CN116361567A

CN116361567A - Data processing method and system applied to cloud office

Info

Publication number: CN116361567A
Application number: CN202310643561.2A
Authority: CN
Inventors: 孙家祥; 李代艳
Original assignee: Taicang City Lvdian Information Technology Co ltd
Current assignee: Taicang City Lvdian Information Technology Co ltd
Priority date: 2023-06-01
Filing date: 2023-06-01
Publication date: 2023-06-30
Anticipated expiration: 2043-06-01
Also published as: CN116361567B

Abstract

The invention provides a data processing method and a data processing system applied to cloud office work, and relates to the technical field of artificial intelligence. In the invention, the candidate group analysis network is subjected to network updating operation to form an updated group analysis network corresponding to the candidate group analysis network; constructing a group distribution network to be analyzed based on office behavior data corresponding to each office person in a plurality of office persons in a target cloud office scene; and carrying out group analysis operation on the group distribution network to be analyzed by utilizing the updated group analysis network so as to output group existence analysis data corresponding to the group distribution network to be analyzed and corresponding target group member analysis data. Based on the above, the reliability of data processing can be improved to some extent.

Description

Data processing method and system applied to cloud office

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a data processing method and system applied to cloud office.

Background

The cloud office principle is to run traditional office software in a web browser in a lean Client (Thin Client) or intelligent Client (Smart Client) mode, so that the purpose of light weight is achieved. With the continuous development of cloud office technology, the world top cloud office application has strong compatibility to the traditional office document format and further shows unprecedented characteristics.

In cloud office, group analysis is required for office staff based on a certain requirement, however, in the prior art, correlation analysis is still performed by adopting the physical position (i.e. office position) of office staff as in the conventional technology, however, for cloud office, the effect of the physical position of office staff is not high, and the reliability of group analysis based on the relationship is not high.

Disclosure of Invention

In view of the foregoing, the present invention aims to provide a data processing method and system applied to cloud office, so as to improve the reliability of data processing to a certain extent.

In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme:

a data processing method applied to cloud office, comprising:

performing network updating operation on the candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network;

constructing a group distribution network to be analyzed based on office behavior data corresponding to each office worker in a plurality of office workers in a target cloud office scene, wherein each distribution network member in the group distribution network to be analyzed is the office worker, member attribute data of the distribution network member is office behavior data corresponding to the office worker, member correlation relations among the distribution network members are determined based on the correlation relations among the member attribute data and correlation relations among at least one other member data, the other member data belong to data except the member attribute data, and the office behavior data is data for recording network behaviors of the office workers in a text form;

And carrying out group analysis operation on the group distribution network to be analyzed by utilizing the updated group analysis network so as to output group existence analysis data corresponding to the group distribution network to be analyzed and target group member analysis data corresponding to the group distribution network to be analyzed, wherein the group existence analysis data are used for representing the possibility of existence of target groups in the group distribution network to be analyzed, the target group member analysis data are used for indicating group members of the target groups in the group distribution network to be analyzed, and the group existence analysis data and the target group member analysis data are used as pushing management basis for carrying out office text data, office image data and/or office voice data.

In some preferred embodiments, in the data processing method applied to cloud office, the step of performing a network update operation on the candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network includes:

extracting typical data for updating a candidate population analysis network and typical population identification data corresponding to the typical data, wherein the typical data comprises a typical population distribution network with typical populations and candidate typical population description vectors, the candidate population analysis network comprises a feature mining sub-network, a population analysis sub-network and a population member analysis sub-unit formed by combining a first number of processing units, and the typical population identification data comprises population existence characteristic data of the typical populations and population member identification data of the typical populations, and the first number is more than 1;

Loading the typical group distribution network to the feature mining sub-network, and performing distribution network mining operation on the typical group distribution network by utilizing the feature mining sub-network to output a typical first distribution network description vector and a typical second distribution network description vector of the typical group distribution network;

loading the candidate representative group description vector and the representative second distribution network description vector to the first number of processing units included in the group member analysis subunit, performing group member analysis operation on the candidate representative group description vector and the representative second distribution network description vector by using the first number of processing units to output first number of candidate group member prediction data corresponding to the first number of processing units, and loading the first number of candidate group member prediction data, the representative first distribution network description vector and the candidate representative group description vector to the feature mining subunit, mining an optimized representative group description vector corresponding to the candidate representative group description vector by using the feature mining subunit, and mining one candidate group member prediction data by using one processing unit;

Loading the optimized typical group description vector to the group analysis sub-network, performing group analysis operation on the optimized typical group description vector by using the group analysis sub-network, and analyzing group existence analysis data of an analysis group corresponding to the optimized typical group description vector according to the group analysis data analyzed by the group analysis sub-network;

loading the optimized representative group description vector and the representative second distribution network description vector to the first number of processing units included in the group member analysis subunit, and performing group member analysis operation on the optimized representative group description vector and the representative second distribution network description vector by using the first number of processing units to output first number of optimized group member prediction data corresponding to the first number of processing units, wherein one processing unit corresponds to one optimized group member prediction data;

analyzing the group member analysis data of the analysis group according to the first number of optimized group member prediction data, and performing network updating operation on the candidate group analysis network according to the group presence characterization data, the group presence analysis data, the group member identification data and the group member analysis data to form an updated group analysis network.

In some preferred embodiments, in the data processing method applied to cloud office, the feature mining sub-network includes a feature mining unit, a feature restoring unit, and a feature converting unit;

the step of loading the typical group distribution network to load into the feature mining sub-network, and performing a distribution network mining operation on the typical group distribution network by using the feature mining sub-network to output a typical first distribution network description vector and a typical second distribution network description vector of the typical group distribution network, includes:

loading the typical group distribution network to the feature mining unit included in the feature mining sub-network, performing distribution network mining operation on the typical group distribution network by using the feature mining unit, and marking the mined distribution network feature information to be a comparison typical distribution network description vector corresponding to the typical group distribution network;

loading the comparison typical distribution network description vector to the feature restoration unit included in the feature mining sub-network, and performing information interpolation operation on the comparison typical distribution network description vector by using the feature restoration unit to form an interpolation typical distribution network description vector corresponding to the comparison typical distribution network description vector;

Extracting, from the interpolated representative distribution network description vectors, a first extracted distribution network description vector for loading to the feature transformation unit and a second extracted distribution network description vector for loading to the first number of processing units;

the first extracted distribution network description vector is labeled as the representative first distribution network description vector, and the second extracted distribution network description vector is labeled as the representative second distribution network description vector.

In some preferred embodiments, in the data processing method applied to cloud office, the feature recovering unit includes a second number of information interpolating subunits, where the second number is greater than 1;

the step of loading the comparison typical distribution network description vector to load into the feature restoration unit included in the feature mining sub-network, and performing information interpolation operation on the comparison typical distribution network description vector by using the feature restoration unit to form an interpolated typical distribution network description vector corresponding to the comparison typical distribution network description vector, includes:

determining an a-th information interpolation subunit and a b-th information interpolation subunit among the second number of information interpolation subunits, a=b-1;

Loading the comparison typical distribution network description vector to the a-th information interpolation subunit, and performing information interpolation operation on the comparison typical distribution network description vector by using the a-th information interpolation subunit to output an a-th interpolation description vector corresponding to the a-th information interpolation subunit;

according to the a-th interpolation description vector, carrying out optimization adjustment operation on the comparison typical distribution network description vector, loading the optimization-adjusted comparison typical distribution network description vector into the b-th information interpolation subunit, and carrying out information interpolation operation on the optimization-adjusted comparison typical distribution network description vector by utilizing the b-th information interpolation subunit to form a b-th interpolation description vector corresponding to the b-th information interpolation subunit;

and according to the optimized and adjusted comparison typical distribution network description vector and the b-th interpolation description vector, analyzing an interpolation typical distribution network description vector corresponding to the comparison typical distribution network description vector.

In some preferred embodiments, in the data processing method applied to cloud office, the interpolated typical distribution network description vector includes the optimized and adjusted comparative typical distribution network description vector and an interpolated description vector corresponding to a target information interpolation subunit, where the target information interpolation subunit is a last information interpolation subunit in the second number of information interpolation subunits, and the optimized and adjusted comparative typical distribution network description vector is obtained based on an a-th interpolated description vector formed by the a-th information interpolation subunit;

The step of extracting, from the interpolated representative distribution network description vectors, a first extracted distribution network description vector for loading to the feature conversion unit and a second extracted distribution network description vector for loading to the first number of processing units, includes:

extracting the optimized and adjusted comparative representative distribution network description vector from the interpolated representative distribution network description vector, and marking the optimized and adjusted comparative representative distribution network description vector as a first extracted distribution network description vector for loading into the feature conversion unit;

and extracting the interpolation description vector corresponding to the target information interpolation subunit from the interpolation typical distribution network description vector, and marking the interpolation description vector corresponding to the target information interpolation subunit as a second extraction distribution network description vector for loading to the first number of processing units.

In some preferred embodiments, in the data processing method applied to cloud office, the feature mining sub-network includes a feature conversion unit;

the step of loading the first number of candidate group member prediction data, the representative first distribution network description vector and the candidate representative group description vector to be loaded into the feature mining sub-network, and mining an optimized representative group description vector corresponding to the candidate representative group description vector by using the feature mining sub-network, includes:

Forming to-be-processed merging data for loading to the feature conversion unit according to the first number of candidate group member prediction data, the representative first distribution network description vector and the candidate representative group description vector;

and loading the merged data to be processed to the feature conversion unit, performing data conversion operation on the merged data to be processed by using the feature conversion unit, and analyzing an optimized typical population description vector corresponding to the candidate typical population description vector according to a data conversion result formed by the data conversion operation.

In some preferred embodiments, in the data processing method applied to cloud office, the feature mining sub-network includes a feature restoration unit, the feature restoration unit includes a second number of information interpolation sub-units, the typical first distribution network description vector includes a third number of interpolation description vectors formed by information interpolation operations performed by a third number of information interpolation sub-units in the second number of information interpolation sub-units, one information interpolation sub-unit is used for performing information interpolation operations to form one interpolation description vector, and a difference value between the second number and the third number is equal to 1;

The step of forming merged data to be processed for loading into the feature transformation unit from the first number of candidate group member prediction data, the representative first distribution network description vector, and the candidate representative group description vector, comprises:

selecting an a-th interpolation description vector from the third number of interpolation description vectors;

selecting an a-th data conversion structure corresponding to the a-th interpolation description vector from a third number of data conversion structures included in the feature conversion unit;

analyzing a standard group member description vector for loading to the feature conversion unit according to the first number of candidate group member prediction data, and analyzing an a-th typical group member description vector corresponding to the a-th data conversion structure according to the standard group member description vector;

in the process of analyzing an a-th typical waiting description vector according to the candidate typical group description vector, marking the a-th typical waiting description vector, the a-th typical group member description vector and the a-th interpolation description vector to be the waiting merging data of the a-th data conversion structure in the feature conversion unit.

In some preferred embodiments, in the data processing method applied to cloud office, the step of loading the merged data to be processed into the feature conversion unit, performing a data conversion operation on the merged data to be processed by using the feature conversion unit, and analyzing an optimized representative group description vector corresponding to the candidate representative group description vector according to a data conversion result formed by the data conversion operation includes:

loading the a-th typical waiting description vector, the a-th typical group member description vector and the a-th interpolation description vector to the a-th data conversion structure of the feature conversion unit, performing data conversion operation on the a-th typical waiting description vector, the a-th typical group member description vector and the a-th interpolation description vector by using the a-th data conversion structure, marking a data conversion result formed by the data conversion operation to mark the b-th typical waiting description vector, and analyzing an optimized typical group description vector corresponding to the candidate typical group description vector according to the b-th typical waiting description vector.

In some preferred embodiments, in the data processing method applied to cloud office, the updating group analysis network includes a feature mining sub-network, a group analysis sub-network, and a group member analysis sub-unit formed by combining a first number of processing units, where the first number is greater than 1, and the step of performing a group analysis operation on the group analysis network to be analyzed by using the updating group analysis network to output group presence analysis data corresponding to the group analysis network to be analyzed and target group member analysis data corresponding to the group analysis network to be analyzed includes:

extracting a reference group description vector;

loading the group distribution network to be analyzed to the feature mining sub-network, and performing distribution network mining operation on the group distribution network to be analyzed by utilizing the feature mining sub-network to output a first distribution network description vector and a second distribution network description vector of the group distribution network to be analyzed;

loading the reference group description vector and the second distribution network description vector to the first number of processing units in the group member analysis subunit, performing group member analysis operation on the reference group description vector and the second distribution network description vector by using the first number of processing units to output first number of pending group member prediction data of the first number of processing units, and loading the first number of pending group member prediction data, the first distribution network description vector and the reference group description vector to load to the feature mining sub-network, and mining an optimized group description vector corresponding to the reference group description vector by using the feature mining sub-network;

Loading the optimized group description vector to the group analysis sub-network, performing group analysis operation on the optimized group description vector by utilizing the group analysis sub-network, and analyzing group existence estimation data of an analysis group corresponding to the optimized group description vector according to the group analysis data analyzed by the group analysis sub-network;

when the population existence estimation data corresponding to the optimized population description vector reflects that the analysis population corresponding to the optimized population description vector belongs to a target population, loading the optimized population description vector and the second distribution network description vector into the first number of processing units included in the population member analysis subunit, and performing population member analysis operation on the optimized population description vector and the second distribution network description vector by using the first number of processing units so as to output first number of target population member prediction data of the first number of processing units, wherein one processing unit corresponds to one target population member prediction data;

and determining target group member analysis data corresponding to the group distribution network to be analyzed based on the first number of target group member prediction data.

The embodiment of the invention also provides a data processing system applied to cloud office, which comprises a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program so as to realize the data processing method applied to cloud office.

According to the data processing method and system applied to cloud office, network updating operation can be performed on the candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network; constructing a group distribution network to be analyzed based on office behavior data corresponding to each office person in a plurality of office persons in a target cloud office scene; and carrying out group analysis operation on the group distribution network to be analyzed by utilizing the updated group analysis network so as to output group existence analysis data corresponding to the group distribution network to be analyzed and corresponding target group member analysis data. Based on the foregoing, since the group analysis can be performed based on the office behavior data corresponding to the office staff, after the action of the physical position of the office staff in the cloud office is reduced, the group analysis can be realized, that is, the office behavior data is utilized to perform the effective group analysis, so that the reliability of data processing can be improved to a certain extent, and the problem of low reliability of data processing (that is, the problem of low reliability of group analysis) caused by the fact that the cloud office still uses the physical position for analysis in the prior art is solved.

In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

FIG. 1 is a block diagram of a data processing system applied to cloud office according to an embodiment of the present invention.

Fig. 2 is a flowchart illustrating steps included in a data processing method applied to cloud office according to an embodiment of the present invention.

Fig. 3 is a schematic diagram of each module included in a data processing device applied to cloud office according to an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

Accordingly, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by one of ordinary skill in the art without undue burden on the person of ordinary skill in the art based on the embodiments of the present invention, are within the scope of the present invention.

As shown in FIG. 1, an embodiment of the present invention provides a data processing system applied to cloud office. Wherein, the data processing system applied to cloud office can comprise a memory and a processor.

In particular, the memory and the processor are electrically connected directly or indirectly to enable transmission or interaction of data. For example, electrical connection may be made to each other via one or more communication buses or signal lines. At least one software functional module (computer program) that may exist in the form of software or firmware may be stored in the memory. The processor may be configured to execute the executable computer program stored in the memory, so as to implement the data processing method applied to cloud office provided by the embodiment of the present invention.

It should be appreciated that in some possible embodiments, the Memory may be, but is not limited to, random Access Memory (RAM), read Only Memory (ROM), programmable Read-Only Memory (PROM), erasable Programmable Read-Only Memory (EPROM), electrically Erasable Programmable Read-Only Memory (EEPROM), and the like.

It should be appreciated that in some possible embodiments, the Processor may be a general purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), a System on Chip (SoC), etc.; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

It should be appreciated that in some possible embodiments, the data processing system applied to cloud offices may be a server with data processing capabilities.

With reference to fig. 2, the embodiment of the invention further provides a data processing method applied to cloud office, which can be applied to the data processing system applied to cloud office. The method steps defined by the flow related to the data processing method applied to cloud office can be realized by the data processing system applied to cloud office. The specific flow shown in fig. 2 will be described in detail below.

And step S110, performing network updating operation on the candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network.

In an embodiment of the present invention, the data processing system applied to cloud office may perform a network update operation on a candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network. The candidate group analysis network can be an initially built neural network, or can be a candidate neural network after updating the initially built neural network to a certain extent.

Step S120, a group distribution network to be analyzed is constructed based on office behavior data corresponding to each office person in a plurality of office persons in a target cloud office scene.

In an embodiment of the present invention, the data processing system applied to cloud office may construct a group distribution network to be analyzed based on office behavior data corresponding to each of a plurality of office workers in a target cloud office scene. Each member of the group distribution network to be analyzed is the office staff, the member attribute data of the member of the distribution network is office behavior data corresponding to the office staff, the member correlation relationship among the members of the distribution network is determined based on the correlation relationship among the member attribute data and the correlation relationship among at least one other member data, and the other member data belongs to data other than the member attribute data, for example, the member attribute data can be information such as age, gender, position, working time, school and the like of the office staff. In addition, in the group distribution network to be analyzed, the member correlation relationships among the distribution network members can be used for representing the distances among the distribution coordinates of the distribution network members, for example, the closer the member correlation relationships are, the smaller the distances among the distribution coordinates of the corresponding distribution network members can be. In addition, the office behavior data refers to data for recording network behaviors of office workers in the form of text, such as office behaviors.

And step S130, carrying out group analysis operation on the group distribution network to be analyzed by utilizing the updated group analysis network so as to output group existence analysis data corresponding to the group distribution network to be analyzed and target group member analysis data corresponding to the group distribution network to be analyzed.

In this embodiment of the present invention, the data processing system applied to cloud office may perform a group analysis operation on the group distribution network to be analyzed by using the updated group analysis network, so as to output group presence analysis data corresponding to the group distribution network to be analyzed and target group member analysis data corresponding to the group distribution network to be analyzed. The population presence analysis data is used for representing the possibility of the existence of a target population in the population distribution network to be analyzed, for example, 0 represents the absence of the target population, 1 represents the existence of the target population, the target population member analysis data is used for indicating the population members of the target population existing in the population distribution network to be analyzed, for example, the population members such as the population member 1, the population member 4, the population member 7 and the like belong to the target population, and the target population is the population with the target population relationship among the population members, so that when the candidate population analysis network is subjected to network updating operation, the typical population in the corresponding typical population distribution network should have the target population relationship, and the updated population analysis network formed by updating can learn the target population relationship. In addition, the group presence analysis data and the target group member analysis data are used as pushing management basis for office text data, office image data and/or office voice data. For example, when pushing operations of office text data, office image data and/or office voice data are required to be performed on a target group having the target group relationship, a target group may be determined based on the group presence analysis data and the target group member analysis data, and then data pushing operations may be performed on group members included in the target group together.

Based on the foregoing (i.e., step S110, step S120 and step S130), since the group analysis can be performed based on the office behavior data corresponding to the office staff, so that after the effect of the physical location of the office staff in the cloud office is reduced, the group analysis can also be implemented, i.e., the office behavior data is utilized to perform the effective group analysis, so that the reliability of data processing can be improved to a certain extent, and the problem of low reliability of data processing (i.e., the problem of low reliability of group analysis) caused by the fact that the cloud office still uses the physical location for analysis in the prior art is improved.

It should be understood that, in some possible embodiments, step S110 in the foregoing may further include the following:

extracting typical data for updating a candidate population analysis network and typical population identification data corresponding to the typical data, wherein the typical data comprises a typical population distribution network with typical populations (the typical population distribution network is constructed in a previous related description) and candidate typical population description vectors (the candidate typical population description vectors for constituting the typical data are configured vectors when the candidate population analysis network is updated at the beginning, the candidate typical population description vectors are used as network parameters of the candidate population analysis network, when each round of updating the candidate population analysis network, the candidate typical population description vectors constituting the typical data are updated, the updated candidate typical population description vectors are used as candidate typical population description vectors in new typical data to iteratively update the candidate population analysis network, namely, the candidate typical population description vectors in subsequent typical data are formed in the network updating process of the candidate population analysis network, the candidate typical population description vectors can be used as reference population description vectors when the candidate population analysis network is trained at the end, the candidate population analysis network comprises characteristic feature groups, the first representative population is formed by the characteristic population, and the characteristic population is more than the first representative population is formed by the typical population, and the characteristic population is represented by the characteristic population of the first representative population, and the characteristic population is formed by the characteristic population of the typical population analysis network;

Loading the typical group distribution network to the feature mining sub-network, and performing distribution network mining operation on the typical group distribution network by utilizing the feature mining sub-network to output a typical first distribution network description vector and a typical second distribution network description vector of the typical group distribution network, namely performing mining operation of key information;

Loading the optimized typical population description vector into the population analysis sub-network, carrying out population analysis operation on the optimized typical population description vector by utilizing the population analysis sub-network, and analyzing population existence analysis data of an analysis population corresponding to the optimized typical population description vector according to the population analysis data analyzed by the population analysis sub-network, wherein the population existence analysis data is used for representing the possibility of existence of a corresponding population;

loading the optimized representative group description vector and the representative second distribution network description vector to be loaded into the first number of processing units included in the group member analysis subunit, performing group member analysis operation on the optimized representative group description vector and the representative second distribution network description vector by using the first number of processing units, so as to output first number of optimized group member prediction data corresponding to the first number of processing units, wherein one processing unit corresponds to one of the optimized group member prediction data, and a manner of determining the first number of optimized group member prediction data of the first number of processing units can be consistent with a manner of determining the first number of candidate group member prediction data;

Analyzing the population member analysis data of the analysis population according to the first number of optimized population member prediction data, for example, the population member analysis data includes the first number of optimized population member prediction data, and performing a network update operation on the candidate population analysis network according to the population presence characterization data, the population presence analysis data, the population member identification data, and the population member analysis data to form an updated population analysis network.

It should be understood that, in some possible embodiments, the feature mining sub-network may include a feature mining unit, a feature restoration unit, and a feature conversion unit, based on which the step of loading the exemplary population distribution network to load into the feature mining sub-network, performing a distribution network mining operation on the exemplary population distribution network by using the feature mining sub-network to output an exemplary first distribution network description vector and an exemplary second distribution network description vector of the exemplary population distribution network may further include the following:

loading the typical group distribution network to the feature mining unit included in the feature mining sub-network, performing a distribution network mining operation on the typical group distribution network by using the feature mining unit, and marking the mined distribution network feature information to be a comparative typical distribution network description vector corresponding to the typical group distribution network, wherein the feature mining unit can be an encoding neural network;

It should be appreciated that in some possible embodiments, the feature recovery unit may include a second number of information interpolation (interpolation) sub-units, where the second number is greater than 1, based on which the step of loading the comparison exemplary distribution network description vector to load into the feature recovery unit included in the feature mining sub-network, and performing an information interpolation operation on the comparison exemplary distribution network description vector by using the feature recovery unit to form an interpolated exemplary distribution network description vector corresponding to the comparison exemplary distribution network description vector may further include the following steps:

Determining, among the second number of information interpolation subunits, an a-th information interpolation subunit and a b-th information interpolation subunit, a=b-1, the a-th information interpolation subunit being any one of the information interpolation subunits, the b-th information interpolation subunit being the latter one of the a-th information interpolation subunits;

loading the comparison typical distribution network description vector to the a-th information interpolation subunit, and performing information interpolation operation on the comparison typical distribution network description vector by using the a-th information interpolation subunit to output an a-th interpolation description vector corresponding to the a-th information interpolation subunit, such as a 1-th interpolation description vector corresponding to the 1-st information interpolation subunit;

according to the a-th interpolation description vector, performing an optimization adjustment operation on the comparison typical distribution network description vector, for example, the a-th interpolation description vector can be directly used as the comparison typical distribution network description vector after optimization adjustment, or the a-th interpolation description vector and the comparison typical distribution network description vector are subjected to superposition and other operations to obtain the comparison typical distribution network description vector after optimization adjustment, and the comparison typical distribution network description vector after optimization adjustment is loaded to be loaded into the b-th information interpolation subunit, and the b-th information interpolation subunit is utilized to perform an information interpolation operation on the comparison typical distribution network description vector after optimization adjustment to form the b-th interpolation description vector corresponding to the b-th information interpolation subunit;

It should be understood that, in some possible embodiments, the interpolated representative distribution network description vector may include the optimized comparison representative distribution network description vector and an interpolated description vector corresponding to a target information interpolation subunit, where the target information interpolation subunit is a last information interpolation subunit in the second number of information interpolation subunits, and the optimized comparison representative distribution network description vector is obtained based on an a-th interpolated description vector formed by the a-th information interpolation subunit, based on which, in the interpolated representative distribution network description vector, the steps of extracting a first extracted distribution network description vector for loading into the feature conversion unit and a second extracted distribution network description vector for loading into the first number of processing units may further include:

Wherein it should be understood that, in some possible embodiments, the first number of processing units may include a first processing unit corresponding to a principal member of the population of the typical population and a second processing unit corresponding to a non-principal member of the population of the typical population, based on which the step of loading the candidate representative population description vector and the representative second distribution network description vector into the first number of processing units included in the population member analysis subunit, performing a population member analysis operation on the candidate representative population description vector and the representative second distribution network description vector by using the first number of processing units to output a first number of candidate population member prediction data corresponding to the first number of processing units may further include the following steps:

Loading the candidate representative group description vector and the representative second distribution network description vector to the first processing unit, performing a first conversion fusion operation on the candidate representative group description vector and the representative second distribution network description vector by using the first processing unit to form a group main member description vector corresponding to the group main member, for example, the candidate representative group description vector may be first subjected to a linear mapping operation to obtain a corresponding linear mapping vector, then the linear mapping vector and the representative second distribution network description vector may be aggregated, for example, superposition or stitching is performed to obtain an aggregate description vector, finally, a convolution operation may be performed on the aggregate description vector to obtain a group main member description vector, and, according to the group main member description vector, the candidate group member prediction data corresponding to the first processing unit is predicted to be output, that is, the analysis prediction of the main member may be performed according to the group main member description vector, specifically, a soft tmax function may be included, for example, the probability function may be calculated to be larger than a first group main member, for example, and the first group may be calculated to be larger than a first group member, and then a first group may be calculated to be a main member;

Loading the candidate representative group description vector and the representative second distribution network description vector to the second processing unit, performing a second conversion fusion operation on the candidate representative group description vector and the representative second distribution network description vector by using the second processing unit to form a group non-main member description vector corresponding to the group non-main member, wherein the candidate representative group description vector can be subjected to linear mapping operation firstly to obtain a corresponding linear mapping vector, mapping parameters of each step can be different from those of the previous step respectively as network parameters of a neural network, the mapping parameters can be correspondingly updated, then the linear mapping vector and the representative second distribution network description vector can be aggregated, such as superposition or splicing, to obtain an aggregate description vector, finally, the aggregate description vector can be subjected to convolution operation, parameters of the convolution operation can also be different from those of the previous steps to obtain a group non-main member description vector, and, according to the group non-main member description vector, the prediction output of the second processing unit can be subjected to linear mapping operation to obtain a corresponding linear mapping vector, and the second processing unit can be used as a prediction value, such as a prediction value, and the non-main member can be calculated to be smaller than a prediction value, such as a prediction value, and a non-main group can be calculated, and a non-main member can be calculated based on the prediction value, and a non-main group is calculated, and a prediction value can be calculated based on the non-main group, and a prediction value is calculated based on the prediction value;

The method further includes performing a fusion operation on the candidate group member prediction data corresponding to the first processing unit and the candidate group member prediction data corresponding to the second processing unit to form a first number of candidate group member prediction data of the first number of processing units, that is, the first number of candidate group member prediction data may include candidate group member prediction data corresponding to the first processing unit and candidate group member prediction data corresponding to the second processing unit, and in other embodiments, other multi-level classification may be performed on the group members, such as a first importance, a second importance, a third importance, and a fourth importance, and configuration may be performed according to actual requirements, and accordingly, group member identification data of the typical group may also have identification of a group main member and identification of a non-group main member, so as to correspondingly perform analysis optimization.

It should be appreciated that in some possible embodiments, the feature mining sub-network may include a feature transformation unit, based on which the step of loading the first number of candidate group member prediction data, the representative first distribution network description vector, and the candidate representative group description vector to load into the feature mining sub-network, and the step of mining the optimized representative group description vector corresponding to the candidate representative group description vector using the feature mining sub-network may further include the following:

Forming to-be-processed merging data for loading to the feature conversion unit according to the first number of candidate group member prediction data, the typical first distribution network description vector and the candidate typical group description vector, namely merging data (vectors) of three aspects;

It should be appreciated that in some possible embodiments, the feature mining sub-network may include a feature restoration unit, the feature restoration unit includes a second number of information interpolation sub-units, the exemplary first distribution network description vector includes a third number of interpolation description vectors formed by information interpolation operations performed by a third number of information interpolation sub-units of the second number of information interpolation sub-units, the second number of information interpolation sub-units may be cascade connected, the third number of information interpolation sub-units may be a foremost third number of information interpolation sub-units of the second number of information interpolation sub-units, an information interpolation sub-unit is configured to perform information interpolation operations to form an interpolation description vector, a difference between the second number and the third number is equal to 1, based on which the step of forming the to-be-processed merging data for loading to the feature conversion unit may include:

Selecting an a-th interpolation description vector from the third number of interpolation description vectors, wherein the a-th interpolation description vector can be any interpolation description vector;

selecting an a-th data conversion structure corresponding to the a-th interpolation description vector from a third number of data conversion structures included in the feature conversion unit, wherein the third number of data conversion structures may be cascade-connected, and thus the a-th data conversion structure may be determined;

according to the first number of candidate group member prediction data, analyzing a standard group member description vector for loading to the feature conversion unit, for example, a mapping vector of the first number of candidate group member prediction data in a feature space can be used as a standard group member description vector, and according to the standard group member description vector, an a-th typical group member description vector corresponding to the a-th data conversion structure can be analyzed, for example, the standard group member description vector can be directly used as an a-th typical group member description vector, or the standard group member description vector can be processed to obtain the a-th typical group member description vector;

In the process of analyzing the a-th representative waiting description vector according to the candidate representative group description vector (for example, the candidate representative group description vector may be regarded as a first representative waiting description vector), the a-th representative waiting description vector, the a-th representative group member description vector and the a-th interpolation description vector are marked to be waiting merging data of the a-th data conversion structure in the feature conversion unit, and it may be understood that each data conversion structure in the feature conversion unit corresponds to one waiting merging data.

It should be understood that, in some possible embodiments, the step of loading the merged data to be processed to be loaded into the feature conversion unit, performing a data conversion operation on the merged data to be processed by using the feature conversion unit, and analyzing an optimized representative population description vector corresponding to the candidate representative population description vector according to a data conversion result formed by the data conversion operation may further include the following steps:

loading the a-th typical waiting description vector, the a-th typical group member description vector and the a-th interpolation description vector to the a-th data conversion structure of the feature conversion unit, performing data conversion operation on the a-th typical waiting description vector, the a-th typical group member description vector and the a-th interpolation description vector by using the a-th data conversion structure, marking a data conversion result formed by the data conversion operation to mark the b-th typical waiting description vector, and analyzing an optimized typical group description vector corresponding to the candidate typical group description vector according to the b-th typical waiting description vector, for example, the data conversion result output by the last data conversion structure can be used as the formed data conversion result.

Wherein, it should be understood that, in some possible embodiments, the step of performing the data conversion operation on the a-th exemplary pending description vector, the a-th exemplary group member description vector and the a-th interpolation description vector may further include the following:

multiplying the a-th representative group member description vector and the transpose vector of the a-th interpolation description vector to obtain a first multiplication result, and then adding the a-th representative group member description vector and the first multiplication result to obtain a first addition result;

and multiplying the first addition result and the a-th interpolation description vector to obtain a second multiplication result, and then adding the second multiplication result and the a-th typical description vector to be processed to output a corresponding data conversion result.

Wherein it should be understood that in some possible embodiments, the optimized representative population description vectors may be a fourth number, each optimized representative population description vector may correspond to a population member analysis data and a population presence analysis data, based on which the step of analyzing the population member analysis data of the analysis population based on the first number of optimized population member prediction data, and performing a network update operation on the candidate population analysis network based on the population presence characterization data, the population presence analysis data, the population member identification data, and the population member analysis data to form an updated population analysis network may further include the steps of:

Analyzing group member analysis data of the analysis group according to the first number of optimized group member prediction data, wherein the group member analysis data comprises the first number of optimized group member prediction data, namely, the optimized group member prediction data are combined together;

analyzing the group member analysis data corresponding to the group member identification data in the group member analysis data corresponding to the fourth number of optimized typical group description vectors, and marking the optimized typical group description vectors corresponding to the group member analysis data corresponding to the group member identification data to be marked as comparison typical group description vectors;

marking the group member identification data to be typical group member actual data corresponding to a comparison typical group description vector, and marking the group presence characterization data to be typical group presence actual data corresponding to the comparison typical group description vector;

extracting non-population presence characterization data (i.e., data characterizing members other than the population), and tagging the non-population presence characterization data to tag representative population presence actual data (i.e., indicating the absence of a representative population) corresponding to an optimized representative population description vector other than the comparative representative population description vector of the fourth plurality of optimized representative population description vectors;

Calculating and outputting a first index of network updating errors according to the group member analysis data corresponding to the comparison representative group description vector and the corresponding representative group member actual data, and calculating and outputting a second index of network updating errors according to the group presence analysis data corresponding to the fourth number of optimization representative group description vectors and the corresponding representative group presence actual data;

and updating and adjusting the network parameters of the candidate group analysis network based on the first network updating error index and the second network updating error index to form an updated group analysis network, for example, the network parameters of the candidate group analysis network can be updated and adjusted along the direction of reducing the first network updating error index and the second network updating error index.

It should be understood that, in some possible embodiments, step S130 in the foregoing may further include the following:

extracting reference group description vectors (as described in the previous relation);

loading the group distribution network to be analyzed to the feature mining sub-network, and performing distribution network mining operation on the group distribution network to be analyzed by utilizing the feature mining sub-network to output a first distribution network description vector and a second distribution network description vector (as related description);

Loading the reference group description vector and the second distribution network description vector to the first number of processing units in the group member analysis subunit, performing a group member analysis operation on the reference group description vector and the second distribution network description vector with the first number of processing units to output a first number of pending group member prediction data of the first number of processing units, and loading the first number of pending group member prediction data, the first distribution network description vector, and the reference group description vector to load to the feature mining sub-network, mining an optimized group description vector corresponding to the reference group description vector with the feature mining sub-network (as described in the foregoing related description);

loading the optimized group description vector to the group analysis sub-network, performing group analysis operation on the optimized group description vector by using the group analysis sub-network, and analyzing group existence estimation data (such as related description) of an analysis group corresponding to the optimized group description vector according to the group analysis data analyzed by the group analysis sub-network;

and determining target group member analysis data corresponding to the group distribution network to be analyzed based on the first number of target group member prediction data, wherein the target group member analysis data can comprise the first number of target group member prediction data.

With reference to fig. 3, an embodiment of the present invention further provides a data processing device applied to cloud office, which may be applied to the above data processing system applied to cloud office. The data processing device applied to cloud office may include the following software functional modules:

The network updating module is used for carrying out network updating operation on the candidate group analysis network so as to form an updated group analysis network corresponding to the candidate group analysis network;

the group distribution network construction module is used for constructing a group distribution network to be analyzed based on office behavior data corresponding to each office worker in a plurality of office workers in a target cloud office scene, each distribution network member in the group distribution network to be analyzed is the office worker, member attribute data of the distribution network member is office behavior data corresponding to the office worker, member correlation among the distribution network members is determined based on the correlation among the member attribute data and correlation among at least one other member data, and the other member data belong to data except the member attribute data;

the group analysis module is used for carrying out group analysis operation on the group distribution network to be analyzed by utilizing the updated group analysis network so as to output group existence analysis data corresponding to the group distribution network to be analyzed and target group member analysis data corresponding to the group distribution network to be analyzed, wherein the group existence analysis data is used for representing the possibility of existence of target groups in the group distribution network to be analyzed, and the target group member analysis data is used for indicating group members of the target groups in the group distribution network to be analyzed.

In summary, according to the data processing method and system applied to cloud office provided by the invention, network updating operation can be performed on the candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network; constructing a group distribution network to be analyzed based on office behavior data corresponding to each office person in a plurality of office persons in a target cloud office scene; and carrying out group analysis operation on the group distribution network to be analyzed by utilizing the updated group analysis network so as to output group existence analysis data corresponding to the group distribution network to be analyzed and corresponding target group member analysis data. Based on the foregoing, since the group analysis can be performed based on the office behavior data corresponding to the office staff, after the action of the physical position of the office staff in the cloud office is reduced, the group analysis can be realized, that is, the office behavior data is utilized to perform the effective group analysis, so that the reliability of data processing can be improved to a certain extent, and the problem of low reliability of data processing (that is, the problem of low reliability of group analysis) caused by the fact that the cloud office still uses the physical position for analysis in the prior art is solved.

The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A data processing method applied to cloud office, comprising:

2. The data processing method applied to cloud office as claimed in claim 1, wherein the step of performing a network update operation on the candidate group analysis network to form an updated group analysis network corresponding to the candidate group analysis network comprises:

3. The data processing method applied to cloud office as claimed in claim 2, wherein the feature mining sub-network comprises a feature mining unit, a feature restoring unit and a feature converting unit;

4. The data processing method applied to cloud office as claimed in claim 3, wherein the feature reduction unit comprises a second number of information interpolation subunits, the second number being greater than 1;

5. The data processing method applied to cloud office as claimed in claim 4, wherein the interpolated canonical distribution network description vector includes the optimized comparison canonical distribution network description vector and an interpolated description vector corresponding to a target information interpolation subunit, the target information interpolation subunit being a last information interpolation subunit of the second number of information interpolation subunits, the optimized comparison canonical distribution network description vector being obtained based on an a-th interpolated description vector formed by the a-th information interpolation subunit;

6. The data processing method applied to cloud office as claimed in claim 2, wherein the feature mining sub-network comprises a feature conversion unit;

7. The data processing method applied to cloud office as claimed in claim 6, wherein the feature mining sub-network comprises a feature restoration unit, the feature restoration unit comprises a second number of information interpolation sub-units, the typical first distribution network description vector comprises a third number of interpolation description vectors formed by information interpolation operations performed by a third number of information interpolation sub-units among the second number of information interpolation sub-units, one information interpolation sub-unit is used for performing information interpolation operations to form one interpolation description vector, and a difference between the second number and the third number is equal to 1;

8. The data processing method for cloud office application as claimed in claim 7, wherein the step of loading the merged data to be processed to be loaded into the feature conversion unit, performing a data conversion operation on the merged data to be processed by using the feature conversion unit, and analyzing an optimized representative group description vector corresponding to the candidate representative group description vector according to a data conversion result formed by the data conversion operation comprises:

9. The data processing method applied to cloud office as claimed in any one of claims 1 to 8, wherein the updated group analysis network includes a feature mining sub-network, a group analysis sub-network, and a group member analysis sub-unit formed by combining a first number of processing units, the first number being greater than 1, the step of performing a group analysis operation on the group distribution network to be analyzed using the updated group analysis network to output group presence analysis data corresponding to the group distribution network to be analyzed and target group member analysis data corresponding to the group distribution network to be analyzed, comprising:

extracting a reference group description vector;

10. A data processing system for use in cloud offices, comprising a processor and a memory, the memory being for storing a computer program, the processor being for executing the computer program to implement the data processing method for use in cloud offices of any of claims 1-9.