standard to sort and allocate tasks.

[0005] In the process of realizing the present disclosure, the inventors discovered through research that the above-mentioned sorting method based on the single-objective model only considers a single goal for task sorting, and does not comprehensively consider the impact of the task completion on the achievement of the ultimate goal of the project. Tasks with a high ranking for a single goal may not be of much help in achieving the ultimate goal of the project. For example, in a housing task system, only the probability of whether the predicted task is completed or not output by the estimation model is used to sort tasks, and tasks with a high ranking may not be of much help in achieving the ultimate goal of the project, which is housing transaction. Therefore, the method of sorting and allocating tasks based on the single-objective model sorting method is not good in achieving project goals.

[0006] Therefore, it is crucial to reasonably allocate the project tasks to simultaneously achieve task completion and achievement of project goals after task completion. Summary of the invention

[0007] The embodiments of the present disclosure provide a task prediction processing method and apparatus, equipment, product and medium to improve the comprehensive effect of task completion and achievement of project goals after task completion.

[0008] According to one aspect of an embodiment of the present disclosure, a task prediction processing method is provided, including:

[0009] For the task to be processed and any candidate task executor among multiple candidate task executors, static features and time series features are obtained; wherein the static features include: task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed; the time series features include: behavior features of the candidate task executor and change features of the actor of the task to be processed within a preset recent time period;

[0010] Utilizing a neural network model, based on the static features and the temporal features, the comprehensive probability of any candidate task executor completing multiple objectives is predicted, so as to determine the task executor of the task to be processed from the multiple candidate task executors based on the comprehensive probability; wherein, completing multiple objectives includes: completing the task to be processed and completing the associated objectives of the task to be processed.

[0011] In another embodiment of the method disclosed above, the method of using a neural network model to predict the comprehensive probability of any candidate task performer completing multiple goals based on the static features and the temporal features includes:

[0012] Inputting the static feature into a deep neural network, and outputting a first feature through the deep neural network;

[0013] Inputting the time series feature into a recursive neural network, and outputting a second feature through the recursive neural network;

[0014] Fusing the first feature and the second feature to obtain a fused feature;

[0015] The comprehensive probability of any candidate task executor completing the multiple objectives is predicted based on the fusion features.

[0016] In another embodiment of the above method based on the present disclosure, it also includes:

[0017] Based on the comprehensive probabilities of the multiple candidate task executors completing multiple objectives, the to-be-processed task is assigned to the candidate task executor with the highest comprehensive probability.

[0018] In another embodiment of the above method based on the present disclosure, there are multiple tasks to be processed;

[0019] The step of obtaining static features and temporal features for the task to be processed and any candidate task executor among the plurality of candidate task executors includes:

[0020] For each of the plurality of tasks to be processed, and for each of the tasks to be processed and the plurality of candidate task executors, obtain static features and temporal features;

[0021] The method of using a neural network model to predict the comprehensive probability of any candidate task executor completing multiple objectives based on the static features and the temporal features includes:

[0022] Using a neural network model, based on the static features and the temporal features, predict the comprehensive probability of any candidate task executor completing the multiple objectives corresponding to any task to be processed;

[0023] The method further comprises:

[0024] Based on the order of the comprehensive probabilities of the multiple candidate task executors completing the multiple pending tasks from high to low and the preset task allocation rules, the task executors of each of the multiple pending tasks are determined respectively, and each pending task is allocated to the determined task executors.

[0025] In another embodiment of the method disclosed above, the pending task is a task in a housing task system; the comprehensive probability of completing multiple goals includes: the comprehensive probability of completing the pending task and completing the housing transaction;

[0026] The task characteristics include any one or more of the following characteristics: task type, task magnitude;

[0027] The executor characteristics include any one or more of the following characteristics: the broker's age, the broker's education level, the broker's length of service;

[0028] The actor characteristics include any one or more of the following characteristics: house characteristics and house owner characteristics; the house characteristics include any one or more of the following characteristics: house price, house location, house nature; the owner characteristics include any one or more of the following characteristics: whether the owner has changed houses, whether the owner has bought a house;

[0029] The behavioral characteristics of the candidate task performers include any one or more of the following characteristics: the type and number of tasks performed by the agent;

[0030] The change characteristics of the actor of the task to be processed include any one or more of the following characteristics: price changes of the house, and the number of times each task type is executed on the house.

[0031] In another embodiment of the above method of the present disclosure, the neural network model is obtained by training through the following steps:

[0032] Acquire at least one training sample; wherein each training sample includes static features and temporal features corresponding to a task sample and a task executor sample, and the training sample has a label indicating whether the task sample is completed and a label indicating whether the associated target of the task sample is completed;

[0033] Using a neural network model, based on the static features and time series features of the training samples, the comprehensive probability of the task performer sample corresponding to the training samples completing multiple goals is predicted to obtain a comprehensive probability prediction value; wherein the task performer sample completing multiple goals includes: completing the task sample and completing the associated goals of the task sample;

[0034] Using a preset multi-objective loss function, based on whether the task sample of each training sample in the at least one training sample is completed, whether the associated target of the task sample is completed, and the comprehensive probability prediction value, the loss function value corresponding to each training sample is calculated;

[0035] The neural network model is trained based on the loss function value corresponding to the at least one training sample.

[0036] In another embodiment of the above method based on the present disclosure, the multi-objective loss function is:

[0037]

[0038] Among them, L(w) loss function value, is the comprehensive probability prediction value, y is the label of whether the task sample is completed, z is the label of whether the associated target of the task sample is completed, C is the multi-target loss weight, and the value of C is a real number greater than 0.

[0039] In another embodiment of the method disclosed above, the task sample is a task sample in a housing task system; the task types of the task sample include any one or more of the following categories: bargaining, showing, virtual reality VR showing, VR house explanation, voice follow-up, face-to-face interview, on-site inspection of the house, and getting the key.

[0040] In another embodiment of the above method of the present disclosure, the value of C is determined based on the task type of the task sample:

[0041]

[0042] Wherein, c is a preset fixed weight value, and the value of λ is a real number greater than 1.

[0043] According to another aspect of the embodiments of the present disclosure, a task prediction processing device is provided, comprising:

[0044] A feature acquisition module is used to acquire static features and time series features for the task to be processed and any candidate task executor among multiple candidate task executors; wherein the static features include: task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed; the time series features include: behavior features of the candidate task executor and change features of the actor of the task to be processed within a preset recent time period;

[0045] A prediction module is used to use a neural network model to predict the comprehensive probability of any candidate task executor completing multiple goals based on the static features and the temporal features, so as to determine the task executor of the task to be processed from the multiple candidate task executors based on the comprehensive probability; wherein, completing multiple goals includes: completing the task to be processed and completing the associated goals of the task to be processed.

[0046] In another embodiment of the above device according to the present disclosure, the prediction module includes:

[0047] A deep neural network, configured to receive the static feature and output a first feature;

[0048] A long short-term memory recursive neural network, used for receiving the time series feature and outputting a second feature;

[0049] A feature fusion unit, used for fusing the first feature and the second feature to obtain a fused feature;

[0050] A prediction unit is used to predict the comprehensive probability of any candidate task executor completing the multiple objectives based on the fusion features.

[0051] In another embodiment of the above device based on the present disclosure, it also includes:

[0052] The allocation module is used to allocate the to-be-processed task to the candidate task executor with the highest comprehensive probability based on the comprehensive probability of the multiple candidate task executors completing multiple objectives.

[0053] In another embodiment of the above device based on the present disclosure, there are multiple tasks to be processed;

[0054] The feature acquisition module is specifically used to acquire static features and temporal features for any task to be processed among the multiple tasks to be processed, and for any task to be processed and any candidate task executor among the multiple candidate task executors;

[0055] The prediction module is specifically used to predict the comprehensive probability of any candidate task executor completing the multiple objectives corresponding to any task to be processed based on the static features and the temporal features by using a neural network model;

[0056] The device also includes:

[0057] The allocation module is used to determine the task executor of each of the multiple pending tasks based on the order of the comprehensive probabilities of the multiple candidate task executors completing the multiple pending tasks from high to low and the preset task allocation rules, and allocate each pending task to the determined task executor.

[0058] In another embodiment of the above device based on the present disclosure, the pending task is a task in a housing task system; the comprehensive probability of completing multiple goals includes: the comprehensive probability of completing the pending task and completing the housing transaction;

[0059] The task characteristics include any one or more of the following characteristics: task type, task magnitude;

[0060] The executor characteristics include any one or more of the following characteristics: the broker's age, the broker's education level, the broker's length of service;

[0061] The actor characteristics include any one or more of the following characteristics: house characteristics and house owner characteristics; the house characteristics include any one or more of the following characteristics: house price, house location, house nature; the owner characteristics include any one or more of the following characteristics: whether the owner has changed houses, whether the owner has bought a house;

[0062] The behavioral characteristics of the candidate task performers include any one or more of the following characteristics: the type and number of tasks performed by the agent;

[0063] The change characteristics of the actor of the task to be processed include any one or more of the following characteristics: price changes of the house, and the number of times each task type is executed on the house.

[0064] In another embodiment of the above device based on the present disclosure, it also includes:

[0065] A sample acquisition module is used to acquire at least one training sample; wherein each training sample includes static features and temporal features corresponding to a task sample and a task executor sample, and the training sample has a label indicating whether the task sample is completed and a label indicating whether the associated target of the task sample is completed;

[0066] The prediction module is further used to use a neural network model to predict the comprehensive probability of the task performer sample corresponding to the training sample completing multiple goals based on the static characteristics and time series characteristics of the training sample, and obtain a comprehensive probability prediction value; wherein the task performer sample completing multiple goals includes: completing the task sample and completing the associated goals of the task sample;

[0067] The training module is used to use a preset multi-objective loss function to calculate the loss function value corresponding to each training sample based on whether the task sample of each training sample in the at least one training sample has completed the label, whether the associated target of the task sample has completed the label and the comprehensive probability prediction value; and train the neural network model based on the loss function value corresponding to the at least one training sample.

[0068] In another embodiment of the above device based on the present disclosure, the multi-objective loss function is:

[0069]

[0070] Among them, L(w) loss function value, is the comprehensive probability prediction value, y is the label of whether the task sample is completed, z is the label of whether the associated target of the task sample is completed, C is the multi-target loss weight, and the value of C is a real number greater than 0.

[0071] In another embodiment of the above-mentioned device based on the present disclosure, the task sample is a task sample in a housing task system; the task types of the task sample include any one or more of the following categories: bargaining, showing, virtual reality VR showing, VR house explanation, voice follow-up, face-to-face interview, on-site inspection of the house, and getting the key.

[0072] In another embodiment of the above device according to the present disclosure, the value of C is determined based on the task type of the task sample:

[0073]

[0074] Wherein, c is a preset fixed weight value, and the value of λ is a real number greater than 1.

[0075] According to another aspect of the embodiments of the present disclosure, an electronic device is provided, including:

[0076] Memory for storing computer programs;

[0077] The processor is used to execute the computer program stored in the memory, and when the computer program is executed, the task prediction processing method described in any of the above embodiments of the present disclosure is implemented.

[0078] According to another aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the task prediction processing method described in any of the above embodiments of the present disclosure is implemented.

[0079] According to another aspect of an embodiment of the present disclosure, a computer program product is provided, including a computer program / instruction, which, when executed by a processor, implements the task prediction processing method described in any of the above embodiments of the present disclosure.

[0080] Based on the task prediction processing method and apparatus, device, product and medium provided by the above-mentioned embodiments of the present disclosure, by obtaining static features (including task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed) and time series features (including behavioral features of the candidate task executor and change features of the actor of the task to be processed in a preset recent time period) for the task to be processed and any candidate task executor among multiple candidate task executors, then, using a neural network model, based on the static features and time series features, predict the comprehensive probability of any candidate task executor completing multiple goals (including completing the task to be processed and completing its associated goals), so as to determine the task executor of the task to be processed from multiple candidate task executors based on the comprehensive probability. Thus, the embodiments of the present disclosure can comprehensively consider the impact of task completion and the achievement of associated goals (such as the final goal of the project) after the task is completed, to determine the task executor and guide the issuance of tasks, so as to simultaneously achieve the comprehensive effect of task completion and the achievement of project goals after the task is completed, which helps to improve the efficiency of task execution and the overall efficiency of the project.

[0081] The technical solution of the present disclosure is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] The accompanying drawings, which constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0083] The present disclosure may be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0084] Figure 1 The present invention is a flowchart of an embodiment of a task prediction processing method.

[0085] Figure 2 The flowchart is another embodiment of the task prediction processing method disclosed in the present invention.

[0086] Figure 3 The present invention is a flowchart of another embodiment of the task prediction processing method.

[0087] Figure 4 The figure is a schematic diagram of the structure of an embodiment of the task prediction processing device disclosed in the present invention.

[0088] Figure 5 It is a structural schematic diagram of another embodiment of the task prediction processing device disclosed in the present invention.

[0089] Figure 6 is a block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0090] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure unless otherwise specifically stated.

[0091] Those skilled in the art can understand that the terms "first" and "second" in the embodiments of the present disclosure are only used to distinguish different steps, devices or modules, etc., and neither represent any specific technical meaning nor indicate the necessary logical order between them.

[0092] It should also be understood that in the embodiments of the present disclosure, “plurality” may refer to two or more than two, and “at least one” may refer to one, two, or more than two.

[0093] It should also be understood that any component, data or structure mentioned in the embodiments of the present disclosure can generally be understood as one or more, unless explicitly limited or otherwise indicated in the context.

[0094] In addition, the term "and / or" in the present disclosure is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present disclosure generally indicates that the associated objects before and after are in an "or" relationship.

[0095] It should also be understood that the description of the various embodiments in the present disclosure focuses on the differences between the various embodiments, and the same or similar aspects thereof can be referenced to each other, and for the sake of brevity, they will not be described one by one.

[0096] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0097] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0098] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered as part of the specification.

[0099] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0100] The embodiments of the present disclosure can be applied to electronic devices such as terminal devices, computer systems, servers, etc., which can operate with many other general or special computing system environments or configurations. Examples of well-known terminal devices, computing systems, environments and / or configurations suitable for use with electronic devices such as terminal devices, computer systems, servers, etc. include, but are not limited to: personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network personal computers, small computer systems, large computer systems, and distributed cloud computing technology environments including any of the above systems, etc.

[0101] Electronic devices such as terminal devices, computer systems, servers, etc. can be described in the general context of computer system executable instructions (such as program modules) executed by computer systems. Generally, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in a distributed cloud computing environment, where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules can be located on local or remote computing system storage media including storage devices.

[0102] Figure 1 FIG. 1 is a flowchart of an embodiment of the task prediction processing method disclosed in the present invention. Figure 1 As shown, the task prediction processing method of this embodiment includes:

[0103] 102 , for the task to be processed and any candidate task executor among the plurality of candidate task executors, obtain static features and temporal features.

[0104] Among them, static features may include, but are not limited to, the following three dimensions of features: task features of the task to be processed (i.e., the features of the task itself), executor features of the candidate task executors, and actor features of the task to be processed. Temporal features may include, but are not limited to, the following two dimensions of features: behavioral features of the candidate task executors and change features of the actor of the task to be processed within a preset recent time period (e.g., within the last 7 days, within the last 30 days, etc.).

[0105] 104. Using a neural network model, based on the above static features and time series features, predict the comprehensive probability of any candidate task executor completing multiple goals, so as to determine the task executor of the task to be processed from multiple candidate task executors based on the comprehensive probability.

[0106] Completing multiple goals includes: completing the pending task and completing the associated goals of the pending task. The associated goals are not limited and can be the next stage goal directly associated with the pending task, or the final goal of the project to which the pending task belongs, etc.

[0107] Based on the task prediction processing method provided by the above-mentioned embodiment of the present disclosure, by obtaining static features (including task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed) and time series features (including behavioral features of the candidate task executor and change features of the actor of the task to be processed in a preset recent time period) for any candidate task executor among the task to be processed and multiple candidate task executors, and then using a neural network model, based on the static features and time series features, predict the comprehensive probability of any candidate task executor completing multiple goals (including completing the task to be processed and completing its associated goals), so as to determine the task executor of the task to be processed from multiple candidate task executors based on the comprehensive probability. Thus, the embodiment of the present disclosure can comprehensively consider the impact of task completion and the achievement of associated goals (such as the final goal of the project) after the task is completed, to determine the task executor and guide the issuance of tasks, so as to simultaneously achieve the comprehensive effect of task completion and the achievement of project goals after the task is completed, which helps to improve the efficiency of task execution and the overall efficiency of the project.

[0108] Figure 2 FIG. 1 is a flowchart of another embodiment of the task prediction processing method disclosed in the present invention. Figure 2 As shown, in Figure 1 Based on the illustrated embodiment, 104 may include:

[0109] At 1042, the static feature is input into a deep neural network (DNN), and the DNN outputs a first feature.

[0110] Through DNN, the static features of the input can be extracted to obtain the first features.

[0111] 1044, input the time series feature into a recurrent neural network (RNN), and output the second feature through the RNN.

[0112] RNN is a type of neural network that can be used to process time series data. Time series data refers to data collected at different time points. This type of data reflects the state or degree of change of a thing or phenomenon over time.

[0113] Optionally, in a possible implementation, a Long Short Term Memory (LSTM) network is a deformation network of an RNN, and a traditional RNN, LSTM network or other neural network with time series data processing capability can be used as an RNN to learn the input time series feature to obtain a second feature. Since the RNN has a feature memory function, the input time series feature can be learned, the change state or degree of the time series feature over time can be analyzed, and the second feature can be obtained and output, that is, the output second feature can reflect the change state or degree of the time series feature over time.

[0114] 1046, the first feature and the second feature are fused to obtain a fused feature.

[0115] The way to fuse the first feature and the second feature may be to directly concatenate or add the first feature and the second feature, or to fuse the first feature and the second feature in other ways, which is not limited in the embodiments of the present disclosure.

[0116] Optionally, in a possible implementation, the first feature and the second feature may be input into a neural network including a fully connected layer, and through the fully connected layer, the first feature and the second feature may be directly concatenated or fused in other ways to obtain a fused feature.

[0117] 1048, predicting the comprehensive probability of any corresponding candidate task executor completing multiple goals based on the fusion feature.

[0118] Optionally, in a possible implementation, an activation function, such as softmax, sigmoid, etc., can be used to predict the comprehensive probability of completing multiple goals based on the fused features, and a numerical value in the value range [0,1] can be obtained as the comprehensive probability of completing multiple goals for any corresponding candidate task executor.

[0119] Based on this embodiment, DNN can be used to further learn and extract features of static features, and the characteristics of RNN for processing time series data can be used to learn and extract features of temporal features, and the first feature and the second feature can be fused. The comprehensive probability of candidate task executors completing multiple goals can be predicted based on the obtained fused features. In this way, static features such as task characteristics of the task to be processed, executor characteristics of the candidate task executors, and characteristics of the actor of the task to be processed, as well as temporal features such as behavioral characteristics of the candidate task executors and change characteristics of the actor of the task to be processed in a preset recent time period can be comprehensively considered to accurately predict the comprehensive probability of candidate task executors completing multiple goals, thereby improving the accuracy of the prediction results of the comprehensive probability of candidate task executors completing multiple goals.

[0120] Optionally, in the above embodiment of the present disclosure, after predicting the comprehensive probability of any candidate task executor completing multiple objectives, the task to be processed can be assigned to the candidate task executor with the highest comprehensive probability based on the comprehensive probability of multiple candidate task executors completing multiple objectives.

[0121] Specifically, the candidate task executor with the highest comprehensive probability of completing multiple objectives can be directly selected from the comprehensive probabilities of multiple candidate task executors completing multiple objectives, and the task to be processed can be assigned to the candidate task executor with the highest comprehensive probability; or, the comprehensive probabilities of multiple candidate task executors completing multiple objectives can be sorted in descending order, and then the task to be processed can be assigned to the candidate task executor with the highest ranking.

[0122] Based on this embodiment, it can be applied to the scenario of a single processing task and multiple candidate task executors. After predicting the comprehensive probability of each candidate task executor completing multiple goals based on the disclosed embodiment, the candidate task executor with the highest comprehensive probability is selected to perform the task, thereby achieving the best optimization of the task execution efficiency and the overall efficiency of the project.

[0123] Optionally, in the above embodiment of the present disclosure, when there are multiple tasks to be processed, the above operations 102-104 can be performed for any of the multiple tasks to be processed, and the comprehensive probability of any candidate task executor completing the multiple goals corresponding to any task to be processed can be obtained. Specifically, in 102, static features and time series features can be obtained for any of the multiple tasks to be processed, any of the tasks to be processed and any of the multiple task executors. In 104, a neural network model is used to predict the comprehensive probability of any candidate task executor completing the multiple goals corresponding to any task to be processed based on the static features and time series features corresponding to any of the tasks to be processed and any of the candidate task executors. Afterwards, the task executors of each of the multiple tasks to be processed can be determined based on the order of the comprehensive probabilities of multiple task executors completing the multiple tasks to be processed from high to low and the preset task allocation rules, and each task to be processed can be allocated to the determined task executor.

[0124] Specifically, in one possible implementation method, the comprehensive probabilities of multiple candidate task executors completing multiple pending tasks can be sorted in descending order, and based on preset task assignment rules, for example, the maximum number of tasks that can be assigned to each candidate task executor is limited to N (for example, the maximum number of assigned tasks cannot be greater than 3), N is an integer greater than 0, and in descending order of comprehensive probability, candidate task executors with higher comprehensive probabilities are selected as task executors for each pending task, and each pending task is assigned to a determined task executor.

[0125] Based on the maximum number N of tasks that can be assigned to each candidate task executor, the tasks that can be assigned to each candidate task executor can be restricted. For example, in order of comprehensive probability from high to low, if M (M is greater than N) pending tasks all correspond to the same candidate task executor, then N pending tasks are selected in order of comprehensive probability from high to low and issued to the candidate task executor. For the other MN tasks, the corresponding candidate task executors are selected in order of comprehensive probability from high to low and the maximum number N of tasks that can be assigned to each candidate task executor.

[0126] Based on this embodiment, when candidate task executors with higher comprehensive probabilities are selected in order from high to low comprehensive probabilities as task executors for each task to be processed, the tasks that can be assigned to each candidate task executor are restricted based on the maximum number N of tasks that can be assigned to each candidate task executor. This can avoid the same candidate task executor being assigned tasks that exceed its execution capacity, while achieving the comprehensive effect of completing the task and achieving the project goals after the task is completed, ensuring that the task can be effectively executed.

[0127] Optionally, the task prediction processing method of any embodiment of the present disclosure can be used for processing tasks in any field and any project. For example, when the embodiment of the present disclosure is applied to task management in a house source task system, the pending tasks in the above embodiment are tasks in the house source task system, and accordingly, the comprehensive probability of completing multiple goals includes the comprehensive probability of completing the pending tasks and completing the house source transaction. Task characteristics, for example, may include but are not limited to any one or more of the following characteristics: task type, task amplitude, etc. Executor characteristics, for example, may include but are not limited to any one or more of the following characteristics: broker's age, broker's education, broker's length of service, etc. Action party characteristics, for example, may include but are not limited to any one or more of the following characteristics: house source characteristics and house source owner characteristics, etc. House source characteristics, for example, may include but are not limited to any one or more of the following characteristics: house source price, house source location, house source nature, etc. Owner characteristics, for example, may include but are not limited to any one or more of the following characteristics: whether the owner has changed houses, whether the owner has bought a house, etc. The behavior characteristics of the candidate task performers may include, but are not limited to, any one or more of the following characteristics: the type and number of tasks performed by the broker (e.g., the number of times the property has been shown in the past seven days, the bargaining range, the number of VR house presentations), etc. The change characteristics of the actor of the task to be processed may include, but are not limited to, any one or more of the following characteristics: price changes of the property (e.g., price changes of the property in the past seven days), the number of times each type of task has been performed on the property (e.g., the number of times the property has been shown in the past seven days), etc. Among them, each task type may include, but is not limited to, any one or more of the following characteristics: bargaining, showing, virtual reality (VR) showing, VR house presentation, voice follow-up, face-to-face interview, property inspection, key acquisition, etc. The task range is the range under the corresponding task type, such as bargaining for 400,000, bargaining for 600,000, VR house presentation 8 times, 10 times, showing 8 times, 10 times, etc.

[0128] In practical applications, relevant information of pending tasks, relevant information of candidate task executors, and relevant information of the parties acting on pending tasks can be obtained from an information database for docking with a task dispatching system used to implement task allocation and distribution, and the above-mentioned static features and temporal features can be extracted from the above-mentioned information. When extracting corresponding task features (including static features and temporal features) from relevant information, if the corresponding information is expressed as numbers, the numbers can be directly extracted as corresponding task features; if the corresponding information is expressed as text, the text can be converted into corresponding task features. As an example, existing word embedding methods, such as one-hot algorithms, word to vector (word2vec) algorithms, etc., can be used to convert text into corresponding features. The embodiments of the present disclosure do not limit the specific implementation method for extracting task features.

[0129] Based on this embodiment, static features (task characteristics such as task type and task magnitude, executor characteristics such as broker's age, broker's education level, broker's length of service, property characteristics and property owner characteristics, and characteristics of the party waiting to process the task) and time series features (behavioral characteristics such as the type and number of tasks performed by the broker within a preset recent time period, price changes of the property, the number of times each type of task is performed on the property, and change characteristics of the party waiting to process the task) can be obtained for the pending tasks and multiple candidate brokers. The comprehensive probability of completing the pending tasks and completing the property transaction can be predicted, and the broker for the pending tasks can be determined from multiple candidate brokers. The comprehensive effect of task completion and property transaction after task completion can be achieved simultaneously, which helps to improve the task execution efficiency and the overall efficiency of the property task system.

[0130] In addition, in the embodiment of the present disclosure, the neural network model can be obtained by pre-training with training samples before 102. Figure 3 FIG. 1 is a flowchart of another embodiment of the task prediction processing method disclosed in the present invention. Figure 3 As shown, in this embodiment, the neural network model is obtained by training through the following operations:

[0131] 202. Obtain at least one training sample.

[0132] Among them, each training sample includes static features and temporal features corresponding to a task sample and a task executor sample, and each training sample has a label of whether the task sample is completed and a label of whether the associated target of the task sample is completed, that is, it has multiple labels. The static features and temporal features corresponding to a task sample and a task executor sample in each training sample, the specific features included can refer to the records of static features and temporal features in the above embodiments of the present disclosure, and will not be repeated here.

[0133] 204, using the neural network model, based on the static features and time series features of each training sample, predict the comprehensive probability of the task executor sample corresponding to each training sample completing the multiple goals, and obtain a comprehensive probability prediction value.

[0134] Among them, the task executor sample completes multiple goals including: completing the task sample and completing the associated goals of the task sample.

[0135] 206. Using a preset multi-objective loss function, based on whether the task sample of each training sample in the at least one training sample is completed, whether the associated target of the task sample is completed, and the comprehensive probability prediction value, calculate the loss function value corresponding to each training sample.

[0136] 208. Train the neural network model based on the loss function value corresponding to the at least one training sample.

[0137] Wherein, the above operations 202-208 or 204-208 are iteratively executed operations, and the training of the neural network model is realized by iteratively executing the above operations 202-208 or 204-208, that is, optimizing and adjusting the parameters of the neural network model until the preset training completion conditions are met to obtain a trained neural network model. The preset training completion conditions, for example, may include but are not limited to any one or more items: the number of iterative training of the neural network model (that is, the number of iterative executions of the above operations 202-208 or 204-208) reaches a preset number of times (for example, 200 times), the loss function value corresponding to the above at least one training sample is less than or equal to a preset threshold, and so on. Wherein, the loss function value corresponding to the above at least one training sample is less than or equal to the preset threshold, for example, it may be that the loss function value corresponding to each training sample is less than or equal to the preset threshold, or the average of the loss function values ​​corresponding to at least one training sample is less than or equal to the preset threshold, and the embodiments of the present disclosure do not limit this.

[0138] The loss function is a way to measure the predicted value of the neural network model and the actual value (label) through a mathematical formula. The design and construction of the loss function is the core design point related to the performance of the neural network model.

[0139] The embodiment of the present disclosure proposes a multi-objective loss function. In a possible implementation, the multi-objective loss function may be:

[0140]

[0141] In the above formula (1), L(w) is the loss function value corresponding to each training sample calculated using the multi-objective loss function, which can also be called the neural network model loss; is the comprehensive probability prediction value output by the neural network model; y is the real label of whether the task sample is completed (i.e., the actual value). In practical applications, the specific value can be 1 or 0. The value of the task sample completion label can be set to 1, and the value of the task sample incomplete label can be set to 0; z is the label of whether the associated target of the task sample is completed. In practical applications, the specific value can be 1 or 0. The value of the task sample associated target completion label can be set to 1, and the value of the task sample associated target incomplete label can be set to 0; C is the multi-target loss weight, and the value of C is a real number greater than 0.

[0142] In the above formula (1), It represents the prediction loss of the comprehensive probability prediction value of the output of the neural network model for completing the task sample and the associated target of the task sample. Represents the prediction loss of the comprehensive probability prediction value output for the completed task sample. Based on the multi-objective loss function The influence of samples that complete the task and the associated targets of the task on the prediction loss of the neural network model is strengthened, and the penalty for prediction errors of such samples is increased. Therefore, after the training of the neural network model is completed, the accuracy of the comprehensive probability of predicting the task corresponding to the task sample and completing the associated targets of the task can be improved.

[0143] For example, in the house task system, one column of labels y is the label of whether the task sample is completed, and the other column of labels z is the label of whether the house is sold. Both columns of labels are applied to the above multi-objective loss function to realize the learning of multi-objective information by the neural network model. After the neural network model training is completed, the comprehensive probability of completing the task corresponding to the task sample and completing the associated goals of the task can be accurately predicted.

[0144] Optionally, when the embodiments of the present disclosure are applied to task management in a house task system, the task samples in the above embodiments may be task samples in the house task system. Accordingly, the task types of the task samples may include, but are not limited to, any one or more of the following: bargaining, showing, VR showing, VR house explanation, voice follow-up, face-to-face interview, house inspection, getting the key, etc.

[0145] By adjusting the value of C in the above formula (2), we can obtain a neural network model that can meet the requirements of both the task completion rate and the associated goal achievement rate after the task is completed, thereby realizing a task comprehensive probability prediction algorithm that combines multiple goals. Therefore, how to select the value of C is crucial.

[0146] Optionally, in a possible implementation, the value of C is determined based on the task type of the task sample, for example, it can be determined in the following manner:

[0147]

[0148] In the above formula (2), c is a preset fixed weight value, the value of c is greater than 1, and the value of λ is a real number greater than 1.

[0149] In most task systems, tasks can often be divided into several different task types, and each task type has its own significant characteristics. Based on this embodiment, the value of the multi-objective loss weight C can be adjusted according to different task types, so that the trained neural network model can be applied to various task types in the task system, and can reasonably weigh the completion of tasks of different task types and the impact on the achievement of related goals after the completion of the tasks, so as to achieve the comprehensive effect of completing the tasks and achieving the final project goals after the completion of the tasks.

[0150] For example, in the house task system, the task types can be divided into eight types: bargaining, showing, VR showing, VR house explanation, voice follow-up, face-to-face interview, house inspection, and key collection. Among them, the bargaining task has a more obvious role in promoting the goal of house transaction compared with other task types, but because the bargaining task is often difficult, the predicted task completion probability is not high. Therefore, for the bargaining task, its weight can be penalized. Through multiple offline tests, the optimal values ​​of λ and c can be obtained, so as to train a neural network model that integrates multiple indicators.

[0151] The inventor of the present disclosure, based on a certain number of tasks, has conducted multiple offline tests and found that when the value of λ is in the range [1,2] and the value of c is in the range [30,80], under the same other conditions, task allocation based on the comprehensive probability predicted by the neural network model can achieve the optimization of the task completion rate while meeting the requirement of the task completion rate. The task completion rate refers to the ratio of the number of completed tasks to the number of all tasks in all tasks participating in the offline test. The task completion rate refers to the ratio of the number of completed tasks and the number of associated targets of the completed tasks to the number of completed tasks.

[0152] When the embodiments of the present disclosure are applied to a housing task system, they can be implemented in the following ways:

[0153] Obtain a task list of all tasks to be assigned, and select one task to be assigned from the task list in turn as the task to be processed in the above embodiment;

[0154] Get all candidate brokers in the broker list;

[0155] Obtain relevant information of pending tasks, relevant information of all candidate brokers and relevant information of the parties to the pending tasks from the information database, and extract static features and temporal features from the above information. Static features include: task characteristics such as the type and magnitude of the pending tasks, executor characteristics such as the age, education and length of service of the brokers, and characteristics of the properties and owners of the properties; and characteristics of the parties to the pending tasks. Temporal features include: behavioral characteristics such as the types and times of tasks performed by brokers in the past 7 days, price changes of properties, and the number of times each type of task is performed on properties; and characteristics of changes in the parties to the pending tasks.

[0156] Using the trained neural network model, based on the above embodiment, predict the comprehensive probability of each candidate broker completing the pending tasks and completing the house transaction;

[0157] After predicting the comprehensive probability of each candidate broker completing each pending task and completing the housing transaction in the above manner, all candidate brokers are sorted in descending order based on their comprehensive probabilities for all pending tasks;

[0158] Based on the maximum number N (for example, 3) of tasks that can be assigned to the same candidate broker, candidate brokers with higher comprehensive probabilities are selected in descending order as brokers for each task to be processed, and each task to be processed is assigned to a determined broker.

[0159] Any task prediction processing method provided in the embodiments of the present disclosure may be executed by any appropriate device with data processing capabilities, including but not limited to: terminal devices and servers. Alternatively, any task prediction processing method provided in the embodiments of the present disclosure may be executed by a processor, such as the processor executing any task prediction processing method mentioned in the embodiments of the present disclosure by calling corresponding instructions stored in a memory. This will not be described in detail below.

[0160] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0161] Figure 4 FIG. 1 is a schematic diagram of a structure of an embodiment of a task prediction processing device disclosed in the present invention. The device of this embodiment can be used to implement the above-mentioned embodiments of the task prediction processing method disclosed in the present invention. Figure 4 As shown, the task prediction processing device of this embodiment includes: a feature acquisition module 302 and a prediction module 304. Among them:

[0162] The feature acquisition module 302 is used to acquire static features and time series features for the task to be processed and any candidate task executor among the multiple candidate task executors. The static features may include: task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed; the time series features may include: behavior features of the candidate task executor and change features of the actor of the task to be processed within a preset recent time period.

[0163] The prediction module 304 is used to use a neural network model to predict the comprehensive probability of any candidate task performer completing multiple goals based on the static features and the time series features, so as to determine the task performer of the task to be processed from multiple candidate task performers based on the comprehensive probability. Wherein, the completion of multiple goals includes: completing the task to be processed and completing the associated goals of the task to be processed.

[0164] Based on the task prediction processing device provided by the above-mentioned embodiment of the present disclosure, by obtaining static features (including task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed) and time series features (including behavioral features of the candidate task executor and change features of the actor of the task to be processed in a preset recent time period) for the task to be processed and any candidate task executor among multiple candidate task executors, then, using a neural network model, based on the static features and time series features, predict the comprehensive probability of any candidate task executor completing multiple goals (including completing the task to be processed and completing its associated goals), so as to determine the task executor of the task to be processed from multiple candidate task executors based on the comprehensive probability. Therefore, the embodiment of the present disclosure can comprehensively consider the impact of task completion and the achievement of associated goals (such as the final goal of the project) after the task is completed, to determine the task executor and guide the issuance of tasks, so as to simultaneously achieve the comprehensive effect of task completion and the achievement of project goals after the task is completed, which helps to improve the efficiency of task execution and the overall efficiency of the project.

[0165] Figure 5 FIG. 1 is a schematic diagram of another embodiment of the task prediction processing device disclosed in the present invention. Figure 5 As shown, in Figure 4 On the basis of the illustrated embodiment, in the task prediction processing device of this embodiment, the prediction module 302 may include: DNN3022, for receiving the static feature and outputting the first feature; RNN3024, for receiving the time series feature and outputting the second feature; a feature fusion unit 3026, for fusing the first feature and the second feature to obtain a fused feature; and a prediction unit 3028, for predicting the comprehensive probability of any candidate task executor completing the multiple objectives based on the fused feature.

[0166] Optionally, see Figure 5 In the task prediction processing device of the above embodiment, an allocation module 306 may also be included, which is used to allocate the task to be processed to the candidate task executor with the highest comprehensive probability based on the comprehensive probability of the multiple candidate task executors completing multiple goals.

[0167] Optionally, in a possible implementation, there are multiple tasks to be processed. Accordingly, the feature acquisition module 302 is specifically used to obtain static features and time series features for any task to be processed among the multiple tasks to be processed, and for any task to be processed and any candidate task executor among the multiple candidate task executors; the prediction module 304 is specifically used to use a neural network model to predict the comprehensive probability of any candidate task executor completing the multiple goals corresponding to any task to be processed based on the static features and the time series features. Accordingly, see Figure 5 In the task prediction and processing device of the above embodiment, the allocation module 306 is used to determine the task executor of each of the multiple tasks to be processed based on the order from high to low of the comprehensive probabilities corresponding to the completion of the multiple tasks to be processed by the multiple candidate task executors and the preset task allocation rules, and assign each task to be processed to the determined task executor.

[0168] Optionally, the task prediction processing device of any embodiment of the present disclosure can be used for processing tasks in any field and any project. For example, when the embodiment of the present disclosure is applied to task management in a housing task system, the pending tasks in the above embodiment are tasks in the housing task system, and accordingly, the comprehensive probability of completing multiple goals includes the comprehensive probability of completing the pending tasks and completing the housing transaction. Task characteristics may include, but are not limited to, any one or more of the following characteristics: task type, task range, etc. Executor characteristics may include, but are not limited to, any one or more of the following characteristics: broker's age, broker's education, broker's length of service, etc. Actor characteristics may include, but are not limited to, any one or more of the following characteristics: housing characteristics and housing owner characteristics, etc. Housing characteristics may include, but are not limited to, any one or more of the following characteristics: housing price, housing location, housing nature, etc. Owner characteristics may include, but are not limited to, any one or more of the following characteristics: whether the owner changes houses, whether the owner has bought a house, etc. Behavioral characteristics of candidate task executors may include, but are not limited to, any one or more of the following characteristics: the type and number of tasks performed by the broker, etc. The change characteristics of the actor of the task to be processed may include, but are not limited to, any one or more of the following characteristics: price changes of the house, the number of times each task type is executed for the house, etc. Among them, each task type may include, but is not limited to, any one or more of the following characteristics: bargaining, showing, VR showing, VR house explanation, voice follow-up, face-to-face interview, house inspection, key collection, etc.

[0169] Optionally, see Figure 5 In the task prediction processing device of the above embodiment, a sample acquisition module 308 and a training module may also be included.

[0170] The sample acquisition module 308 is used to acquire at least one training sample; wherein each training sample includes static features and temporal features corresponding to a task sample and a task executor sample, and each training sample has a label of whether the task sample is completed and a label of whether the associated target of the task sample is completed.

[0171] Correspondingly, in this embodiment, the prediction module 304 can also be used to utilize a neural network model to predict the comprehensive probability of the task executor sample corresponding to the training sample completing multiple goals based on the static characteristics and time series characteristics of the training sample, and obtain a comprehensive probability prediction value; wherein, the task executor sample completing multiple goals includes: completing the task sample and completing the associated goals of the task sample.

[0172] The training module 310 is used to calculate the loss function value corresponding to each training sample based on whether the task sample of each training sample in the at least one training sample is completed, whether the associated target of the task sample is completed, and the comprehensive probability prediction value using a preset multi-objective loss function; and train the neural network model based on the loss function value corresponding to the at least one training sample.

[0173] Optionally, in a possible implementation, the multi-objective loss function may be:

[0174]

[0175] In the above formula (1), L(w) is the loss function value corresponding to each training sample calculated using the multi-objective loss function. is the comprehensive probability prediction value, y is the label of whether the task sample is completed, z is the label of whether the associated target of the task sample is completed, C is the multi-target loss weight, and the value of C is a real number greater than 0.

[0176] Optionally, when the embodiments of the present disclosure are applied to task management in a house task system, the task samples in the above embodiments may be task samples in the house task system. Accordingly, the task types of the task samples may include, but are not limited to, any one or more of the following: bargaining, showing, VR showing, VR house explanation, voice follow-up, face-to-face interview, house inspection, getting the key, etc.

[0177] Optionally, in a possible implementation, the value of C is determined based on the task type of the task sample, for example, it can be determined in the following manner:

[0178]

[0179] In the above formula (2), c is a preset fixed weight value, the value of c is greater than 1, and the value of λ is a real number greater than 1.

[0180] In addition, an embodiment of the present disclosure further provides an electronic device, including:

[0181] Memory for storing computer programs;

[0182] The processor is used to execute the computer program stored in the memory, and when the computer program is executed, the task prediction processing method described in any of the above embodiments of the present disclosure is implemented.

[0183] Below, reference Figure 6 The electronic device according to the embodiment of the present disclosure is described. The electronic device may be any one or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the collected input signals from them.

[0184] Figure 6 FIG. 1 is a block diagram of an electronic device according to an embodiment of the present disclosure. Figure 6 As shown, the electronic device includes one or more processors and memory.

[0185] The processor may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0186] The memory may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor may execute the program instructions to implement the task prediction processing method of each embodiment of the present disclosure described above and / or other desired functions.

[0187] In one example, the electronic device may further include: an input device and an output device, and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0188] In addition, the input device may also include, for example, a keyboard, a mouse, and the like.

[0189] The output device can output various information to the outside, including determined distance information, direction information, etc. The output device can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0190] Of course, to simplify, Figure 6Only some of the components related to the present disclosure in the electronic device are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device may further include any other appropriate components.

[0191] In addition to the above-mentioned methods and devices, an embodiment of the present disclosure may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the task prediction processing method according to various embodiments of the present disclosure described in the above part of this specification.

[0192] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0193] In addition, an embodiment of the present disclosure may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enables the processor to execute the steps of the task prediction processing method according to various embodiments of the present disclosure described in the above part of this specification.

[0194] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0195] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0196] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0197] The block diagrams of the devices, apparatuses, equipment, and systems involved in this disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including," "comprising," "having," and the like are open words, referring to "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or," and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0198] The method and apparatus of the present disclosure may be implemented in many ways. For example, the method and apparatus of the present disclosure may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present disclosure are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present disclosure may also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present disclosure. Therefore, the present disclosure also covers a recording medium storing a program for executing the method according to the present disclosure.

[0199] It should also be noted that in the apparatus, device and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0200] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0201] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A task prediction processing method, characterized in that: include: For the task to be processed and any candidate task executor among multiple candidate task executors, static features and time series features are obtained; wherein the static features include: task features of the task to be processed, executor features of the candidate task executor and actor features of the task to be processed; the time series features include: behavior features of the candidate task executor and change features of the actor of the task to be processed within a preset recent time period; the task to be processed is a task in a housing source task system, and the actor features include at least one of housing source features and housing source owner features; A neural network model is used to predict the comprehensive probability of any candidate task executor completing multiple goals based on the static features and the temporal features, so as to determine the task executor of the pending task from the multiple candidate task executors based on the comprehensive probability; wherein, the completion of multiple goals includes: completing the pending task and completing the associated goals of the pending task; the associated goals include at least one of the goals of the next stage directly associated with the pending task and the final goals of the project to which the pending task belongs; and the comprehensive probability of completing multiple goals includes: the comprehensive probability of completing the pending task and completing the housing transaction.

2. The method according to claim 1, characterized in that The method of using a neural network model to predict the comprehensive probability of any candidate task executor completing multiple objectives based on the static features and the temporal features includes: Inputting the static feature into a deep neural network, and outputting a first feature through the deep neural network; Inputting the time series feature into a recursive neural network, and outputting a second feature through the recursive neural network; Fusing the first feature and the second feature to obtain a fused feature; The comprehensive probability of any candidate task executor completing the multiple objectives is predicted based on the fusion features.

3. The method according to claim 1, characterized in that Also includes: Based on the comprehensive probabilities of the multiple candidate task executors completing multiple objectives, the to-be-processed task is assigned to the candidate task executor with the highest comprehensive probability.

4. The method according to claim 1, characterized in that: There are multiple tasks to be processed; The step of obtaining static features and temporal features for the task to be processed and any candidate task executor among the plurality of candidate task executors includes: For each of the plurality of tasks to be processed, and for each of the tasks to be processed and the plurality of candidate task executors, obtain static features and temporal features; The method of using a neural network model to predict the comprehensive probability of any candidate task executor completing multiple objectives based on the static features and the temporal features includes: Using a neural network model, based on the static features and the temporal features, predict the comprehensive probability of any candidate task executor completing the multiple objectives corresponding to any task to be processed; The method further comprises: Based on the order of the comprehensive probabilities of the multiple candidate task executors completing the multiple pending tasks from high to low and the preset task allocation rules, the task executors of each of the multiple pending tasks are determined respectively, and each pending task is allocated to the determined task executors.

5. The method according to any one of claims 1 to 4, characterized in that: The task characteristics include any one or more of the following characteristics: task type, task magnitude; The executor characteristics include any one or more of the following characteristics: the broker's age, the broker's education level, the broker's length of service; The house source characteristics include any one or more of the following characteristics: house source price, house source location, and house source nature; the owner characteristics include any one or more of the following characteristics: whether the owner has changed houses, whether the owner has bought a house; The behavioral characteristics of the candidate task performers include any one or more of the following characteristics: the type and number of tasks performed by the agent; The change characteristics of the actor of the task to be processed include any one or more of the following characteristics: price changes of the house, and the number of times each task type is executed on the house.

6. The method according to any one of claims 1 to 4, characterized in that: The neural network model is also trained by the following steps: Acquire at least one training sample; wherein each training sample includes static features and temporal features corresponding to a task sample and a task executor sample, and the training sample has a label indicating whether the task sample is completed and a label indicating whether the associated target of the task sample is completed; Using a neural network model, based on the static features and time series features of the training samples, the comprehensive probability of the task performer sample corresponding to the training samples completing multiple goals is predicted to obtain a comprehensive probability prediction value; wherein the task performer sample completing multiple goals includes: completing the task sample and completing the associated goals of the task sample; Using a preset multi-objective loss function, based on whether the task sample of each training sample in the at least one training sample is completed, whether the associated target of the task sample is completed, and the comprehensive probability prediction value, the loss function value corresponding to each training sample is calculated; The neural network model is trained based on the loss function value corresponding to the at least one training sample.

7. The method according to claim 6, characterized in that The multi-objective loss function is: in, The loss function value, is the comprehensive probability prediction value, Is the label of whether the task sample is completed? is the label of whether the associated target of the task sample is completed, is the multi-objective loss weight, The value of is a real number greater than 0.

8. The method according to claim 7, characterized in that The task samples are task samples in the housing task system; the task types of the task samples include any one or more of the following categories: bargaining, showing, virtual reality VR showing, VR house explanation, voice follow-up, face-to-face interview, on-site inspection of the house, and getting the keys.

9. The method according to claim 8, characterized in that The value of C is determined based on the task type of the task sample: Among them, c is the preset fixed weight value, The value of is a real number greater than 1.

10. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to execute a computer program stored in the memory, and when the computer program is executed, implement the method described in any one of claims 1 to 9.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method described in any one of claims 1 to 9 is implemented.

12. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method described in any one of claims 1 to 9 is implemented.

Images