CN112801144B

CN112801144B - Resource allocation method, device, computer equipment and storage medium

Info

Publication number: CN112801144B
Application number: CN202110037824.6A
Authority: CN
Inventors: 陈弘; 牛犇; 张莉; 吴志成
Original assignee: Ping An Technology Shenzhen Co Ltd
Current assignee: Ping An Technology Shenzhen Co Ltd
Priority date: 2021-01-12
Filing date: 2021-01-12
Publication date: 2021-09-28
Anticipated expiration: 2041-01-12
Also published as: CN112801144A

Abstract

The invention relates to the technical field of artificial intelligence, and provides a resource allocation method, a resource allocation device, computer equipment and a storage medium, wherein the resource allocation method comprises the following steps: carrying out feature classification on data in the data tables of the multiple target objects by using a feature classification model to obtain multiple feature classes corresponding to each data table; constructing a characteristic category data matrix for the corresponding target object according to the plurality of characteristic categories corresponding to each data table; predicting to obtain a plurality of evaluation categories corresponding to each target object and an evaluation probability corresponding to each evaluation category based on the characteristic category data matrix by using a multi-task multi-output prediction model; generating a characteristic evaluation probability matrix according to the plurality of characteristic categories of each target object and the evaluation probability corresponding to each evaluation category; clustering the target objects according to the characteristic evaluation probability matrixes to obtain a plurality of target object clusters; and allocating resources for each target object according to the plurality of target object clusters. The invention can solve the problem of resource distribution imbalance and has high resource distribution efficiency.

Description

Resource allocation method, device, computer equipment and storage medium

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a resource allocation method, a resource allocation device, computer equipment and a storage medium.

Background

With the acceleration of enterprise information construction, enterprises are aware of the importance of resources more and more, and less resources are allocated to business units with poor performance, so that the resources of the enterprises can be saved, redundant resources are allocated to business units with excellent performance, and more opportunities are provided for the business units with excellent performance.

In the conventional resource allocation system, resources are allocated based on performance data of business units by training a machine learning model using the machine learning model. Thus, some business units with strong comprehensive service capability but low performance data are allocated with less resources, so that the accuracy of resource allocation is low; further, based solely on performance data, a resource allocation imbalance problem may also arise.

Disclosure of Invention

In view of the above, it is necessary to provide a resource allocation method, device, computer device and storage medium, which can allocate resources to a plurality of target objects in batch and have high resource allocation efficiency; and the same resources are distributed to the target objects in the same target object cluster through clustering, so that the problem of unbalanced resource distribution is solved.

A first aspect of the present invention provides a resource allocation method, including:

carrying out feature classification on data in the data tables of the multiple target objects by using a feature classification model to obtain multiple feature classes corresponding to each data table;

constructing a characteristic category data matrix for the corresponding target object according to the plurality of characteristic categories corresponding to each data table;

predicting by using a multi-task multi-output prediction model based on the characteristic category data matrix of each target object to obtain a plurality of evaluation categories corresponding to each target object and evaluation probabilities corresponding to each evaluation category;

generating a characteristic evaluation probability matrix according to the plurality of characteristic categories of each target object and the evaluation probability corresponding to each evaluation category;

clustering the target objects according to the characteristic evaluation probability matrixes to obtain a plurality of target object clusters;

and allocating resources to each target object according to the plurality of target object clusters.

In an optional embodiment, the constructing a feature class data matrix for the corresponding target object according to the plurality of feature classes corresponding to each data table includes:

for each target object, acquiring a plurality of data corresponding to the same characteristic category;

splicing a plurality of data corresponding to the same characteristic category to obtain a characteristic category data vector;

constructing a first feature category data matrix according to feature category data vectors corresponding to a plurality of feature categories;

and respectively carrying out transverse normalization processing and longitudinal normalization processing on the first characteristic category data matrix to obtain a second characteristic category data matrix.

In an optional embodiment, after obtaining the second feature class data matrix, the method further includes:

calculating the weight of each data index in each feature category data vector in the second feature category data matrix;

updating the corresponding feature category data vector according to the weight to obtain a first feature category data vector;

extracting a plurality of cross feature data based on data in the first feature class data vector;

calculating a correlation between each cross feature data and the data of each data index in the corresponding first feature class data vector;

screening out a plurality of target cross feature data according to the correlation;

updating the corresponding first feature type data vector based on the plurality of target cross feature data to obtain a second feature type data vector;

a target feature class data matrix is generated based on the plurality of second feature class data vectors for each target object.

In an optional embodiment, the generating a feature evaluation probability matrix according to the plurality of feature classes of each target object and the evaluation probability corresponding to each evaluation class includes:

for each feature category, acquiring the maximum evaluation probability in the corresponding multiple evaluation probabilities;

setting the evaluation probabilities except the maximum evaluation probability in the plurality of evaluation probabilities as preset values;

and generating a characteristic evaluation probability matrix according to the maximum evaluation probability corresponding to each characteristic category and the preset values.

In an optional embodiment, the clustering the target objects according to the feature evaluation probability matrices to obtain target object clusters includes:

clustering the characteristic evaluation probability matrixes to obtain a plurality of characteristic evaluation probability matrix clusters;

and determining the target object corresponding to the characteristic evaluation probability matrix in each characteristic evaluation probability matrix cluster to obtain a target object cluster.

In an optional embodiment, the allocating resources to each target object according to the plurality of target object clusters includes:

calculating an average evaluation probability matrix of the evaluation probability matrix cluster corresponding to each target object cluster;

calculating an evaluation score according to the average evaluation probability matrix;

performing reverse ordering on the target object clusters according to the evaluation scores;

and allocating corresponding resources for the target objects in the multiple target object clusters after the reverse sequencing, wherein the same resources are allocated for the target objects in the same target object cluster.

In an optional embodiment, said calculating a rating score according to said average rating probability matrix comprises:

and calculating the weighted average sum between the average evaluation probability of each target object cluster and the feature class weight of the corresponding feature class to obtain the evaluation score of each target object cluster.

A second aspect of the present invention provides a resource allocation apparatus, the apparatus comprising:

the classification module is used for performing feature classification on data in the data tables of the multiple target objects by using the feature classification model to obtain multiple feature classes corresponding to each data table;

the construction module is used for constructing a characteristic category data matrix for the corresponding target object according to the plurality of characteristic categories corresponding to each data table;

the prediction module is used for predicting based on the characteristic category data matrix of each target object by using a multitask and multi-output prediction model to obtain a plurality of evaluation categories corresponding to each target object and evaluation probability corresponding to each evaluation category;

the generating module is used for generating a characteristic evaluation probability matrix according to the plurality of characteristic categories of each target object and the evaluation probability corresponding to each evaluation category;

the clustering module is used for clustering the target objects according to the characteristic evaluation probability matrixes to obtain a plurality of target object clusters;

and the allocation module is used for allocating resources to each target object according to the plurality of target object clusters.

A third aspect of the invention provides a computer apparatus comprising a processor for implementing the resource allocation method when executing a computer program stored in a memory.

A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the resource allocation method.

In summary, the resource allocation method, apparatus, computer device and storage medium according to the present invention perform feature classification on data in data tables of multiple target objects by using a feature classification model to obtain multiple feature classes corresponding to each data table, and implement accurate subdivision of feature classes of data in data tables of multiple target objects, so as to construct a feature class data matrix for a corresponding target object according to the multiple feature classes corresponding to each data table, and construct a feature class data matrix, considering an association relationship between data of multiple feature classes of a target enterprise, when performing prediction based on the feature class data matrix of each target object by using a multi-task multi-output prediction model, multiple evaluation classes corresponding to each target object and an evaluation probability corresponding to each evaluation class can be obtained, because the data in the data tables include data of multiple feature dimensions, the evaluation probability of the target object is calculated by comprehensively measuring data of a plurality of characteristic dimensions, so that the evaluation probability of the target object is high in accuracy, a characteristic evaluation probability matrix is generated according to a plurality of characteristic categories of each target object and the evaluation probability corresponding to each evaluation category, the target objects are clustered according to the characteristic evaluation probability matrices, namely, the target objects are clustered, and finally resources are distributed to the target objects according to the target object clusters. The invention can carry out resource allocation on a plurality of target objects in batches, and the resource allocation efficiency is high; the data of a plurality of characteristic categories are comprehensively measured, the evaluation probability is calculated based on the data of each characteristic category, the target object is more comprehensively evaluated, and finally resources are allocated to the target object based on the evaluation probabilities of all the characteristic categories, so that the resource allocation accuracy is higher; the same resources are distributed to the target objects in the same target object cluster through clustering, and the problem of resource distribution imbalance is solved.

Drawings

Fig. 1 is a flowchart of a resource allocation method according to an embodiment of the present invention.

Fig. 2 is a structural diagram of a resource allocation apparatus according to a second embodiment of the present invention.

Fig. 3 is a schematic structural diagram of a computer device according to a third embodiment of the present invention.

Detailed Description

In order that the above objects, features and advantages of the present invention can be more clearly understood, a detailed description of the present invention will be given below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and features of the embodiments may be combined with each other without conflict.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

The resource allocation method provided by the embodiment of the invention is executed by computer equipment, and correspondingly, the resource allocation device runs in the computer equipment.

Fig. 1 is a flowchart of a resource allocation method according to an embodiment of the present invention. The resource allocation method specifically includes the following steps, and the order of the steps in the flowchart may be changed and some may be omitted according to different requirements.

And S11, performing feature classification on the data in the data tables of the target objects by using the feature classification model to obtain a plurality of feature categories corresponding to each data table.

The target object may be an enterprise, a business unit in the enterprise, or a team in the business unit. For example, when the target object is a business, the evaluation levels of the multiple businesses in the next period are predicted according to the data tables of the multiple businesses in the current period. And when the target object is a business unit, predicting the evaluation levels of the multiple enterprises in the next period according to the data table of the multiple business units in the current period.

Wherein, the data table comprises a plurality of data in the current period. The period can be monthly, quarterly or yearly.

The feature classification model is a pre-trained model for performing feature classification on a plurality of data of a target object. The computer equipment can acquire a plurality of historical data aiming at each feature type, then a marking tool is used for marking each historical data with a corresponding feature type, a historical data pair is constructed based on each historical data and the corresponding feature type, and a plurality of historical data pairs are used as training data to train the deep neural network to obtain a feature classification model. The feature categories may include, but are not limited to: performance categories, behavior categories, and quality categories.

And calling the feature classification model by the computer equipment to perform feature classification on each data in the data table of each target object to obtain the feature class of each data, and further obtaining a plurality of feature classes corresponding to each target object. Illustratively, assume that the data table of the target object includes: after the characteristic classification is carried out by using the characteristic classification model, the customer satisfaction data and the number of the increased members data are divided into quality categories, the number of the arrived meetings data and the number of the invited members data are divided into behavior data, and the increase rate data and the renewal rate data are divided into performance categories.

And S12, constructing a feature type data matrix for the corresponding target object according to the plurality of feature types corresponding to each data table.

Each target object corresponds to one data table, each data table corresponds to a plurality of feature types, each feature type corresponds to a plurality of data, a plurality of feature type data vectors are constructed based on the plurality of feature types corresponding to the target object and the plurality of data corresponding to each feature type, and evaluation type prediction can be conveniently carried out subsequently based on the plurality of feature type data vectors of each target object.

Because the feature class data vectors of different feature classes have large differences, in order to facilitate the subsequent rapid prediction by using a multi-task multi-output prediction model, each line in the feature class data matrix is normalized in the transverse direction. And the difference between different data is large, and in order to further accelerate the prediction of the multi-task and multi-output prediction model, each column in the characteristic class data matrix is normalized in the longitudinal direction.

In this optional embodiment, the first feature type data matrix is subjected to transverse normalization processing and longitudinal normalization processing, so as to obtain a second feature type data matrix, which has uniform dimensions in both the transverse direction and the longitudinal direction, thereby improving the prediction efficiency of the multi-task multi-output prediction model.

The XGBOOST model may be used to compute a weight for each data index in the feature class data vector. Each feature type data vector comprises data of a plurality of data indexes, the data of each data index is constructed into an index data vector, a plurality of index data vectors corresponding to each feature type data vector are input into the XGB OST model for training, the XGB OST model can be used for calculating the weights of the plurality of index data vectors, and therefore the weights of the plurality of data indexes are obtained. The larger the weight of the data index is, the more forward the corresponding data plays a role in the model, and the prediction accuracy of the model can be improved. The smaller the weight of the data index is, the more negative the corresponding data plays a role in the model, and the prediction accuracy of the model is reduced.

The data corresponding to the data indexes with the weight larger than the preset weight threshold value in each feature category data vector can be obtained, or the weights are sorted in a reverse order, and the data of the data indexes corresponding to the front preset K weights in the weights after the reverse sorting are obtained, so that the data corresponding to the data indexes with the weight smaller than or equal to the preset weight threshold value in each feature category data vector are removed, and the updating of each feature category data vector is realized.

An attention factor decomposition (AFM) or a multi-layer perceptron may be used to extract a plurality of cross feature data from the data in the first feature class data vector, and the extracted cross feature data includes both the cross combination between the low-order features and the high-order features, so that the extracted cross feature data has more comprehensive feature representation, thereby avoiding the loss of features, and thus improving the accuracy of the evaluation class prediction when the evaluation class prediction is performed based on the cross feature data, thereby improving the prediction accuracy of the integrated prediction model.

Since not all of the plurality of cross feature data are data that is advantageous for model prediction, the correlation between each two cross feature data is calculated by calculating a correlation coefficient between each two cross feature data. And the cross feature data with high correlation is reserved, and the cross feature data with low correlation is removed, so that the first feature class data vector is updated. Since the updated cross feature data in the second feature category data vectors are all data beneficial to improving the model prediction accuracy, a target feature category data matrix is generated according to the plurality of second feature category data vectors of each target object, and the evaluation category prediction is performed based on the target feature category data matrix, so that the accuracy of the evaluation category prediction can be further improved.

And S13, performing prediction based on the characteristic type data matrix of each target object by using the multitask and multi-output prediction model to obtain a plurality of evaluation types corresponding to each target object and an evaluation probability corresponding to each evaluation type.

The multi-task and multi-output prediction model is a plurality of pre-trained evaluation categories used for outputting each target object and evaluation probabilities corresponding to the evaluation categories.

The Bi-LSTM neural network framework may be used to train a multi-task multi-output prediction model, for example, obtaining historical data corresponding to each task, inputting the historical data corresponding to each task into the Bi-LSTM neural network framework for iterative training, obtaining predicted task identifiers of a plurality of tasks output by each iterative training, calculating residuals between the predicted task identifiers of the plurality of tasks and actual task identifiers of the plurality of tasks, and achieving a training goal by continuously reducing residuals generated by the iterative training. The task can be a quality task, a performance task and a behavior task, wherein the quality task can correspond to four good task identifiers, the performance task can correspond to four good task identifiers, and the behavior task can correspond to four good task identifiers.

And inputting the multiple feature type data vectors of each target object into the multi-task multi-output prediction model for prediction to obtain multiple evaluation types corresponding to each feature type of each target object and evaluation probability corresponding to each evaluation type. Illustratively, inputting a quality feature category data vector, a performance feature category data vector and a behavior feature category data vector of a target object into the multi-task multi-output prediction model, and outputting through the multi-task multi-output prediction model: the evaluation probabilities of the good mid-to-poor evaluation category and the good mid-to-poor evaluation category corresponding to the quality feature category, the evaluation probabilities of the good mid-to-poor evaluation category and the good mid-to-poor evaluation category corresponding to the performance feature category, and the evaluation probabilities of the good mid-to-poor evaluation category and the good mid-to-poor evaluation category corresponding to the behavior feature category.

S14, a feature evaluation probability matrix is generated based on the plurality of feature types for each target object and the evaluation probability corresponding to each evaluation type.

And constructing a plurality of feature types of the target object and corresponding feature type evaluation probabilities into a feature evaluation probability matrix.

Each row in the characteristic evaluation probability matrix corresponds to one characteristic category, each row element value corresponds to the evaluation category probability under the characteristic category, each column in the characteristic evaluation probability matrix corresponds to one evaluation category, and each column element value corresponds to the evaluation category probability of different characteristic categories under the evaluation categories.

Wherein the preset value may be 0.

Exemplary, as shown in the following table:

	P1	P2	P3	P4
					T1	X1
T2		X2
					T3		X3

wherein Ti represents the ith characteristic category, and Pi represents the ith evaluation category; x1 represents the maximum rating probability of the 1 st feature class, the maximum rating probability corresponding to the rating class P1; x2 represents the maximum rating probability of the 2 nd feature class, the maximum rating probability corresponding to the rating class P2; x3 represents the maximum rating probability for the 3 rd feature class, which corresponds to the rating class P4.

In this optional embodiment, by reserving the maximum evaluation probability under each feature category and setting the corresponding other evaluation probabilities to 0, the feature evaluation probability matrix of each target object can be simplified, so that when the plurality of target objects are clustered according to the plurality of feature evaluation probability matrices, a plurality of target object clusters can be quickly obtained, and the efficiency of calculating the evaluation level of the target object is further improved.

And S15, clustering the target objects according to the feature evaluation probability matrixes to obtain a plurality of target object clusters.

After the evaluation probability matrix is simplified, the target objects can be clustered according to the feature evaluation probability matrices.

The multiple feature evaluation probability matrixes can be subjected to clustering analysis by using a K-means clustering algorithm to obtain multiple feature evaluation probability matrix clusters, and each feature evaluation probability matrix cluster comprises a plurality of feature evaluation probability matrixes.

Each feature evaluation probability matrix corresponds to one target object, and the corresponding target object is determined according to the feature evaluation probability matrix in the feature evaluation probability matrix cluster, so that the target objects are clustered, and the evaluation level of each target object is determined according to the clustered target object cluster.

In this optional embodiment, the accurate clustering of the plurality of target objects is realized by clustering the plurality of feature evaluation probability matrices, the target objects with the same evaluation level are grouped into the same class, and the users with different evaluation levels are grouped into different classes.

S16, allocating resources for each target object according to the target object clusters.

And calculating the evaluation level of each target object cluster, and determining the evaluation level of the target object cluster as the evaluation level of all target objects in the target object cluster.

and distributing corresponding resources for the target objects in the plurality of target object clusters after the reverse sequencing.

Each target object cluster corresponds to one evaluation probability matrix cluster, each evaluation probability matrix cluster comprises a plurality of evaluation probability matrixes, the evaluation probability matrixes are the same in size, element values of the same positions in the evaluation probability matrixes are accumulated, and the element values are divided by the number of the evaluation probability matrixes to obtain the average evaluation probability of the corresponding positions, so that the average evaluation probability matrix is obtained.

After the average evaluation probability matrix of each target object cluster is calculated, according to the preset feature class weight, calculating the weighted average sum between the average evaluation probability and the feature class weight of the corresponding feature class to obtain the evaluation score of each target object cluster.

The higher the evaluation score is, the stronger the comprehensive service capability of the corresponding target object cluster is indicated, and the lower the evaluation score is, the worse the comprehensive service capability of the corresponding target object cluster is indicated. And performing reverse ordering on the evaluation scores, thereby realizing the reverse ordering of the target object clusters. After the reverse ordering, the stronger the comprehensive service capability of the target object cluster arranged in front, the more resources are allocated, and the worse the comprehensive service capability of the target object cluster arranged in back, the less resources are allocated. And target objects in the same target object cluster are allocated with the same resource.

In summary, the data in the data tables of multiple target objects are subjected to feature classification by using the feature classification model to obtain multiple feature classes corresponding to each data table, so as to realize accurate subdivision of the feature classes of the data in the data tables of multiple target objects, so that a feature class data matrix is constructed for the corresponding target object according to the multiple feature classes corresponding to each data table, the constructed feature class data matrix takes into account the incidence relation among the data of the multiple feature classes of the target enterprise, when a multi-task multi-output prediction model is used for prediction based on the feature class data matrix of each target object, multiple evaluation classes corresponding to each target object and the evaluation probability corresponding to each evaluation class can be obtained, and since the data in the data tables comprises the data of multiple feature dimensions, that is, the data of multiple feature dimensions are comprehensively measured to calculate the evaluation probability of the target object, therefore, the evaluation probability accuracy of the target objects is high, the feature evaluation probability matrix is generated according to the plurality of feature types of each target object and the evaluation probability corresponding to each evaluation type, the plurality of target objects are clustered according to the plurality of feature evaluation probability matrices, namely, the clustering of the plurality of target objects is realized, and finally, resources are distributed to each target object according to the plurality of target object clusters.

The invention can carry out resource allocation on a plurality of target objects in batches, and the resource allocation efficiency is high; the data of a plurality of characteristic categories are comprehensively measured, the evaluation probability is calculated based on the data of each characteristic category, the target object is more comprehensively evaluated, and finally resources are allocated to the target object based on the evaluation probabilities of all the characteristic categories, so that the resource allocation accuracy is higher; the same resources are distributed to the target objects in the same target object cluster through clustering, and the problem of resource distribution imbalance is solved.

It is emphasized that the plurality of data tables may be stored in the nodes of the blockchain in order to further ensure privacy and security of the plurality of data tables.

In some embodiments, the resource allocation apparatus 20 may include a plurality of functional modules made up of computer program segments. The computer program of each program segment in the resource allocation apparatus 20 can be stored in a memory of a computer device and executed by at least one processor to perform the functions of resource allocation (described in detail in fig. 1).

In this embodiment, the resource allocation apparatus 20 may be divided into a plurality of functional modules according to the functions performed by the apparatus. The functional module may include: a classification module 201, a construction module 202, a prediction module 203, a generation module 204, a clustering module 205, and an assignment module 206. The module referred to herein is a series of computer program segments capable of being executed by at least one processor and capable of performing a fixed function and is stored in memory. In the present embodiment, the functions of the modules will be described in detail in the following embodiments.

The classification module 201 is configured to perform feature classification on data in data tables of multiple target objects by using a feature classification model, so as to obtain multiple feature classes corresponding to each data table.

The building module 202 is configured to build a feature class data matrix for the corresponding target object according to the plurality of feature classes corresponding to each data table.

In an optional embodiment, the constructing module 202 constructs a feature class data matrix for the corresponding target object according to the plurality of feature classes corresponding to each data table, including:

The prediction module 203 is configured to perform prediction based on the feature class data matrix of each target object by using a multi-task multi-output prediction model, and obtain a plurality of evaluation classes corresponding to each target object and an evaluation probability corresponding to each evaluation class.

The generating module 204 is configured to generate a feature evaluation probability matrix according to the plurality of feature categories of each target object and the evaluation probability corresponding to each evaluation category.

In an optional embodiment, the generating module 204 generates the feature evaluation probability matrix according to the plurality of feature classes of each target object and the evaluation probability corresponding to each evaluation class, including:

Wherein the preset value may be 0. Exemplary, as shown in the following table:

	P1	P2	P3	P4
					T1	X1
T2		X2
					T3		X3

The clustering module 205 is configured to cluster the plurality of target objects according to the plurality of feature evaluation probability matrices to obtain a plurality of target object clusters.

In an optional embodiment, the clustering module 205 clusters the target objects according to the feature evaluation probability matrices to obtain target object clusters includes:

The allocating module 206 is configured to allocate resources to each target object according to the plurality of target object clusters.

In an optional embodiment, the allocating module 206 allocates resources for each target object according to the plurality of target object clusters, including:

Fig. 3 is a schematic structural diagram of a computer device according to a third embodiment of the present invention. In the preferred embodiment of the present invention, the computer device 3 includes a memory 31, at least one processor 32, at least one communication bus 33, and a transceiver 34.

It will be appreciated by those skilled in the art that the configuration of the computer device shown in fig. 3 does not constitute a limitation of the embodiments of the present invention, and may be a bus-type configuration or a star-type configuration, and that the computer device 3 may include more or less hardware or software than those shown, or a different arrangement of components.

In some embodiments, the computer device 3 is a device capable of automatically performing numerical calculation and/or information processing according to instructions set or stored in advance, and the hardware includes but is not limited to a microprocessor, an application specific integrated circuit, a programmable gate array, a digital processor, an embedded device, and the like. The computer device 3 may also include a client device, which includes, but is not limited to, any electronic product capable of interacting with a client through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device, for example, a personal computer, a tablet computer, a smart phone, a digital camera, etc.

It should be noted that the computer device 3 is only an example, and other electronic products that are currently available or may come into existence in the future, such as electronic products that can be adapted to the present invention, should also be included in the scope of the present invention, and are included herein by reference.

In some embodiments, the memory 31 has stored therein a computer program which, when executed by the at least one processor 32, implements all or part of the steps of the resource allocation method as described. The Memory 31 includes a Read-Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), a One-time Programmable Read-Only Memory (OTPROM), an electronically Erasable rewritable Read-Only Memory (Electrically-Erasable Programmable Read-Only Memory (EEPROM)), an optical Read-Only disk (CD-ROM) or other optical disk Memory, a magnetic disk Memory, a tape Memory, or any other medium readable by a computer capable of carrying or storing data.

Further, the computer-readable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.

The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.

In some embodiments, the at least one processor 32 is a Control Unit (Control Unit) of the computer device 3, connects various components of the entire computer device 3 by using various interfaces and lines, and executes various functions and processes data of the computer device 3 by running or executing programs or modules stored in the memory 31 and calling data stored in the memory 31. For example, the at least one processor 32, when executing the computer program stored in the memory, implements all or part of the steps of the resource allocation method described in the embodiments of the present invention; or to implement all or part of the functionality of the resource allocation means. The at least one processor 32 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips.

In some embodiments, the at least one communication bus 33 is arranged to enable connection communication between the memory 31 and the at least one processor 32 or the like.

Although not shown, the computer device 3 may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 32 through a power management device, so as to implement functions of managing charging, discharging, and power consumption through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The computer device 3 may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.

The integrated unit implemented in the form of a software functional module may be stored in a computer-readable storage medium. The software functional module is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a computer device, or a network device) or a processor (processor) to execute parts of the methods according to the embodiments of the present invention.

In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.

The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.

In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.

It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned. Furthermore, it is obvious that the word "comprising" does not exclude other elements or that the singular does not exclude the plural. A plurality of units or means recited in the present invention can also be implemented by one unit or means through software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.

Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for resource allocation, the method comprising:

constructing a feature category data matrix for the corresponding target object according to a plurality of feature categories corresponding to each data table, wherein the feature category data matrix comprises a plurality of feature category data vectors;

inputting a plurality of feature type data vectors of each target object into a multi-task multi-output prediction model for prediction to obtain a plurality of evaluation types corresponding to each feature type of each target object and an evaluation probability corresponding to each evaluation type;

2. The method of claim 1, wherein the constructing a feature class data matrix for the corresponding target object according to the plurality of feature classes corresponding to each data table comprises:

3. The method of resource allocation according to claim 2, wherein after obtaining the second feature class data matrix, the method further comprises:

4. The method of claim 1, wherein the generating a feature evaluation probability matrix according to the plurality of feature classes of each target object and the evaluation probability corresponding to each evaluation class comprises:

5. The method of claim 4, wherein the clustering the plurality of target objects according to the plurality of feature evaluation probability matrices to obtain a plurality of target object clusters comprises:

6. The resource allocation method according to any one of claims 1 to 5, wherein the allocating resources for each target object according to the plurality of target object clusters comprises:

7. The method of claim 6, wherein said calculating a rating score based on said average rating probability matrix comprises:

8. An apparatus for resource allocation, the apparatus comprising:

the construction module is used for constructing a feature category data matrix for the corresponding target object according to a plurality of feature categories corresponding to each data table, and the feature category data matrix comprises a plurality of feature category data vectors;

the prediction module is used for inputting a plurality of characteristic category data vectors of each target object into the multi-task multi-output prediction model for prediction to obtain a plurality of evaluation categories corresponding to each characteristic category of each target object and evaluation probabilities corresponding to each evaluation category;

9. A computer device, characterized in that the computer device comprises a processor for implementing the resource allocation method according to any one of claims 1 to 7 when executing a computer program stored in a memory.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method for resource allocation according to any one of claims 1 to 7.