CN116258356B - Work order dispatching method and device based on WaveNet and related medium - Google Patents
Work order dispatching method and device based on WaveNet and related medium Download PDFInfo
- Publication number
- CN116258356B CN116258356B CN202310548729.1A CN202310548729A CN116258356B CN 116258356 B CN116258356 B CN 116258356B CN 202310548729 A CN202310548729 A CN 202310548729A CN 116258356 B CN116258356 B CN 116258356B
- Authority
- CN
- China
- Prior art keywords
- vector
- user portrait
- attribute
- vectors
- task feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 45
- 239000013598 vector Substances 0.000 claims abstract description 186
- 238000012545 processing Methods 0.000 claims abstract description 36
- 238000011176 pooling Methods 0.000 claims description 27
- 238000007781 pre-processing Methods 0.000 claims description 14
- 230000004913 activation Effects 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000002123 temporal effect Effects 0.000 claims description 9
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000003213 activating effect Effects 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 8
- 230000004927 fusion Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 238000010561 standard procedure Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Operations Research (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a worksheet dispatching method, a device and a related medium based on WaveNet, wherein the method comprises the following steps: acquiring work order data and a plurality of user data to be received, generating task feature attributes according to the work order data, and correspondingly generating a plurality of user portrait attributes according to the plurality of user data; respectively carrying out vectorization processing on the task feature attribute and each user portrait attribute to obtain a task feature vector and a plurality of corresponding initial user portrait vectors; generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels; and respectively calculating the similarity of the task feature vector and each target user portrait vector, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as dispatch objects. The embodiment of the application can improve the accuracy, the instantaneity and the success rate of the dispatching process.
Description
Technical Field
The application relates to the technical field of work order dispatch, in particular to a work order dispatch method and device based on WaveNet and a related medium.
Background
In recent years, along with the digital transformation of the property industry, intelligent worksheets are widely applied to multiple scenes of property management, provide unified report, periodic maintenance tasks, task allocation and scheduling and other integral solutions for the fields of customer assets, space and the like, simultaneously better realize the quantification of service results, and provide very important basis and guarantee for the statistics of the later worksheet completion rate and the calculation of payroll.
However, the present property work order distribution is still realized based on certain specific rules, but when special situations exist, such as employee departure, holiday, abnormal working, etc., the method of distributing the work order to a working group or organization and post at this time can reduce the accuracy, so that the work order may not be successfully distributed to the effective users, and the employee has too long time to rob the unit.
Disclosure of Invention
The embodiment of the application provides a work order dispatching method, a device, computer equipment and a storage medium based on WaveNet, aiming at improving the accuracy, the instantaneity and the success rate of work order dispatching.
In a first aspect, an embodiment of the present application provides a method for dispatching a work order based on WaveNet, including:
acquiring work order data and a plurality of user data to be received, generating task feature attributes according to the work order data, and correspondingly generating a plurality of user portrait attributes according to the plurality of user data;
respectively carrying out vectorization processing on the task feature attribute and each user portrait attribute to obtain a task feature vector and a plurality of corresponding initial user portrait vectors;
generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels;
and respectively calculating the similarity of the task feature vector and each target user portrait vector, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as dispatch objects, wherein N is more than or equal to 1.
In a second aspect, an embodiment of the present application provides a worksheet dispatching device based on WaveNet, including:
the data acquisition unit is used for acquiring the work order data and a plurality of user data to be received, generating task characteristic attributes according to the work order data and correspondingly generating a plurality of user portrait attributes according to the plurality of user data;
the vectorization processing unit is used for respectively vectorizing the task characteristic attribute and each user portrait attribute to obtain a task characteristic vector and a plurality of corresponding initial user portrait vectors;
the label generating unit is used for generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels;
and the user selection unit is used for calculating the similarity of the task feature vector and each target user portrait vector respectively, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as dispatch objects, wherein N is more than or equal to 1.
In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored on the memory and capable of running on the processor, where the processor implements the WaveNet-based work order dispatch method according to the first aspect when the processor executes the computer program.
In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, where a computer program is stored, where the computer program is executed by a processor to implement the method for dispatching a worksheet based on WaveNet according to the first aspect.
The embodiment of the application provides a worksheet dispatching method, a worksheet dispatching device, computer equipment and a storage medium based on WaveNet, wherein the method comprises the following steps: acquiring work order data and a plurality of user data to be received, generating task feature attributes according to the work order data, and correspondingly generating a plurality of user portrait attributes according to the plurality of user data; respectively carrying out vectorization processing on the task feature attribute and each user portrait attribute to obtain a task feature vector and a plurality of corresponding initial user portrait vectors; generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels; and respectively calculating the similarity of the task feature vector and each target user portrait vector, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as dispatch objects, wherein N is more than or equal to 1. The embodiment of the application firstly collects the user and the work order data, automatically generates the user portrait attribute and the task feature attribute, abstracts the user portrait attribute and the task feature attribute into vectors, then generates personalized labels for the user by utilizing an improved WaveNet model, calculates the users similar to the task feature by using cosine similarity, and selects a plurality of users with the top ranking as recommended users of the work order, thereby improving the accuracy, the real-time performance and the success rate in the property dispatch process.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a worksheet dispatching method based on a WaveNet according to an embodiment of the present application;
fig. 2 is a schematic sub-flowchart of a worksheet dispatching method based on WaveNet according to an embodiment of the present application;
FIG. 3 is a schematic block diagram of a WaveNet-based work order dispensing device according to an embodiment of the present application;
fig. 4 is a sub-schematic block diagram of a worksheet dispatching device based on WaveNet according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
Referring to fig. 1, an embodiment of the present application provides a worksheet dispatching method based on WaveNet, which specifically includes: steps S101 to S104.
S101, acquiring work order data and a plurality of user data to be received, generating task feature attributes according to the work order data, and correspondingly generating a plurality of user portrait attributes according to the plurality of user data;
in this step, after the work order is created, task feature engineering is invoked to generate corresponding task features for the work order, where the task feature attribute may specifically include information such as time, space, and content of the work order. The time attribute may include task creation time, task response time, planned starting time, planned ending time, actual starting time, actual ending time, standard man-hour, actual man-hour, order receiving time length, proceeding time length, acceptance time length, evaluation time length, etc.; spatial attributes may include spatial location, geographic location, coordinates, belonging properties, and the like; the job ticket content may include task type, specialty type, flow status, priority, whether to charge, number of SOPs (Standard Operating Procedure, i.e., standard job procedures), SOP classifications, number of steps, etc.
And when the user triggers the events such as order receiving, work starting or work order task completion, the user portrait engineering is called, and user portrait attributes are generated for the user. The user portrait attributes may include three major attributes of user portrait time, space, and specialty. Wherein the time attribute may include: the latest active time, the latest active time period, the latest work efficiency, the accumulated standard working hours of the work orders in progress on the same day, the accumulated standard working hours of the work orders in progress, the accumulated actual working hours of the work orders in progress, and the like; the spatial attributes may include the latest activity coordinates, the latest active property, the province at which the space is located, the geographic location at which the space is located, etc.; the specialty properties may include maneuver properties, specialty tags, specialty tag weight values, specialty class tags, specialty class tag weight values, and the like.
S102, respectively carrying out vectorization processing on the task feature attributes and each user portrait attribute to obtain task feature vectors and corresponding multiple initial user portrait vectors;
in this step, as shown in fig. 2, the step S102 includes: steps S201 to S204.
S201, performing time attribute vectorization representation on the task feature attributes according to the following formula:
；
wherein ,tj Representing a first temporal attribute vector, t c Representing the time of creating the work order, s represents the operation standard of the work order, and u represents the emergency degree;
s202, acquiring longitude and latitude coordinates of the task feature attributes, and converting the longitude and latitude coordinates into n-dimensional vectors by using a longitude and latitude distance calculation method to obtain space attribute vectorization representation of the task feature attributes:
；
wherein ,sj Representing a first spatial attribute vector, s 1 、s 2 、s n Representing the converted dimension value;
s203, carrying out content attribute vectorization representation on the task feature attribute by adopting a word vector model:
；
wherein ,cj Representing a first content attribute vector, c 1 、c 2 、c m Respectively representing dimension values converted by the word vector model;
s204, summarizing the first time attribute vector, the first space attribute vector and the first content attribute vector into the task feature vector。
In addition, the step S102 further includes:
and carrying out time attribute vectorization representation on the user portrait attributes according to the following steps:
；
wherein ,representing a second temporal attribute vector, "> and />Representing the earliest and latest timestamps or dates, respectively, in the dataset;
acquiring longitude and latitude coordinates of the user portrait attribute, and converting the longitude and latitude coordinates into two-dimensional coordinate vectors by using a longitude and latitude distance calculation method to obtain a space attribute vectorization representation of the task feature attribute:
；
wherein ,representing a second spatial attribute vector, "> and />Representing longitude and latitude, respectively;
and carrying out content attribute vectorization representation on the user portrait attributes by adopting a word vector model:
；
wherein ,representing a second content attribute vector->Representing a weight matrix, +.>Representing content +.>Mapping into word vectors;
summarizing the second temporal attribute vector, the second spatial attribute vector, and the second content attribute vector into the user portrait vector。
On the one hand, the time attribute, the space attribute and the content attribute in the task feature attribute are correspondingly converted into a first time attribute vector, a first space attribute vector and a first content attribute vector, so that the task feature vector is obtained in a summarizing mode; on the other hand, the user portrait vector is obtained by correspondingly converting the time attribute, the space attribute and the professional attribute in the user portrait attribute into a second time attribute vector, a second space attribute vector and a second content attribute vector.
S103, generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels;
the method specifically comprises the following steps:
performing data preprocessing on the initial user portrait vector, wherein the data preprocessing comprises data cleaning and feature extraction;
inputting the initial user portrait vector subjected to data preprocessing into a WaveNet model, and sequentially carrying out convolution processing and pooling processing on the initial user portrait vector by utilizing a convolution layer and a pooling layer in the WaveNet model;
carrying out information flow adjustment processing on the output result of the convolution layer by utilizing a residual error module;
and carrying out classification prediction on the output result of the pooling layer through a softmax function, and generating the personalized tag based on the classification prediction result.
Specifically, the information flow adjustment processing is performed on the output result of the convolution layer by using the residual error module, including:
and carrying out information flow regulation processing on the output result of the convolution layer by adopting a gate control activating unit in the residual block according to the following steps:
；
wherein z is the output result of the gating activation unit, tanh and sigmoid are a filtering gate and a learning gate respectively, x is the input data of the convolution layer, and w f and wg The weighting coefficients corresponding to x inside tanh and sigmoid are represented, respectively.
The WaveNet network model is based on a convolutional neural network, can perform parallel operation, and has high training and predicting speeds. In order to solve the problem of low multi-classification accuracy of the personalized labels of users, the embodiment applies a WaveNet model to the field of user portrait label generation. To extract the data features more fully, a fused pooling with a maximum, average pooling output parallel is used as the pooling layer of the model. And in the feature extraction stage, extracting deep features by a fusion pooling layer obtained by parallelly connecting the maximum pooling and the average pooling.
Specifically, the acquired user feature vector is first input to the WaveNet model, while the residual technique is applied to the model in order to optimize the use of the extended convolution. The residual block output is:
；
where x is the input feature of the expanded convolution (i.e., the user feature vector),is the output of the unit inside the residual block.
The residual block internal unit is a gating activation unit, and tan h and sigmoid are respectively used as a filtering gate and a learning gate and are used as an activation function. The output of the gating activation expansion convolution is:
；
wherein , and />Respectively indicate-> and />And the weight coefficient corresponding to x is internally calculated.
Stacking the residual blocks to obtain the following output:
；
when the pooling fusion and batch standardization treatment are carried out, the fusion pooling process is as follows:
；
h is the input of the upper network into the fusion pooling layer;output for maximum pooling; />Output for average pooling; />To connect the maximum pooling in parallel with the average pooling.
Batch inputs are:
；
calculating a batch processing data average value:
；
calculating batch data variance:
；
normalization processing:
；
dimensional change and offset:
；
wherein ,representing the output of the normalization process; />Representing the output of the BN transformation.
And then, calculating the probability of the classified result by using a softmax function, namely, sending the output of the model to the softmax function, so that a personalized label is generated for the user according to the classified result output by the softmax function, and a corresponding more suitable work order is recommended for the user according to the personalized label of the user. The softmax multi-class calculation formula is:
；
wherein ,output result as softmax function, < ->，/>,…,/>Is a parameter in the softmax function, < ->Normalization factor as a function of softmax.
S104, calculating the similarity of the task feature vector and each target user portrait vector, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as dispatch objects, wherein N is more than or equal to 1.
Before the step S103, the method includes:
and extracting common features from the task feature vectors by using a PCA dimension reduction method to obtain target task feature vectors.
In the PCA dimension reduction method, the commonality feature refers to the correlation or collinearity structure that exists in the original dataset. These commonalities can be interpreted as the direction in which the dominant pattern or variance present in the data is greatest. By applying PCA dimension reduction techniques, these common features in the original dataset can be extracted to obtain a smaller but still representative feature set, the features of which are called principal components. In this way, the data set can be simplified and help discover the underlying structure of the data.
The step S103 includes:
for each target user portrait vector, calculating the similarity between the target task feature vector and the target user portrait vector according to the following formula:
；
wherein ,representing a target user portrait vector u i And target task feature vector w j Similarity of->Representing the dot product of the vector, ">Representation vector->Is (are) mould>Representation vector->Is a mold of (a).
The similarity score between each user and the work order task can be obtained by calculating the similarity between the user portrait vector (the target user portrait vector in this step) and the task feature vector (the target task feature vector in this step) using cosine similarity. The similarity scores may then be ranked, e.g., from high to low or low to high, with the top N users being the recommended users for the worksheet.
Of course, in other embodiments, the similarity between the target task feature vector and the target user representation vector, such as Euclidean distance, etc., may also be calculated in a manner.
The embodiment firstly collects the user and the work order data, automatically generates the user portrait attribute and the task feature attribute, abstracts the user portrait attribute and the task feature attribute into vectors, then generates personalized labels for the user by utilizing an improved WaveNet model, calculates users similar to task feature matching by using cosine similarity, and selects users with a plurality of digits before ranking as recommended users of the work order, thereby improving the accuracy, the instantaneity and the success rate in the property dispatch process.
Fig. 3 is a schematic block diagram of a worksheet dispatching device 300 based on WaveNet according to an embodiment of the present application, where the device 300 includes:
a data obtaining unit 301, configured to obtain a worksheet data and a plurality of user data to be received, generate task feature attributes according to the worksheet data, and generate a plurality of user portrait attributes according to a plurality of user data;
vectorization processing unit 302, configured to perform vectorization processing on the task feature attribute and each user portrait attribute, to obtain a task feature vector and a corresponding plurality of initial user portrait vectors;
a tag generating unit 303, configured to generate a personalized tag for each of the initial user portrait vectors by using the improved WaveNet model, and generate a target user portrait vector by combining the initial user portrait vector and the corresponding personalized tag;
and a user selection unit 304, configured to calculate the similarity for the task feature vector and each of the target user portrait vectors, and select users corresponding to the first N target user portrait vectors with the highest similarity as the serving objects, where N is greater than or equal to 1.
In one embodiment, as shown in fig. 4, the vectorization processing unit 302 includes:
a first time representation unit 401, configured to perform a time attribute vectorization representation on the task feature attribute according to the following formula:
；
wherein ,tj Representing a first temporal attribute vector, t c Representing the time of creating the work order, s represents the operation standard of the work order, and u represents the emergency degree;
the first spatial representation unit 402 is configured to obtain latitude and longitude coordinates of the task feature attribute, and convert the latitude and longitude coordinates into n-dimensional vectors by using a latitude and longitude distance calculation method, so as to obtain a spatial attribute vectorized representation of the task feature attribute:
；
wherein ,sj Representing a first spatial attribute vector, s 1 、s 2 、s n Representing the converted dimension value;
a first content representation unit 403, configured to perform content attribute vectorization representation on the task feature attribute by using a word vector model:
；
wherein ,cj Representing a first content attribute vector, c 1 、c 2 、c m Respectively representing dimension values converted by the word vector model;
a first vector summarization unit 404 configured to summarize the first time attribute vector, the first space attribute vector, and the first content attribute vector into the task feature vector。
In an embodiment, the vectorization processing unit 302 further includes:
and the second time representation unit is used for carrying out time attribute vectorization representation on the user portrait attributes according to the following steps:
；
wherein ,representing a second temporal attribute vector, "> and />Representing the earliest and latest timestamps or dates, respectively, in the dataset;
the second spatial representation unit is used for acquiring longitude and latitude coordinates of the user portrait attribute, converting the longitude and latitude coordinates into two-dimensional coordinate vectors by using a longitude and latitude distance calculation method, and obtaining the spatial attribute vectorization representation of the task feature attribute:
；
wherein ,representing a second spatial attribute vector, "> and />Representing longitude and latitude, respectively;
the second content representation unit is used for carrying out content attribute vectorization representation on the user portrait attribute by adopting a word vector model:
；
wherein ,representing a second content attribute vector->Representing a weight matrix, +.>Representing content +.>Mapping into word vectors;
a second vector summarization unit for summarizing the second time attribute vector, the second space attribute vector and the second content attribute vector into the user portrait vector。
In an embodiment, the tag generation unit 303 includes:
the data preprocessing unit is used for carrying out data preprocessing on the initial user portrait vector, wherein the data preprocessing comprises data cleaning and feature extraction;
the convolution pooling unit is used for inputting the initial user portrait vector subjected to data preprocessing into a WaveNet model, and carrying out convolution processing and pooling processing on the initial user portrait vector by utilizing a convolution layer and a pooling layer in the WaveNet model in sequence;
the residual processing unit is used for carrying out information flow adjustment processing on the output result of the convolution layer by utilizing the residual module;
and the classification prediction unit is used for carrying out classification prediction on the output result of the pooling layer through a softmax function and generating the personalized tag based on the classification prediction result.
In an embodiment, the residual processing unit comprises:
the gating activation unit is used for carrying out information flow adjustment processing on the output result of the convolution layer by adopting the gating activation unit in the residual block according to the following formula:
；
wherein z is the output result of the gating activation unit, tanh and sigmoid are a filtering gate and a learning gate respectively, x is the input data of the convolution layer, and w f and wg The weighting coefficients corresponding to x inside tanh and sigmoid are represented, respectively.
In one embodiment, the user selection unit 304 includes:
the similarity calculation unit is used for calculating the similarity between the target task feature vector and the target user portrait vector according to the following formula for each target user portrait vector:
；
wherein ,representing a target user portrait vector u i And target task feature vector w j Similarity of->Representing the dot product of the vector, ">Representation vector->Is (are) mould>Representation vector->Is a mold of (a).
Since the embodiments of the apparatus portion and the embodiments of the method portion correspond to each other, the embodiments of the apparatus portion are referred to the description of the embodiments of the method portion, and are not repeated herein.
The embodiment of the present application also provides a computer readable storage medium having a computer program stored thereon, which when executed can implement the steps provided in the above embodiment. The storage medium may include: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The embodiment of the application also provides a computer device, which can comprise a memory and a processor, wherein the memory stores a computer program, and the processor can realize the steps provided by the embodiment when calling the computer program in the memory. Of course, the computer device may also include various network interfaces, power supplies, and the like.
In the description, each embodiment is described in a progressive manner, and each embodiment is mainly described by the differences from other embodiments, so that the same similar parts among the embodiments are mutually referred. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section. It should be noted that it will be apparent to those skilled in the art that various modifications and adaptations of the application can be made without departing from the principles of the application and these modifications and adaptations are intended to be within the scope of the application as defined in the following claims.
It should also be noted that in this specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Claims (8)
1. A worksheet dispatching method based on WaveNet is characterized by comprising the following steps:
acquiring work order data and a plurality of user data to be received, generating task feature attributes according to the work order data, and correspondingly generating a plurality of user portrait attributes according to the plurality of user data;
respectively carrying out vectorization processing on the task feature attribute and each user portrait attribute to obtain a task feature vector and a plurality of corresponding initial user portrait vectors;
generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels;
respectively calculating the similarity of the task feature vector and each target user portrait vector, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as dispatch objects, wherein N is more than or equal to 1;
the generating personalized labels for each initial user portrait vector by using the improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels comprises the following steps:
performing data preprocessing on the initial user portrait vector, wherein the data preprocessing comprises data cleaning and feature extraction;
inputting the initial user portrait vector subjected to data preprocessing into a WaveNet model, and sequentially carrying out convolution processing and pooling processing on the initial user portrait vector by utilizing a convolution layer and a pooling layer in the WaveNet model;
carrying out information flow adjustment processing on the output result of the convolution layer by utilizing a residual error module;
carrying out classified prediction on the output result of the pooling layer through a softmax function, and generating the personalized tag based on the classified prediction result;
the information flow adjustment processing is carried out on the output result of the convolution layer by utilizing the residual error module, and the information flow adjustment processing comprises the following steps:
and carrying out information flow regulation processing on the output result of the convolution layer by adopting a gate control activating unit in the residual block according to the following steps:
；
wherein z is the output result of the gating activation unit, tanh and sigmoid are a filtering gate and a learning gate respectively, x is the input data of the convolution layer, and w f and wg The weighting coefficients corresponding to x inside tanh and sigmoid are represented, respectively.
2. The method for dispatching a worksheet based on WaveNet according to claim 1, wherein said vectorizing the task feature attribute and each user portrait attribute to obtain a task feature vector and a corresponding plurality of initial user portrait vectors, respectively, comprises:
and carrying out time attribute vectorization representation on the task feature attributes according to the following steps:
；
wherein ,tj Representing a first temporal attribute vector, t c Representing the time of creating the work order, s represents the operation standard of the work order, and u represents the emergency degree;
acquiring longitude and latitude coordinates of the task feature attributes, and converting the longitude and latitude coordinates into n-dimensional vectors by using a longitude and latitude distance calculation method to obtain space attribute vectorization representation of the task feature attributes:
；
wherein ,sj Representing a first spatial attribute vector, s 1 、s 2 、s n Representing the converted dimension value;
and carrying out content attribute vectorization representation on the task feature attribute by adopting a word vector model:
；
wherein ,cj Representing a first content attribute vector, c 1 、c 2 、c m Respectively representing dimension values converted by the word vector model;
summarizing the first temporal attribute vector, the first spatial attribute vector, and the first content attribute vector into the task feature vector。
3. The method for dispatching a worksheet based on WaveNet according to claim 2, wherein said vectorizing the task feature attribute and each user portrait attribute to obtain a task feature vector and a corresponding plurality of initial user portrait vectors, respectively, further comprises:
and carrying out time attribute vectorization representation on the user portrait attributes according to the following steps:
；
wherein ,representing a second temporal attribute vector, "> and />Representing the earliest and latest time stamps or days, respectively, in a datasetA period;
acquiring longitude and latitude coordinates of the user portrait attribute, and converting the longitude and latitude coordinates into two-dimensional coordinate vectors by using a longitude and latitude distance calculation method to obtain a space attribute vectorization representation of the task feature attribute:
；
wherein ,representing a second spatial attribute vector, "> and />Representing longitude and latitude, respectively;
and carrying out content attribute vectorization representation on the user portrait attributes by adopting a word vector model:
；
wherein ,representing a second content attribute vector->Representing a weight matrix, +.>Representing content +.>Mapping into word vectors;
summarizing the second temporal attribute vector, the second spatial attribute vector, and the second content attribute vector into the user portrait vector。
4. The method for assigning a worksheet based on WaveNet as set forth in claim 1, wherein before said step of calculating a similarity for each of said task feature vector and each of said target user portrayal vectors, respectively, comprises:
and extracting common features from the task feature vectors by using a PCA dimension reduction method to obtain target task feature vectors.
5. The WaveNet-based work order dispatch method of claim 4, wherein said computing similarity for each of said task feature vector and each of said target user representation vectors, respectively, comprises:
for each target user portrait vector, calculating the similarity between the target task feature vector and the target user portrait vector according to the following formula:
；
wherein ,representing a target user portrait vector u i And target task feature vector w j Similarity of->Representing the dot product of the vector, ">Representation vector->Is (are) mould>Representation vector->Is a mold of (a).
6. A WaveNet-based work order dispatch device, comprising:
the data acquisition unit is used for acquiring the work order data and a plurality of user data to be received, generating task characteristic attributes according to the work order data and correspondingly generating a plurality of user portrait attributes according to the plurality of user data;
the vectorization processing unit is used for respectively vectorizing the task characteristic attribute and each user portrait attribute to obtain a task characteristic vector and a plurality of corresponding initial user portrait vectors;
the label generating unit is used for generating personalized labels for each initial user portrait vector by utilizing an improved WaveNet model, and generating target user portrait vectors by combining the initial user portrait vectors and the corresponding personalized labels;
the user selection unit is used for calculating the similarity of the task feature vector and each target user portrait vector respectively, and selecting users corresponding to the first N target user portrait vectors with the highest similarity as a dispatch object, wherein N is more than or equal to 1;
the tag generation unit includes:
the data preprocessing unit is used for carrying out data preprocessing on the initial user portrait vector, wherein the data preprocessing comprises data cleaning and feature extraction;
the convolution pooling unit is used for inputting the initial user portrait vector subjected to data preprocessing into a WaveNet model, and carrying out convolution processing and pooling processing on the initial user portrait vector by utilizing a convolution layer and a pooling layer in the WaveNet model in sequence;
the residual processing unit is used for carrying out information flow adjustment processing on the output result of the convolution layer by utilizing the residual module;
the classification prediction unit is used for carrying out classification prediction on the output result of the pooling layer through a softmax function and generating the personalized tag based on the classification prediction result;
the residual processing unit includes:
the gating activation unit is used for carrying out information flow adjustment processing on the output result of the convolution layer by adopting the gating activation unit in the residual block according to the following formula:
；
wherein z is the output result of the gating activation unit, tanh and sigmoid are a filtering gate and a learning gate respectively, x is the input data of the convolution layer, and w f and wg The weighting coefficients corresponding to x inside tanh and sigmoid are represented, respectively.
7. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the WaveNet-based work order dispatch method of any one of claims 1 to 5 when the computer program is executed.
8. A computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when executed by a processor, the computer program implements the WaveNet-based work order dispatch method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310548729.1A CN116258356B (en) | 2023-05-16 | 2023-05-16 | Work order dispatching method and device based on WaveNet and related medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310548729.1A CN116258356B (en) | 2023-05-16 | 2023-05-16 | Work order dispatching method and device based on WaveNet and related medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116258356A CN116258356A (en) | 2023-06-13 |
| CN116258356B true CN116258356B (en) | 2023-08-25 |
Family
ID=86686562
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310548729.1A Active CN116258356B (en) | 2023-05-16 | 2023-05-16 | Work order dispatching method and device based on WaveNet and related medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116258356B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116882681A (en) * | 2023-07-11 | 2023-10-13 | 武汉市万睿数字运营有限公司 | Multi-target scheduling method, device and equipment for property work orders and storage medium |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105824912A (en) * | 2016-03-15 | 2016-08-03 | 平安科技（深圳）有限公司 | Personalized recommending method and device based on user portrait |
| CN109003027A (en) * | 2018-07-16 | 2018-12-14 | 江苏满运软件科技有限公司 | A kind of management method and system of label of drawing a portrait |
| CN109840287A (en) * | 2019-01-31 | 2019-06-04 | 中科人工智能创新技术研究院(青岛)有限公司 | A kind of cross-module state information retrieval method neural network based and device |
| CN110832596A (en) * | 2017-10-16 | 2020-02-21 | 因美纳有限公司 | Deep convolutional neural network training method based on deep learning |
| CN111442778A (en) * | 2019-01-16 | 2020-07-24 | 百度在线网络技术（北京）有限公司 | Travel scheme recommendation method, device and equipment and computer readable storage medium |
| CN111737516A (en) * | 2019-12-23 | 2020-10-02 | 北京沃东天骏信息技术有限公司 | Interactive music generation method and device, intelligent sound box and storage medium |
| CN111881296A (en) * | 2020-07-31 | 2020-11-03 | 深圳市万物云科技有限公司 | Work order processing method based on community scene and related components |
| CN114332778A (en) * | 2022-03-08 | 2022-04-12 | 深圳市万物云科技有限公司 | Intelligent alarm work order generation method and device based on people stream density and related medium |
| CN114418641A (en) * | 2022-01-20 | 2022-04-29 | 四川众信佳科技发展有限公司 | Automatic dispatching method and device for stock customer access work order and electronic equipment |
| CN115048174A (en) * | 2022-06-22 | 2022-09-13 | 平安科技（深圳）有限公司 | User-based personalized menu generation method, device, equipment and storage medium |
| CN115100739A (en) * | 2022-06-09 | 2022-09-23 | 厦门国际银行股份有限公司 | Man-machine behavior detection method, system, terminal device and storage medium |
| CN115187005A (en) * | 2022-06-23 | 2022-10-14 | 中国人民公安大学 | Scheduling method, device, equipment and storage medium |
| CN115809796A (en) * | 2023-02-06 | 2023-03-17 | 佰聆数据股份有限公司 | Project intelligent dispatching method and system based on user portrait |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11308439B2 (en) * | 2019-01-22 | 2022-04-19 | Everseen Limited | Goods receipt management system and method |
-
2023
- 2023-05-16 CN CN202310548729.1A patent/CN116258356B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105824912A (en) * | 2016-03-15 | 2016-08-03 | 平安科技（深圳）有限公司 | Personalized recommending method and device based on user portrait |
| CN110832596A (en) * | 2017-10-16 | 2020-02-21 | 因美纳有限公司 | Deep convolutional neural network training method based on deep learning |
| CN109003027A (en) * | 2018-07-16 | 2018-12-14 | 江苏满运软件科技有限公司 | A kind of management method and system of label of drawing a portrait |
| CN111442778A (en) * | 2019-01-16 | 2020-07-24 | 百度在线网络技术（北京）有限公司 | Travel scheme recommendation method, device and equipment and computer readable storage medium |
| CN109840287A (en) * | 2019-01-31 | 2019-06-04 | 中科人工智能创新技术研究院(青岛)有限公司 | A kind of cross-module state information retrieval method neural network based and device |
| CN111737516A (en) * | 2019-12-23 | 2020-10-02 | 北京沃东天骏信息技术有限公司 | Interactive music generation method and device, intelligent sound box and storage medium |
| CN111881296A (en) * | 2020-07-31 | 2020-11-03 | 深圳市万物云科技有限公司 | Work order processing method based on community scene and related components |
| CN114418641A (en) * | 2022-01-20 | 2022-04-29 | 四川众信佳科技发展有限公司 | Automatic dispatching method and device for stock customer access work order and electronic equipment |
| CN114332778A (en) * | 2022-03-08 | 2022-04-12 | 深圳市万物云科技有限公司 | Intelligent alarm work order generation method and device based on people stream density and related medium |
| CN115100739A (en) * | 2022-06-09 | 2022-09-23 | 厦门国际银行股份有限公司 | Man-machine behavior detection method, system, terminal device and storage medium |
| CN115048174A (en) * | 2022-06-22 | 2022-09-13 | 平安科技（深圳）有限公司 | User-based personalized menu generation method, device, equipment and storage medium |
| CN115187005A (en) * | 2022-06-23 | 2022-10-14 | 中国人民公安大学 | Scheduling method, device, equipment and storage medium |
| CN115809796A (en) * | 2023-02-06 | 2023-03-17 | 佰聆数据股份有限公司 | Project intelligent dispatching method and system based on user portrait |
Non-Patent Citations (1)
| Title |
|---|
| Yu YB，et al.A sample-level DCNN for music auto-tagging.《multimedia tools and applications》.2021,第80卷(第8期),11459-11469. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116258356A (en) | 2023-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10936947B1 (en) | Recurrent neural network-based artificial intelligence system for time series predictions | |
| CN107967575B (en) | Artificial intelligence platform system for artificial intelligence insurance consultation service | |
| CN116258356B (en) | Work order dispatching method and device based on WaveNet and related medium | |
| CN113255842B (en) | Vehicle replacement prediction method, device, equipment and storage medium | |
| WO2019214455A1 (en) | Data sequence prediction method and computing apparatus | |
| Wu et al. | A PCA-AdaBoost model for E-commerce customer churn prediction | |
| CN111310462A (en) | User attribute determination method, device, equipment and storage medium | |
| Kedia et al. | Portfolio generation for Indian stock markets using unsupervised machine learning | |
| CN116957304B (en) | Unmanned aerial vehicle group collaborative task allocation method and system | |
| Shen et al. | Equiangular basis vectors | |
| CN116703129B (en) | Intelligent task matching scheduling method and system based on personnel data image | |
| CN117456267A (en) | Class increment learning method based on similarity prototype playback | |
| CN116957260A (en) | Intelligent allocation method for carrier vehicles based on vehicle demand prediction | |
| CN115797795A (en) | Remote sensing image question-answering type retrieval system and method based on reinforcement learning | |
| CN115034837A (en) | Product sales prediction method, equipment and medium based on knowledge graph | |
| Potočnik et al. | Self-organizing neural network-based clustering and organization of production cells | |
| CN117194966A (en) | Training method and related device for object classification model | |
| CN113571198A (en) | Conversion rate prediction method, device, equipment and storage medium | |
| JP2022182628A (en) | Information processing device, information processing method, information processing program, and learning model generation device | |
| Liu et al. | Scalable learning for dispersed knowledge systems | |
| KR20200112495A (en) | Server and method for providing commercial analysis services | |
| CN111582408A (en) | Data processing method, data processing apparatus, storage medium, and electronic device | |
| Motzev et al. | Self-organizing data mining techniques in model based simulation games for business training and education | |
| CN117670066B (en) | Questor management method, system, equipment and storage medium based on intelligent decision | |
| CN113469491B (en) | Flexible workshop operation scheduling method based on reinforcement learning and graph neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |