CN114676420A

CN114676420A - AI and big data combined cloud office information processing method and server

Info

Publication number: CN114676420A
Application number: CN202210235085.6A
Authority: CN
Inventors: 王宇
Original assignee: Individual
Current assignee: Individual
Priority date: 2022-03-10
Filing date: 2022-03-10
Publication date: 2022-06-28

Abstract

The application discloses a cloud office information processing method and a server combining AI and big data. When the method is applied, office business abnormity existing in the target office node is weighed through the target office log, attention type business abnormity or group type business abnormity existing in the target office node is weighed through the cooperative office log, and the multi-mode behavior description of the target office node is determined in a comprehensive analysis mode, so that the aim of improving the richness of the multi-mode behavior description is fulfilled, the beneficial effect of improving the analysis accuracy of the task process of the office node is realized, and the technical problem of lower analysis accuracy of the task process of the office node is further improved to a certain extent.

Description

AI and big data combined cloud office information processing method and server

Technical Field

The application relates to the technical field of AI, big data and cloud office, in particular to a cloud office information processing method and a server combining the AI and the big data.

Background

The continuous development of Artificial Intelligence (AI) has prompted the intelligent upgrade of modern office models, and the intelligent office and cloud office have come into operation. In addition, the big data can ensure the smooth operation of the office business process, thereby effectively improving the office efficiency and saving office resources.

At present, with the continuous expansion of the scales of smart office work and cloud office work, the number and the types of office equipment (office nodes) accessing a cloud office system are more and more, and the security problem of the cloud office work cannot be ignored.

In order to ensure the data information security in the cloud office process, detection and analysis need to be performed on task processes of different office nodes, but it is difficult for related technologies to accurately detect the task processes of the office nodes.

Disclosure of Invention

The embodiment of the application provides a cloud office information processing method and a server combining AI and big data.

In view of this, an aspect of the present application provides a cloud office information processing method combining an AI and big data, which is applied to a cloud office information processing server, and the method includes: acquiring a cooperative office log of a cooperative office node connected with a target office node to be analyzed and a target office log generated by the target office node in a target service time period, wherein the target office log is an office record generated in the office process of the target office node; analyzing the collaborative office logs and the target office logs respectively to obtain derived quantitative descriptions corresponding to the collaborative office logs and business quantitative descriptions corresponding to the target office logs, wherein the derived quantitative descriptions are used for representing attention-type contents between the target office nodes and the connected collaborative office nodes, and the business quantitative descriptions are used for representing operation preference descriptions presented by the target office nodes in the office process; determining a multi-modal behavior description of the target office node according to the derived quantitative description and the business quantitative description; on the premise that the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to a target task process, wherein the target behavior description is the multi-modal behavior description of the office node corresponding to the target task process.

Another aspect of the present application provides a cloud office information processing server, including: a memory for storing executable instructions; and the processor is used for operating the cloud office information processing server to execute the method according to the control of the executable instruction.

According to the technical scheme, the embodiment of the application has the following advantages: in the embodiment of the application, acquiring a cooperative office log of a cooperative office node connected with a target office node to be analyzed and a target office log generated by the target office node in a target service time period, wherein the target office log is an office record generated in the office process of the target office node; respectively analyzing the cooperative office logs and the target office logs to obtain derived quantitative descriptions corresponding to the cooperative office logs and business quantitative descriptions corresponding to the target office logs, wherein the derived quantitative descriptions are used for representing attention-type contents between the target office nodes and the connected cooperative office nodes, and the business quantitative descriptions are used for representing operation preference descriptions presented by the target office nodes in the office process; determining a multi-mode behavior description of the target office node according to the derived quantitative description and the business quantitative description; on the premise that the difference degree between the multi-modal behavior description and the target behavior description of the target office node is not larger than a set threshold value, determining that the target office node corresponds to the target task process, wherein the target behavior description is a multi-modal behavior description of an office node corresponding to the target task process, the office business abnormity existing in the target office node is weighed through the target office log, the attention type business abnormity or the group type business abnormity existing in the target office node is weighed through the cooperative office log, and determines the multi-modal behavior description of the target office node through a comprehensive analysis form, thereby achieving the purpose of improving the richness of the multi-modal behavior description, therefore, the method has the advantages of improving the analysis accuracy of the task process of the office node, and further improving the technical problem of low analysis accuracy of the task process of the office node to a certain extent.

Other features of the present application and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the application and together with the description, serve to explain the principles of the application.

Fig. 1 is a block diagram showing one hardware configuration of a cloud office information processing server that can implement an embodiment of the present application.

Fig. 2 is a block diagram showing another hardware configuration of a cloud office information processing server that can implement an embodiment of the present application.

Fig. 3 is a flowchart illustrating a cloud office information processing method combining AI and big data, which can implement an embodiment of the present application.

Fig. 4 is an architectural diagram illustrating a cloud office information processing system incorporating AI and big data in which embodiments of the present application may be implemented.

Detailed Description

< hardware configuration >

Fig. 1 is a block diagram illustrating a hardware configuration of a cloud office information processing server 100 that can implement an embodiment of the present application, where the cloud office information processing server 100 may include a processor 110 and a memory 120, the memory 120 is used for storing executable instructions, and the processor 110 is used for operating the cloud office information processing server 100 to execute a cloud office information processing method combining AI and big data in the present application according to the control of the executable instructions.

Fig. 2 is a block diagram illustrating another hardware configuration of the cloud office information processing server 100 that can implement an embodiment of the present application, where the cloud office information processing server 100 may include a processor 110, a memory 120, and a cloud office information processing apparatus 400 that combines an AI and big data, where the cloud office information processing apparatus 400 that combines an AI and big data includes at least one software function module that may be stored in the memory 120 in the form of software or firmware (firmware), and the processor 110 executes various function applications and data processing by running software programs and modules stored in the memory 120, for example, the cloud office information processing apparatus 400 that combines an AI and big data in the embodiment of the present application, so as to implement a cloud office information processing method that combines an AI and big data in the embodiment of the present application.

< method examples >

Fig. 3 is a flowchart illustrating a cloud office information processing method combining AI and big data, which can implement an embodiment of the present application, and the cloud office information processing method combining AI and big data can be implemented by the cloud office information processing server 100 shown in fig. 1 or fig. 2, and further can include the technical solutions described in the following steps.

S202, obtaining a cooperative office log of a cooperative office node connected with a target office node to be analyzed, and a target office log generated by the target office node in a target service time period, wherein the target office log is an office record generated in the office process of the target office node.

And S204, analyzing the cooperative office log and the target office log respectively to obtain a derivative quantitative description corresponding to the cooperative office log and a business quantitative description corresponding to the target office log, wherein the derivative quantitative description is used for representing the attention type content between the target office node and the connected cooperative office node, and the business quantitative description is used for representing the operation preference description presented by the target office node in the office process.

And S206, determining the multi-mode behavior description of the target office node according to the derivative quantitative description and the business quantitative description.

And S208, on the premise that the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to the target task progress, wherein the target behavior description is the multi-modal behavior description of the office node corresponding to the target task progress.

For some possible examples, in the embodiment of the present application, the cloud office information processing method combining AI and big data may be applied, but not limited to, in a visual interaction architecture of an office user risk level, for example, a cooperative office log of a cooperative office node connected to a target office node to be analyzed is obtained to balance attention-type business anomaly or group-type business anomaly of the target office node, a target office log generated by the target office node in a target business period is obtained to balance business anomaly of the target office node, so as to construct a more comprehensive and rich office user behavior description, and compare the office user behavior description with an office node corresponding to a target task progress, and if a similarity coefficient of the comparison is not greater than a set threshold, an office node task progress of the target office node may be represented as a target task progress (an abnormal state or a risk state), otherwise, the office node task process of the target office node is judged to be in a normal state (risk-free state).

It is understood that, in the embodiment of the present application, the cooperative office node to which the target office node is connected may refer to, but is not limited to, an associated cooperative office node, a service cooperative office node, a matching type cooperative office node, and the like. The associated cooperative office node is a cooperative office node which is configured and registered by using an office communication address same as that of the target office node, the business cooperative office node is a cooperative office node which generates office business interaction behaviors or generates a target office log with the target office node, and the matching type cooperative office node is a cooperative office node or an offline cooperative office node which has a matching type relationship (such as a binding relationship or a pairing relationship) with the target office node;

Viewed from some exemplary perspective, the target task process may include, but is not limited to, a variety of office exception categories, such as focus-type business exceptions or group-type business exceptions, office privilege exceptions, and the like.

For some possible examples, assume that the collaborative office node to which the target office node is connected is a matching type collaborative office node, the corresponding collaborative office log may be but is not limited to relevant office record information for a matching type collaborative office node, such as the status, number, etc., of the matching-type cooperative office nodes, for example, all the matching-type cooperative office nodes officeone 304 of the determination target office node officeone 302, further obtain the cooperative office logs (such as office interaction frequency, office file amount, etc.) of the matching type cooperative office node officeon 304 and the target office node officeon 302, or the relevant office record information (such as abnormality detection record, reference service/historical office service, etc.) of the matching type collaborative office node officeon 304 is taken as a collaborative office log, the number of the matching type collaborative office node officeon 304 can also be acquired, to evaluate the concerned business abnormality or the group business abnormality of the target cooperative office node officeon 302.

In this embodiment of the present application, the target office log may be, but is not limited to, office records of multiple categories generated in an office process of the target office node, where the categories may include information categories or service categories, for example, the information categories may be divided into event information and interaction records, and the service categories may be further divided according to different level descriptions, for example, the target office log whose service information amount reaches an information amount threshold is determined as a first service category, or the target office log whose office interaction frequency reaches a frequency threshold is determined as a second service category, and the like, which is not limited herein.

In the embodiment of the present application, the target behavior description may be, but not limited to, a multi-modal behavior description (for example, may be understood as a user portrait or a portrait feature) that is already defined as an office node corresponding to the target task process, or may be, but not limited to, a multi-modal behavior description that is estimated as an office node corresponding to the target task process, and an office node corresponding to the target task process is taken as an abnormal office node, assuming that a first AI model (supervised model) is trained by using a sample labeled as whether the office node is an abnormal office node, so as to train to obtain a multi-modal behavior description of the abnormal office node (for example, an office node with risk), and then a second AI model (unsupervised model) is trained by using a sample not labeled as whether the office node is an abnormal office node, for example, so as to train to obtain a multi-modal behavior description that meets the setting requirements as an abnormal office node, wherein, the setting requirement may include but is not limited to at least one of the following: the population quantization (aggregation level) is higher than a first set threshold, the global occupancy (total number of shares) is lower than a second set threshold, the risk value is higher than a third set threshold, etc.

It can be understood that a cooperative office log of a cooperative office node connected to a target office node to be analyzed and a target office log generated by the target office node in a target service time period are obtained, where the target office log is an office record generated in an office process of the target office node; respectively analyzing the cooperative office logs and the target office logs to obtain derived quantitative descriptions corresponding to the cooperative office logs and business quantitative descriptions corresponding to the target office logs, wherein the derived quantitative descriptions are used for representing attention-type contents between the target office nodes and the connected cooperative office nodes, and the business quantitative descriptions are used for representing operation preference descriptions presented by the target office nodes in the office process; determining multi-modal behavior description of the target office node according to the derived quantitative description and the business quantitative description; and on the premise that the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to the target task process, wherein the target behavior description is the multi-modal behavior description of the office node corresponding to the target task process. Furthermore, a derived quantitative description may be understood as an associated feature value or an associated description value.

For some possible examples, alternative detection modes of task progress of the office node are assumed as follows. Acquiring a collaborative office log journel 504 of a collaborative office node connected with a target office node to be analyzed, and a target office log journel 506 generated by the target office node office 502 in a target service time interval, determining a multi-modal behavior description508 of the target office node office 502 according to a derived quantitative description and a service quantitative description under the premise of acquiring the derived quantitative description corresponding to the collaborative office log journel 504 and the service quantitative description corresponding to the target office log journel 506, and further completing process identification of the target office node office 502 according to the multi-modal behavior description508, wherein the identification result is used for representing whether the target office node office 502 corresponds to a target task process, or identifying and analyzing what target task process the target office node office 502 corresponds to on the premise of corresponding to the target task process.

When the technical scheme described in the embodiment of the application is applied, acquiring a cooperative office log of a cooperative office node connected with a target office node to be analyzed and a target office log generated by the target office node in a target service time period, wherein the target office log is an office record generated in the office process of the target office node; analyzing the cooperative office log and the target office log respectively to obtain a derivative quantitative description corresponding to the cooperative office log and a business quantitative description corresponding to the target office log, wherein the derivative quantitative description is used for representing concerned contents between the target office node and the connected cooperative office node, and the business quantitative description is used for representing an operation preference description presented by the target office node in the office process; determining multi-modal behavior description of the target office node according to the derived quantitative description and the business quantitative description; on the premise that the difference between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to the target task process, wherein the target behavior description is the multi-modal behavior description of the office node corresponding to the target task process, balancing office business abnormity existing in the target office node through a target office log, balancing attention type business abnormity or group type business abnormity existing in the target office node through a cooperative office log, and determining the multi-modal behavior description of the target office node through a comprehensive analysis form, so that the aim of improving the richness of the multi-modal behavior description is fulfilled, and the beneficial effect of improving the analysis accuracy of the office node task process is achieved.

For some independently implementable technical solutions, determining the multi-modal behavior description of the cooperative office node connected to the target office node according to the derived quantitative description and the business quantitative description may include the following contents.

And S1, integrating the derived quantitative description and the business quantitative description to obtain a target quantitative description.

And S2, determining the target quantization description as the multi-modal behavior description of the target office node.

Or, S3, determining the derived quantitative description as the first behavioral description of the target office node.

And S4, determining the business quantification description as a second behavior description of the target office node, wherein the multi-modal behavior description of the target office node comprises the first behavior description and the second behavior description.

It is understood that, in the embodiment of the present application, in consideration of flexibility of processing of the behavior description, how to determine the multi-modal behavior description of the target office node may be decided by itself, for example, the multi-modal behavior description may be, but is not limited to, an integrated feature of the derived quantitative description and the business quantitative description, and may also include a first behavior description and a second behavior description corresponding to the derived quantitative description and the business quantitative description that are not affected by each other. The implementation has the advantages that in the subsequent process of identifying and analyzing the task process of the office node, the identification and analysis of the behavior description of the whole can improve the analysis accuracy, and the identification and analysis of two behavior descriptions which are not influenced mutually can avoid data information loss in the integration process, so that the accuracy of the identification and analysis of the task process of the office node is improved.

It can be understood that, the derived quantization descriptions and the business quantization descriptions are integrated (fused) to obtain a target quantization description; determining the target quantitative description as a multi-modal behavior description of the target office node; or, determining the derived quantitative description as a first behavior description of the target office node; and determining the business quantification description as a second behavior description of the target office node, wherein the multi-modal behavior description of the target office node comprises the first behavior description and the second behavior description. The behavior description may be a feature vector or a feature map.

When the technical scheme described in the embodiment of the application is applied, the derived quantization description and the business quantization description are integrated to obtain a target quantization description; determining the target quantization description as a multi-modal behavior description of the target office node; or, determining the derived quantitative description as a first behavior description of the target office node; and determining the service quantitative description as a second behavior description of the target office node, wherein the multi-modal behavior description of the target office node comprises the first behavior description and the second behavior description, so that the aim of flexibly acquiring the multi-modal behavior description is fulfilled, and the effect of improving the flexibility of acquiring the multi-modal behavior description is realized.

For some independently implementable technical solutions, determining that the target office node corresponds to the target task process may include the following:

and Step1, under the premise that the target quantization description is determined to be the multi-modal behavior description of the target office node, and the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to the target task progress.

Or the like, or a combination thereof,

and Step2, on the premise that the derived quantitative description is determined to be a first behavior description, and the difference degree between the first behavior description and a first target description is not larger than a first preset difference degree, determining that the target office node corresponds to a first target task progress, wherein the first target description is attention-type content between the target office node and a connected cooperative office node.

And Step3, on the premise that the business quantification description is determined to be a second behavior description, and the difference between the second behavior description and a second target description is not greater than a second preset difference, determining that the target office node corresponds to a second target task progress, wherein the second target description is an operation preference description presented by the target office node in the office process. Wherein the operation preference description can be understood as an operation behavior feature.

In the embodiment of the application, a corresponding office node task progress detection method is determined according to different behavior description acquisition modes, for example, a target quantization description is determined as a multi-modal behavior description of a target office node, a difference degree between the target quantization description and a first target description is compared, and if the difference degree is not greater than a first preset difference degree, the target quantization description is considered to be close to the first target description. In addition, on the premise that the difference between the first behavior description and the first target description is not greater than the first preset difference, further according to the size of the difference, more detailed information corresponding to the first target task process of the target office node is determined, for example, the first target task process may be further refined into different evaluation levels, and the evaluation levels are inversely or positively correlated with the difference, so that a more easily understood and detailed recognition result may be provided to the user.

It can be understood that, for example, on the premise that the derived quantitative description is determined as the first behavioral description, the difference between the first behavioral description and the second target description is compared, and if the difference is not greater than the second preset difference, the derived quantitative description is considered to be close to the second target description, that is, it indicates that there may be an attention-type business abnormality or a group-type business abnormality in the target office node, and it is determined that the target office node corresponds to the attention-type business abnormality or the group-type business abnormality (the first target task process); for example, on the premise that the business quantitative description is determined as the second behavior description, the difference degree between the second behavior description and the third target description is compared, if the difference degree is not greater than the third preset association degree, the business quantitative description is considered to be close to the third target description, that is, the business abnormality may exist in the target office node, and the target office node is determined to correspond to the business abnormality (the second target task process).

It can be understood that on the premise that the target quantization description is determined as the multi-modal behavior description of the target office node, and the difference between the multi-modal behavior description of the target office node and the target behavior description is not greater than the set threshold, it is determined that the target office node corresponds to the target task process; or, on the premise that the derived quantitative description is determined to be a first behavior description and the difference between the first behavior description and a first target description is not greater than a first preset difference, determining that the target office node corresponds to a first target task progress, wherein the first target description is attention-type content between an office node corresponding to the target task progress and a connected cooperative office node; and on the premise that the business quantification description is determined as a second behavior description and the difference degree between the second behavior description and a second target description is not larger than a second preset difference degree, determining that the target office node corresponds to a second target task progress, wherein the second target description is an operation preference description presented by the office node corresponding to the target task progress in the office process.

When the technical scheme described in the embodiment of the application is applied, on the premise that the target quantization description is determined to be the multi-modal behavior description of the target office node, and the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to the target task progress; or, on the premise that the derived quantitative description is determined as a first behavior description and the difference between the first behavior description and a first target description is not greater than a first preset difference, determining that the target office node corresponds to a first target task process, wherein the first target description is attention-type content between the office node corresponding to the target task process and the connected cooperative office node; and on the premise that the business quantitative description is determined to be the second behavior description and the difference degree between the second behavior description and the second target description is not greater than the second preset difference degree, determining that the target office node corresponds to the second target task process, wherein the second target description is the operation preference description presented by the office node corresponding to the target task process in the office process, so that the aim of flexibly identifying and analyzing the office node task process is fulfilled, and the effect of improving the flexibility of identifying and analyzing the office node task process is achieved.

For some independently implementable technical solutions, the parsing is performed on the cooperative office log and the target office log respectively to obtain a derived quantitative description corresponding to the cooperative office log and a business quantitative description corresponding to the target office log, which may include the following contents.

Step 1, on the premise that the target office log comprises a business interaction record, analyzing the business interaction record to obtain a quantitative description corresponding to the business interaction record, wherein the business interaction record is used for representing the interaction record generated by the target office node in the office process, and the business quantitative description comprises a quantitative description corresponding to the business interaction record.

And 2, analyzing the business item information on the premise that the target office log comprises the business item information to obtain quantitative description corresponding to the business item information, wherein the business interaction record is used for representing the item information generated by the target office node in the office process, and the business quantitative description comprises the quantitative description corresponding to the business item information.

And 3, analyzing the office records of the matched type cooperative office node on the premise that the cooperative office log comprises the office records of the matched type cooperative office node connected with the target office node to obtain quantitative descriptions corresponding to the office records of the matched type cooperative office node, wherein the business quantitative descriptions comprise the quantitative descriptions corresponding to the office records of the matched type cooperative office node.

And 4, on the premise that the cooperative office log comprises office records of the business cooperative office node connected with the target office node, analyzing the office records of the business cooperative office node to obtain quantitative descriptions corresponding to the office records of the business cooperative office node, wherein the business quantitative descriptions comprise quantitative descriptions corresponding to the office records of the business cooperative office node.

It is understood that in the embodiment of the present application, the service interaction record may be, but is not limited to, an office record generated for an office user profile transmission behavior, and the office user profile transmission behavior is exemplarily divided into two directions, i.e., a profile transmission direction and a profile reception direction.

It is understood that, in the embodiment of the present application, the business transaction information may be, but is not limited to, corresponding visual information (e.g., text information) accompanying in an office process, for example, in various office business environments of cloud office, occurrence of an office user profile transmission behavior is accompanied by corresponding business visual information (e.g., profile text).

In the embodiment of the present application, in order to prevent invalid association or invalid matching binding (that is, malicious association or too many office node devices that are bound), the office records of the matching type collaborative office nodes may be used, but not limited to, to detect whether the target office node has an abnormal risk of invalid association or invalid matching binding, for example, assuming that the office records of the matching type collaborative office nodes are the number of the matching type collaborative office nodes, and then on the premise that the number of the matching type collaborative office nodes is greater than or equal to the number threshold, it may be determined that the target office node has an attention-type service abnormality or a group-type service abnormality, and the like.

In the embodiment of the application, the cloud office information processing method combining the AI and the big data can be used for performing identification analysis on the business cooperation office node to obtain an identification result of the office node task process of the business cooperation office node, and then the identification result is used as an office record of the business cooperation office node to improve the analysis accuracy. In addition, in order to improve the efficiency of the identification analysis, the severity of the identification analysis of the business cooperative office node may also be, but not limited to, slowed down, for example, only the target office log of the business cooperative office node is identified and analyzed, but not the business cooperative office log of the business cooperative office node, or only the business cooperative office log of the business cooperative office node is identified and analyzed, but not the target office log of the business cooperative office node.

It can be understood that, on the premise that the target office log includes the service interaction record, the service interaction record is analyzed to obtain a quantitative description corresponding to the service interaction record, wherein the service interaction record is used for representing the interaction record generated by the target office node in the office process, and the service quantitative description includes a quantitative description corresponding to the service interaction record; analyzing the business item information on the premise that the target office log comprises the business item information to obtain quantitative description corresponding to the business item information, wherein the business interaction record is used for representing the item information generated by the target office node in the office process, and the business quantitative description comprises the quantitative description corresponding to the business item information; analyzing the office records of the matched type cooperative office nodes on the premise that the cooperative office logs comprise the office records of the matched type cooperative office nodes connected with the target office nodes to obtain quantitative descriptions corresponding to the office records of the matched type cooperative office nodes, wherein the business quantitative descriptions comprise quantitative descriptions corresponding to the office records of the matched type cooperative office nodes; on the premise that the cooperative office log includes office records of the business cooperative office node connected with the target office node, the office records of the business cooperative office node are analyzed to obtain quantitative descriptions corresponding to the office records of the business cooperative office node, wherein the business quantitative descriptions include quantitative descriptions corresponding to the office records of the business cooperative office node.

When the technical scheme described in the embodiment of the application is applied, on the premise that the target office log comprises the business interaction record, the business interaction record is analyzed to obtain quantitative description corresponding to the business interaction record, wherein the business interaction record is used for representing the interaction record generated by the target office node in the office process, and the business quantitative description comprises the quantitative description corresponding to the business interaction record; on the premise that the target office log comprises the business item information, analyzing the business item information to obtain quantitative description corresponding to the business item information, wherein the business interaction record is used for representing the item information generated by the target office node in the office process, and the business quantitative description comprises quantitative description corresponding to the business item information; analyzing the office records of the matched type cooperative office nodes on the premise that the cooperative office logs comprise the office records of the matched type cooperative office nodes connected with the target office nodes to obtain quantitative descriptions corresponding to the office records of the matched type cooperative office nodes, wherein the business quantitative descriptions comprise quantitative descriptions corresponding to the office records of the matched type cooperative office nodes; on the premise that the cooperative office log comprises office records of the business cooperative office nodes connected with the target office node, the office records of the business cooperative office nodes are analyzed to obtain quantitative descriptions corresponding to the office records of the business cooperative office nodes, wherein the business quantitative descriptions comprise quantitative descriptions corresponding to the office records of the business cooperative office nodes, the purpose of improving the identification and analysis precision of the task process of the office nodes is achieved, and the effect of improving the analysis accuracy of the task process of the office nodes is achieved.

For some independently implementable technical solutions, parsing the service transaction information to obtain a quantitative description corresponding to the service transaction information may include the following contents.

And step 100, cleaning noise information in the service event information, wherein the noise information is visual information used for representing that the association degree with the service content does not meet a third preset association degree in the service event information.

And 200, extracting significance information in the cleaned service item information, wherein the significance information is visual information used for representing that the association degree with the service content in the service item information reaches a fourth preset association degree.

And step 300, analyzing the significance information to obtain quantitative description corresponding to the business item information.

It can be understood that noise information (invalid information or interference information) in the service event information is cleaned (filtered), where the noise information is visual information (text information) in the service event information for representing that the association degree with the service content does not satisfy the third preset association degree; extracting significance information (key information) in the cleaned business item information, wherein the significance information is visual information which is used for representing that the association degree of business contents reaches a fourth preset association degree in the business item information; and analyzing the significance information to obtain the quantitative description corresponding to the business item information.

When the technical scheme described in the embodiment of the application is applied, noise information in the service transaction information is cleaned, wherein the noise information is visual information used for representing that the association degree of the service transaction information and the service content does not meet a third preset association degree; extracting significance information in the cleaned business item information, wherein the significance information is visual information used for representing that the association degree with business contents reaches a fourth preset association degree in the business item information; the significance information is analyzed to obtain the quantitative description corresponding to the business item information, so that the aim of quickly obtaining the quantitative description corresponding to the business item information by using efficient means such as cleaning and extraction is fulfilled, and the effect of improving the mining efficiency of the quantitative description corresponding to the business item information is achieved.

For some independently implementable technical solutions, parsing the significance information to obtain a quantitative description corresponding to the business event information may include the following contents.

(1) And analyzing the significance information to obtain a quantitative description corresponding to the significance information.

(2) And determining the quantitative description corresponding to the significance information as the quantitative description corresponding to the business event information.

Or, (3) matching the significance information with the target information to obtain a matching result of the significance information and the target information, wherein the target information is visual information used for representing that the association degree with the abnormal business content reaches a fifth set association degree.

(4) And analyzing the pairing result to obtain the quantitative description corresponding to the pairing result.

(5) And determining the quantitative description corresponding to the pairing result as the quantitative description corresponding to the service item information.

It can be understood that the significance information is parsed to obtain a quantitative description corresponding to the significance information; determining the quantitative description corresponding to the significance information as the quantitative description corresponding to the business item information; or, matching the significance information with the target information to obtain a matching result of the significance information and the target information, wherein the target information is visual information used for representing that the association degree with the abnormal business content reaches a fifth set association degree; analyzing the pairing result to obtain quantitative description corresponding to the pairing result; and determining the quantitative description corresponding to the pairing result as the quantitative description corresponding to the service item information. Alternatively, the abnormal business may include, but is not limited to, office file stealing, office data tampering, and the like.

For some possible examples, optionally, gambling is taken as an example of abnormal business, and on the premise that a periodically updated gambling business seed visualization word bank exists after reference gambling audit and gambling model deposition, the user monthly business visualization and the reference abnormal visualization are matched in a user unit to detect whether gambling seed visualization, visual evaluation level and visual source model are hit;

when the technical scheme described in the embodiment of the application is applied, the significance information is analyzed to obtain quantitative description corresponding to the significance information; determining the quantitative description corresponding to the significance information as the quantitative description corresponding to the business item information; or, matching the significance information with the target information to obtain a matching result of the significance information and the target information, wherein the target information is visual information used for representing that the association degree with the abnormal business content reaches a fifth set association degree; analyzing the pairing result to obtain quantitative description corresponding to the pairing result; the quantitative description corresponding to the pairing result is determined as the quantitative description corresponding to the service item information, so that the aim of determining the quantitative description corresponding to the service item information by using the pairing result with higher precision is fulfilled, and the effect of improving the accuracy of the quantitative description corresponding to the service item information is achieved.

For some independently implementable technical solutions, after acquiring the cooperative office log of the cooperative office node connected to the target office node to be analyzed and the target office log generated by the target office node in the target service time period, the following contents may be included: importing the collaborative office log and the target office log into an AI neural network, wherein the AI neural network is a machine learning model for analyzing data information, which is obtained by performing feature differentiation analysis on a plurality of training samples; obtaining an analysis result output by the AI neural network, wherein the analysis result is a characteristic differentiation analysis result of the collaborative office log and the target office log; and determining the office node task process to which the target office node belongs according to the analysis result.

It can be understood that the collaborative office log and the target office log are led into an AI neural network, wherein the AI neural network is a machine learning model for analyzing data information, which is obtained by performing feature differentiation analysis on a plurality of training samples; obtaining an analysis result output by the AI neural network, wherein the analysis result is a characteristic differentiation analysis result of the collaborative office log and the target office log; and determining the office node task process to which the target office node belongs according to the analysis result.

For some possible examples, optionally, on the premise that the multi-modal behavior descriptions corresponding to the collaborative office log and the target office log are obtained, clustering is performed on the multi-modal behavior descriptions by using a secondary feature differentiation analysis policy (such as a two-step clustering algorithm).

When the technical scheme described in the embodiment of the application is applied, the cooperative office log and the target office log are led into an AI neural network, wherein the AI neural network is a machine learning model for analyzing data information, which is obtained by performing feature differentiation analysis by using a plurality of training samples; obtaining an analysis result output by the AI neural network, wherein the analysis result is a characteristic differentiation analysis result of the collaborative office log and the target office log; and determining the office node task process to which the target office node belongs according to the analysis result, so that the aim of quickly determining the office node task process to which the target office node belongs by using an unsupervised office record identification model is fulfilled, and the effect of improving the efficiency of determining the office node task process to which the target office node belongs is realized.

For some independently implementable technical solutions, before importing the collaborative office log and the target office log into the AI neural network, the method includes: acquiring a plurality of training samples, wherein the training samples comprise sample collaborative office logs and sample target office logs; circularly implementing the following operations until an AI neural network is obtained; determining a current training sample from a plurality of training samples, and determining a current AI neural network, wherein the current training sample comprises a current sample collaborative office log and a current sample target office log; identifying a current characteristic differentiation analysis result of a current training sample through a current AI neural network; on the premise that the current characteristic differentiation analysis result does not meet the analysis judgment index, acquiring a next training sample as a current training sample; and determining the current AI neural network as the AI neural network on the premise that the current characteristic differentiation analysis result reaches the analysis judgment index.

Optionally, in the embodiment of the present application, since the AI neural network is an unsupervised model architecture, and the second AI model architecture can, but is not limited to, divide data into clusters based on shared description, and does not need to perform data information calibration, that is, a sample that does not need high-quality calibration is represented, compared with an additional office record calibration process experienced by the supervised model architecture, the second AI model architecture is more matched with an office node task process detection and analysis environment, which not only ensures a certain analysis accuracy, but also can ensure higher training efficiency, and thus improve the overall analysis and recognition efficiency of the office node task process.

It can be understood that a plurality of training samples are obtained, wherein the training samples include a sample collaborative office log and a sample target office log; the following operations are performed in a loop until an AI neural network is obtained: determining a current training sample from a plurality of training samples, and determining a current AI neural network, wherein the current training sample comprises a current sample collaborative office log and a current sample target office log; identifying a current characteristic differentiation analysis result of a current training sample through a current AI neural network; on the premise that the current feature differentiation analysis result does not meet the analysis judgment index, acquiring a next training sample as a current training sample; and on the premise that the current characteristic differentiation analysis result reaches an analysis judgment index, determining the current AI neural network as the AI neural network.

When the technical scheme described in the embodiment of the application is applied, a plurality of training samples are obtained, wherein the training samples comprise sample collaborative office logs and sample target office logs; the following operations are performed in a loop until an AI neural network is obtained: determining a current training sample from a plurality of training samples, and determining a current AI neural network, wherein the current training sample comprises a current sample collaborative office log and a current sample target office log; identifying a current characteristic differentiation analysis result of a current training sample through a current AI neural network; on the premise that the current feature differentiation analysis result does not meet the analysis judgment index, acquiring a next training sample as a current training sample; on the premise that the current characteristic differentiation analysis result reaches the analysis and judgment index, the current AI neural network is determined to be the AI neural network, so that the aim of efficiently training the AI neural network is fulfilled, and the effect of improving the overall recognition and analysis efficiency of the office node task process is achieved.

For some independently implementable technical solutions, before determining that the target office node corresponds to the target task process, the following may be further included: acquiring configuration data and reference data of a target office node, wherein the configuration data is used for representing office node configuration information of the target office node, and the reference data is used for representing reference keyword information of the target office node; the configuration data and the reference data are respectively analyzed to obtain configuration quantitative descriptions corresponding to the configuration data and reference quantitative descriptions corresponding to the reference data, wherein the configuration quantitative descriptions are used for representing key configuration contents corresponding to the target office node, and the reference quantitative descriptions are used for representing keyword description contents previously distributed by the target office node; and determining office node task processes to which the target office nodes belong according to the configuration quantitative description, the reference quantitative description, the derivation quantitative description and the business quantitative description, wherein the office node task processes comprise the target task processes.

In the embodiment of the present application, the configuration data may include, but is not limited to, at least one of the following: office node evaluation level, office node distinguishing label, etc. The reference data may include, but is not limited to, at least one of: reference evaluation (history evaluation) of the office node, a reference progress recognition result of the office node, and the like.

In the embodiment of the application, multiple levels of configuration quantitative description, reference quantitative description, service visualization, attention-type service exception or group-type service exception of the target office node can be combined, so as to determine a more comprehensive and richer multi-modal behavior description for the target office node.

It can be understood that configuration data and reference data of the target office node are obtained, wherein the configuration data is used for representing office node configuration information of the target office node, and the reference data is used for representing reference keyword information of the target office node; the configuration data and the reference data are respectively analyzed to obtain configuration quantitative descriptions corresponding to the configuration data and reference quantitative descriptions corresponding to the reference data, wherein the configuration quantitative descriptions are used for representing key configuration contents corresponding to the target office node, and the reference quantitative descriptions are used for representing keyword description contents previously distributed by the target office node; and determining office node task processes to which the target office nodes belong according to the configuration quantization description, the reference quantization description, the derivative quantization description and the business quantization description, wherein the office node task processes comprise the target task processes.

In some independently implementable technical solutions, after determining the multi-modal behavior description of the target office node according to the derived quantitative description and the traffic quantitative description, the method further includes: determining an office guidance message of the target office node when it is determined that the office guidance requirement exists in the target office node based on the multi-modal behavior description.

In some embodiments, determining the office guidance message of the target office node may include the following: acquiring status summary information of a cloud office interaction object in a target office node, and extracting hot spot status content matched with the cloud office interaction object from the status summary information; performing interactive object office habit analysis on the hot spot state content to obtain cloud office habit description used for expressing the cloud office interactive object; when the cloud office interaction object is determined to be in accordance with the office guide index based on the cloud office habit description, acquiring N office interaction assistance projects of cloud office affair categories matched with the cloud office habit description, and transferring office guide messages corresponding to the office interaction assistance projects of the N cloud office affair categories to the target office node so as to output the office guide messages in a cloud office visual interaction unit corresponding to the target office node; and N is a positive integer.

In some independently implementable technical solutions, the technical solutions related to the office guidance message described above can be further implemented by the following technical solutions.

Step S101, acquiring status summary information of the cloud office interaction object in the target office node, and extracting hot spot status content matched with the cloud office interaction object from the status summary information.

In some possible examples, the cloud office information processing server may collect, from an office information database corresponding to the target office node, first office request summary information of the cloud office interaction object in the target office node in a first office period; further, the cloud office information processing server can take the first office request summary information as status summary information of the cloud office interaction object, and can acquire a target summary topic keyword corresponding to the status summary information; further, the cloud office information processing server can extract target hotspot keywords corresponding to the previous cloud office cooperation state of the cloud office interaction object in the first office period from the state summary information; further, the cloud office information processing server can determine the hot spot state content matched with the cloud office interaction object based on the target summary topic keyword and the target hot spot keyword.

Viewed from an exemplary perspective, the status summary information may be an operation log, and correspondingly, the office request summary information may be a log record of device login, and the summary topic keyword may be understood as a title word of the relevant summary information. Further, the hotspot keyword may be understood as a key field of the prior cloud office collaboration state. In addition, the hot spot status content may be understood as office activity data or office activity information.

In some possible embodiments, in a cloud office large environment, the target hotspot keywords extracted from the status summary information of the cloud office interaction object may include at least the following keywords: the service capability description information includes object tag information (e.g., IDxxx) of the cloud office interaction object, member portrayal of the cloud office interaction object in the target office node (e.g., office business level participant name), attribute information (e.g., identity feature information) of the cloud office interaction object, authority emphasis degree (e.g., authority level 3) of the cloud office interaction object in the target office node, and service capability description (e.g., rank 8) of the cloud office interaction object in the target office node.

In some possible embodiments, the target office node in the embodiment of the present application may be exemplified by a collaboration-type target office node, for example, the collaboration-type target office node may be an intelligent device corresponding to an "online office meeting item XXX", and the intelligent device corresponding to the "online office meeting item XXX" may be the above-mentioned target intelligent device. It can be understood that, in the embodiment of the present application, the cloud office information processing server may be a background server of the target office node, and the target office node may perform cooperative operation with regional office equipment corresponding to the cloud office interaction object.

In some possible embodiments, the status summary information of the cloud office interaction object (for example, the cloud office interaction object targetA) collected by the cloud office information processing server from the office information database may be status summary information message-a, status summary information messages-b, …, and status summary information message-n. The status collection step (e.g., in min) corresponding to the status summary information may form a time queue. The status collection step length under the time queue corresponds to a status summary message.

For example, the Timestamp generated by the status summary message-a may be later than the Timestamp generated by the status summary message-b. By analogy, the generation Timestamp of the status summary message- (n-1) may be later than the generation Timestamp of the status summary message-n. It is understood that the Timestamp of the generation of the status summary information message-a may be the time2 time node, i.e. the Timestamp of the first use of the target office node within the status detection step 2.

Similarly, the Timestamp generated by the status summary message-b may be the time1 time node, that is, the Timestamp used when the target office node is used for the first time in the status detection step 1. In this way, the Timestamp generated by other status summarization information may be the Timestamp used by the target office node for the first time within the corresponding status collection step for the cloud office interaction object (e.g., cloud office interaction object targetA), which is not listed here. Based on this, the first office request summary information of the cloud office interaction object in the first office period collected by the cloud office information processing server from the office information database may be collectively regarded as the above status summary information, and the number of the collected first office request summary information is not limited herein.

For convenience of understanding, in the embodiment of the present application, the status summary information in the first office period, which is acquired from the office information database by the cloud office information processing server, is taken as the above status summary information message-a and status summary information message-b, so as to describe a further process of determining hot status content from the status summary information. Further, the following is a scenario trend of status summary information provided in the embodiments of the present application. The status summary information may be status summary information message-b, that is, the status summary information may be used to count the generated collaboration status between the cloud office interaction object and the target office node in the previous status collection step (i.e., within the status detection step 1).

For example, the keyword KEY-A1 may be the summary topic of the status summary information described above. For example, in a feedback scenario (user reflow scenario), the keyword KEY-a1 may be a keyword "ClientRegister", and at this time, the keyword "ClientRegister" set forth in this embodiment may be an aggregation subject described in the status aggregation information generated by the cloud office interaction object in the status detection step 1. In addition, the keyword KEY-B1 may be a keyword of the status summary information, for example, in a feedback scenario (user reflow scenario), the keyword KEY-B1 may be a keyword "OfficeServiceID", in which case the keyword "OfficeServiceID" set forth in the embodiment of the present application may be a status space description recorded in the status summary information generated by the cloud office interaction object in the status detection step1, and so on, which are not listed herein.

It can be understood that after the cloud office information processing server acquires the status summary information, the status summary information may be sorted to select a keyword in the status summary information, where the keyword is matched with a key map unit (office node) that needs to be assisted and guided. For example, the keywords of the key map units that need to be assisted and guided may be identified in the status summary information in advance according to the requirement category, and then the cloud office information processing server may be helped to make sure which keywords (which may be understood as keywords or fields corresponding to office matters or office events) corresponding to the collaboration status in the status summary information need to be assisted and guided, so that the cloud office information processing server may be helped to quickly extract the hotspot status content corresponding to the corresponding requirement category from the status summary information. In other words, in the embodiment of the present application, in the extracted keywords in the status summary information, the keyword that is matched with the corresponding requirement category and carries the identifier may be further used as the hotspot keyword, and the hotspot keyword set forth in the embodiment of the present application is the hotspot status content. For example, in a feedback scenario, the hotspot status content corresponding to the feedback requirement category may specifically include the summary topic keyword "keyword KEY-a 1" and a plurality of hotspot keywords. For example, the keywords KEY-B1, the keywords KEY-B3, the keywords KEY-B5.

Further, it can be understood that, after selecting the summary topic keywords (abbreviated as summary topics) and the keywords matching with the corresponding requirement categories from the status summary information, the cloud office information processing server may further perform index development (interface development) on the summary topics and the keywords provided in the status summary information, that is, may perform optimization processing (for example, may perform encapsulation) on the hot status contents extracted from the status summary information, so as to obtain a content index (for example, may be a service interface) of the hot status contents corresponding to the feedback requirement categories. It can be understood that, in the embodiment of the present application, one requirement category may be allowed to correspond to one content index, so that when a certain collaboration state implemented by a cloud office interaction object in a target office node activates a corresponding requirement category, the cloud office information processing server may perform automatic triggering quickly.

Further, the cloud office information processing server may deploy the developed content index corresponding to the requirement category to an office habit processing thread (for example, the content index may be understood as an office user tag system) matched with the office information database, and then, in the office habit processing thread, the corresponding hot spot state content of the cloud office interaction object in the first office period may be obtained by using the corresponding content index.

It can be understood that, according to the target demand category in the cloud office knowledge graph (which can be understood as a graph office environment), the embodiment of the application can develop the corresponding content index in advance for the key graph unit which needs to be guided in an auxiliary manner under the cloud office knowledge graph. For example, for a target demand category that is a feedback demand category, the hot status content matching the feedback demand category (e.g., hot status content StateContent 1) may be extracted from the status summary information. For another example, for a target requirement category that is a complex office operation requirement category, hot status content (e.g., hot status content StateContent 2) that matches the complex office operation requirement category may be extracted from the status rollup information. For another example, for a target demand category that is a level-of-minutes positive index prompt demand category (activity prompt demand category), then hotspot status content that matches the level-of-seconds positive index prompt event (e.g., hotspot status content StateContent 3) may be extracted from the status rollup information.

For ease of understanding, the following provides an embodiment of index development according to an index optimization strategy for the embodiments of the present application. For example, the index1 is obtained by performing optimization processing on the hot state content (the hot state content may be the hot state content StateContent 2) according to the index optimization policy corresponding to the complexity office operation event when the target requirement category is the complexity office operation event. For example, the summary topic retrieved from the status summary information by the cloud office information processing server may be the summary topic keyword the 1 in the text paragraph 30a, and the summary topic keyword the 1 in the text paragraph 30a may be collectively regarded as the target summary topic keyword retrieved from the status summary information by the embodiment of the present application. Similarly, the keywords selected by the cloud office information processing server from the status summary information may be hot keywords in the text paragraph 30 a. It is to be appreciated that embodiments of the present application can collectively consider the hotspot keywords in the text paragraph 30a as target hotspot keywords obtained from the status rollup information.

For another example, the index2 may be obtained by optimizing another hot-spot status content (the hot-spot status content may be StateContent 3) according to an index optimization policy corresponding to the active index prompting event of the level of minutes when the target demand type is the active index prompting event of the level of minutes. For example, the summary topic obtained from the status summary information by the cloud office information processing server may be the summary topic keyword theme2 in the text paragraph 30b, and the summary topic keyword theme2 in the text paragraph 30a may be collectively regarded as the target summary topic keyword obtained from the status summary information in the embodiment of the present application. Similarly, the keywords selected by the cloud office information processing server from the status summary information may be hot keywords in the text paragraph 30b, for example, the hot keywords may specifically be a minute-second positive index and the like; and the second-level positive index of the cloud office interaction object in the previous state collection step (e.g., last minute) corresponds to a set positive index interval, which may include a first set positive index (e.g., "0.2") and a second set positive index (e.g., "0.9"). That is, in the embodiment of the present application, the hot keywords in the text paragraph 30b can be collectively regarded as the target hot keywords obtained from the status summary information.

Based on this, the cloud office information processing server may deploy the content indexes (e.g., the index1 and the index 2) of the corresponding requirement categories obtained by research and development to the office habit processing thread matched with the office information database, and then may obtain, in the office habit processing thread, the corresponding hotspot state content of the cloud office interaction object in the first office period by using the corresponding content index. It can be understood that the index optimization strategy according to the embodiment of the present application may substantially include the requirement category, the drop index, the influence on the cloud office of the interactive object, the strategy classification label, the dropped event classification label, the drop frequency, the activation time, the incoming keyword, and the like.

In some possible embodiments, the office habit processing thread refers to an intelligent system which is established for a cloud office interaction object and can be used for collecting, managing, analyzing and utilizing cloud office information, and in the office habit processing thread, the cloud office habit description can be regarded as the priority, so that various collaboration states of the interaction object in the cloud office process can be counted through the cloud office habit description, and states of various office collaboration applications (for example, participation applications) can be processed. In addition, it can be understood that the office habit processing thread is also used to provide various types of AI models (for example, an office habit recognition network and the like required in the following step S102), so as to provide support for subsequent interactive object office habit parsing and multi-level interaction.

It can be understood that, after the optimization of the content index corresponding to the target office event is completed, the hot spot information of the target office event may be configured further based on the operation architecture corresponding to the relevant AI technology in the embodiment of the present application. For example, the cloud office information processing server may respond to the event loading behavior of the corresponding visualization interval of the staff corresponding to the operation architecture in the corresponding configuration interface. For example, during the process of configuring the target office event, the staff may add a corresponding operation log index to the visualization section zone-a of the configuration interface, add a summary information name that can be utilized through the operation log index to the visualization section zone-b of the configuration interface, and write an event identification number of the interaction object status event (i.e., the target office event) in the visualization section zone-c of the configuration interface. Therefore, when the interactive object executes a corresponding cloud office cooperation state in the target office node (for example, the use state of the target office node is used during feedback), the index utilization state corresponding to the feedback requirement category is activated, that is, the electronic device (that is, the cloud office information processing server) corresponding to the relevant AI technology obtains an instruction to be automatically activated, so as to automatically trigger the interactive object, and further, a plurality of rounds of interaction strategies (question and answer mechanisms) can be provided according to the set request response tag, so as to help the cloud office interactive object provide a reliable office question and answer mode in the corresponding cloud office map unit of the target office node.

And S102, carrying out interactive object office habit analysis on the hot spot state content to obtain cloud office habit description for expressing the cloud office interactive object.

In some possible examples, the cloud office information processing server may obtain an office habit identification network corresponding to the target office node from the generated office habit processing thread; further, the cloud office information processing server can transmit the hot spot state content to an office habit identification network corresponding to the target office node, and the office habit identification network performs office habit feature identification on the hot spot state content to obtain an office habit feature identification result corresponding to the hot spot state content; further, the cloud office information processing server can generate a cloud office habit description for expressing the cloud office interactive object based on the interactive object state identifier corresponding to the hot spot state content pointed by the office habit feature recognition result. For example, a cloud office habit description may be understood as an office user representation of a cloud office interaction object.

In some possible embodiments, a further process of obtaining, by the cloud office information processing server, an office habit feature recognition result corresponding to the hot spot status content through the hot spot status content and the office habit recognition network may be described as: the cloud office information processing server can transmit the hot spot state content to the office habit recognition network, and the office habit recognition network explained in the embodiment of the application can be a trained office habit recognition network, and at this time, the cloud office information processing server can extract the state key description (such as behavior feature vector) corresponding to the hot spot state content from the feature mining subnet (feature extractor or convolutional layer) in the office habit recognition network, the status key description may further be passed into a discrepancy analysis subnetwork (such as a fully connected layer or classifier) in the office habit recognition network, and the difference analysis subnet in the office habit recognition network continues to perform office habit feature recognition on the hot spot state content, to derive a similarity (or understood as degree of match/relevance/degree of commonality) between the state key descriptions and the standard habit preference descriptions in the difference analysis sub-network; further, the cloud office information processing server may use, in the similarity, a standard habit preference description having a maximum similarity to the state key description as a target standard habit preference description, and use a template state identifier corresponding to the target standard habit preference description as an interactive object state identifier of the state key description; further, the cloud office information processing server can use the interactive object state identification as an office habit feature recognition result corresponding to the office habit recognition network.

It can be understood that, as time changes, the cloud office habit descriptions of the cloud office interaction objects in the target office nodes change along with the time changes, and therefore, in order to obtain the cloud office habit descriptions capable of accurately describing the cloud office interaction objects, the embodiment of the present application can continuously perform periodic cycle training on the office habit recognition model for performing office habit analysis based on the obtained prior office request form in the second office period, so as to ensure that the latest cloud office habit descriptions are used when performing office guidance in the current first office period.

For convenience of understanding, in the embodiment of the application, the untrained office habit recognition models can be uniformly regarded as the original office habit recognition network, and the trained original office habit recognition network can be uniformly regarded as the cloud office habit description. Based on this, it can be understood that before the cloud office information processing server outputs the cloud office habit description of the cloud office interaction object based on the office habit recognition network, the cloud office information processing server needs to train the original office habit recognition network, so that the trained original office habit recognition network is used as the office habit recognition network for testing the cloud office habit description.

For example, the cloud office information processing server may collect second office request summary information of the cloud office interaction object in a second office period from the office information database; the second office period set forth in the embodiments of the present application may be a previous office period of the first office period; further, the cloud office information processing server may extract a previous summary topic keyword matched with the cloud office interaction object and a previous hotspot keyword matched with the previous summary topic keyword from the second office request summary information; further, the cloud office information processing server can take the prior summary topic keywords and the prior hotspot keywords as training state contents, and then can train the original office habit recognition network in the office habit processing thread based on the training state contents, so that the original office habit recognition network after training is taken as the office habit recognition network.

In some possible embodiments, in the present application embodiment, the cloud office habit description may be formed by a series of interactive object state identification information. For example, for the feedback requirement category, the cloud office habit description set forth in this embodiment may specifically include an interactive object state identifier of a first type, and the interactive object state identifier of the first type may be an acquaintance office user feedback tag. For example, when the status summary information acquired by the cloud office information processing server includes the status summary information message-b corresponding to the status collection step T2 and the status summary information message-a corresponding to the status collection step T1, the cloud office information processing server may indirectly obtain that P status collection steps continuously not using the target office node exist between the two used timestamps when it is determined that the used timestamps corresponding to the two status collection steps have used time intervals greater than or equal to a time interval threshold (for example, 10 min).

For example, when the cloud office information processing server does not detect the use state of the interactive object of the cloud office interactive object using the target office node through the target cloud office authentication data in the P state collection step lengths, the first type of interactive object state identifier is obtained through the office habit recognition network, and then the cloud office habit description for describing the cloud office interactive object can be obtained according to the currently obtained interactive object state identifier, so that the following step S103 can be continuously executed in the following process.

For example, if there are P uninterrupted state collection steps between the usage Timestamp (i.e., the time 2) for the target office node described in the status summary information message-a and the usage Timestamp (i.e., the time 1) for the target office node described in the status summary information message-b, P may be greater than or equal to the preset time interval threshold. In other words, the cloud office information processing server can quickly determine that the cloud office interaction object belongs to the feedback interaction object through the cloud office habit description output by the office habit recognition network (the office habit recognition network can be an AI model in any form), that is, the cloud office information processing server can indirectly analyze the usage duration interval between the time1 time node and the time2 time node through the state summary information, and further can determine whether the P state collection step lengths pointed by the usage duration interval are greater than or equal to the duration interval threshold corresponding to the feedback requirement category, and if not, the cloud office interaction object currently using the target office node is represented as not belonging to the feedback interaction object temporarily; in other examples, if the determination is yes, then the characterization indicates that the cloud office interaction object currently using the target office node temporarily belongs to the feedback interaction object. In some possible embodiments, the preset time interval threshold for the feedback requirement category may be 10min, but of course, the value of the preset time interval threshold may also be improved according to the actual office requirement in the cloud office large environment in the embodiment of the present application, and the embodiments of the present application will not be listed herein one by one.

It can be understood that, the cloud office information processing server can learn by using the corresponding content index in the office habit processing thread: the current usage Timestamp (i.e., time 2) of the target office node used by the cloud office interactive object can be 14 times 3/month/4/2019, and the cloud office information processing server can know that the time2 time node of the target office node used by the cloud office interactive object last time is 14 times 3/month/6/2019 by analyzing the status summary information. At this time, the cloud office information processing server may determine that the usage duration interval between the time1 time node and the time2 time node may be 2 days, which indicates that the target office node has not been used for 2 consecutive days before the target office node is used this time.

In other examples, similarly, taking the time1 time node as 2019, 3, 4, and 14 pointing, and the time2 time node of the cloud office interaction object using the target office node last time is 2019, 3, and 14 pointing, the cloud office information processing server may obtain the hot state content in the state summary information by using the corresponding index in the office habit processing thread, and further may determine, based on the hot state content in the state summary information, that the number of intervals (for example, P) of the state collection step between the time1 time node and the time2 time node is 0 day, which indicates that the target office node is not used in 0 day before the target office node is used this time. In other words, this indicates that the cloud office interaction object has been using the target office node for 2 consecutive days.

In other examples, it can be understood that, for the positive index prompt requirement category, the cloud office habit description set forth in the embodiment of the present application may specifically include the above-mentioned second type of interactive object state identifier, and the second type of interactive object state identifier may be a positive index label in a minute-second scale. Similarly, after the cloud office information processing server obtains the second type of interactive object state identifier through the office habit recognition network, another cloud office habit description for expressing the cloud office interactive object may be generated and obtained based on the second type of interactive object state identifier.

In other examples, for the requirement category of the complex office operation, the cloud office habit description set forth in the embodiment of the present application may specifically include the interactive object state identifier of the third type, and the interactive object state identifier of the third type may be an exception tag of the complex office operation. Similarly, after the cloud office information processing server obtains the third kind of interactive object state identifier through the office habit recognition network, another cloud office habit description for expressing the cloud office interactive object may be generated and obtained based on the third kind of interactive object state identifier.

Step S103, when the cloud office interaction object is determined to be in accordance with the office guide index based on the cloud office habit description, acquiring N office interaction assistance projects of cloud office affair categories which are matched with the cloud office habit description, and transferring office guide messages corresponding to the office interaction assistance projects of the N cloud office affair categories to a target office node so as to output the office guide messages in a cloud office visual interaction unit corresponding to the target office node; n is a positive integer.

In some possible examples, the cloud office habit descriptions set forth in the embodiments of the present application may include one or more of the above-mentioned interactive object state identifier of the first type, the interactive object state identifier of the second type, and the interactive object state identifier of the third type, where the kind of the interactive object state identifiers output by the above-mentioned office habit recognition network is not limited in some possible examples.

For example, if a first type of interactive object state identifier exists in the cloud office habit description, and the first type of interactive object state identifier indicates that the interactive object use state of the cloud office interactive object using the target office node through the target cloud office authentication data is not detected in P uninterrupted state collection step lengths, the cloud office information processing server may determine, as the interactive object feedback state, the use state of the cloud office interactive object using the target office node through the target cloud office authentication data in a next state collection step length (i.e., a current state collection step length) of the P uninterrupted state collection step lengths; p may be greater than or equal to a preset usage duration interval; further, the cloud office information processing server can use the demand type corresponding to the feedback state of the interactive object as a feedback demand type based on the feedback state of the interactive object and the target cloud office identity verification data, and when the feedback demand type is detected to be matched with the target demand type in the office guidance indexes, the cloud office interactive object is determined to be in accordance with the office guidance indexes; further, the cloud office information processing server can acquire an auxiliary project distribution network corresponding to the target office node, acquire N office interaction assistance projects of the cloud office item categories matched with the interactive object state identification of the first category in the cloud office habit description through the auxiliary project distribution network, and transfer office guidance messages corresponding to the office interaction assistance projects of the N cloud office item categories to the target office node.

For convenience of understanding, in the embodiment of the present application, the interactive object state identifier including the first class in the cloud office habit description is taken as an example, so as to describe a further process of dropping the office guidance message to the regional office equipment for outputting. Further, after the cloud office habit description for mining the cloud office interaction object is obtained through step S102, when it is determined that the first type of interaction object state identifier exists in the cloud office habit description, it is determined that the cloud office interaction object uses the use state of the target office node through the target cloud office authentication data at this time as the interaction object feedback state within the current state collection step length (i.e., the last state collection step length of the above-mentioned P uninterrupted state collection step lengths), and then the demand type corresponding to the interaction object feedback state can be used as the feedback demand type according to the interaction object feedback state and the target cloud office authentication data.

In some possible embodiments, if the cloud office information processing server detects that the currently determined feedback requirement category belongs to an auto-trigger event, that is, when it detects that the feedback requirement category is adapted to a target requirement category in office guidance indicators under the auto-trigger event, it may be determined that a cloud office interaction object meets the office guidance indicators, and then when an auxiliary item distribution network for the target office node is obtained, office interaction assistance items of N cloud office item categories that are matched with the interactive object state identifier of the first category may be obtained through the auxiliary item distribution network, and then when office guidance information of the office interaction assistance items of the N cloud office item categories is obtained, the office guidance information may be transferred to corresponding regional office equipment.

In some possible embodiments, when the cloud office information processing server obtains the office interaction assistance items of the N cloud office item categories, the N types of office interaction assistance items may be optimized according to the set output style, and then the optimized office interaction assistance items with the set output style may be used as office guidance messages to be released to the regional office equipment.

Further, when the regional office equipment allows the target office node, a cloud office interaction page (i.e., a cloud office visual interaction unit) of the target office node may be obtained, that is, the cloud office interaction page may be the visual interaction unit100 a. The visual interactive unit100a may contain process activation controls for entering different office processes. For example, the process activation control40a may be used to jump the current cloud office knowledgegraph to an office process corresponding to the office rating list of the target office node. As another example, the process activation control40b may be used to jump the current cloud office knowledgegraph to an office process corresponding to the device cluster of the target office node. For another example, the process activation control40c may be configured to jump the current cloud office knowledgegraph to an office process corresponding to a certain business interaction scenario of the target office node. In addition, the process activation control40d may be configured to jump the current cloud office knowledge graph to an office process corresponding to another service interaction scenario of the target office node, where cloud office processes in the visual interaction unit100a are not listed one by one.

In some possible embodiments, the office interaction assistance items of the N cloud office event categories set forth in the embodiments of the present application may be response contents with a certain timeliness prepared in advance for a cloud office map unit (e.g., cloud office map unit node 1) of the cloud office interaction object at a later time (e.g., time3 time node) according to the latest cloud office habit description obtained by the cloud office information processing server in a set office period (e.g., the first office period), so that, when the cloud office interaction object is at the moment that the cloud office knowledge graph of the target office node is run to the cloud office graph unit node1 (e.g., time3 time node), the office guidance message corresponding to the tested response content can be output to realize the automatic triggering of the office guidance message.

It can be understood that, in the embodiment of the present application, the cloud office interaction object may quickly jump the service scene of the target office node from the current cloud office environment to a help-seeking scene by triggering the office state matched with the office guidance message (i.e., triggering the AI control embedded in the target office node), so that the cloud office interaction object may quickly acquire the office assistance response content required at this time (i.e., time3 time node) in the help-seeking scene.

In some possible embodiments, by adopting the embodiments of the present application, office business requirements of the cloud office interaction object under some key map units in the target office node can be intelligently identified through cloud office habit description, so that interactive object tendencies of the cloud office interaction object under the key map units can be intelligently identified, and further possible auxiliary information (i.e., office interaction assistance items of the N cloud office item categories) can be prepared for the cloud office interaction object under the corresponding cloud office map units in advance based on the identified interactive object tendencies, and then the prepared office interaction assistance items of the N cloud office item categories can be optimized, so that office guidance information obtained through optimization can be subsequently output to the regional office equipment for display.

For example, when the cloud office interaction object targetA performs an activation action for a visual interaction area where an intelligent thread for outputting an office guidance message is located, the regional office equipment may respond to the activation action implemented for the intelligent visual interaction area to jump the business scene of the target office node from the visual interaction unit100a (i.e., a cloud office interaction page in a cloud office environment) to the visual interaction unit200a (i.e., a user list in a help scene). It can be understood that, in the embodiment of the present application, the relevant AI technology may be regarded as an AI control of the cloud office interaction object targetA in the target office node, so as to intelligently provide office interaction assistance projects of different levels for the cloud office interaction object targetA under the corresponding cloud office map unit.

For example, for the cloud office interaction object belonging to the feedback interaction object, dynamically changing office events occurring during the period of time in which the cloud office interaction object does not participate (i.e., the usage duration interval formed by the P state collection steps) may be prepared for the cloud office interaction object, that is, office interaction assistance projects of the plurality of cloud office event categories may be issued in advance for the cloud office interaction object carrying the feedback labels of the acquaintance office users.

For another example, in other examples, if a second type of interactive object state identifier exists in the cloud office habit description, and the second type of interactive object state identifier indicates that the global positive index obtained by processing the first type of office collaboration application by the target cloud office authentication data in the single state collection step size of the cloud office interactive object corresponds to the set positive index interval, the cloud office information processing server may use the use state of the target office node as the first target interactive object state by the cloud office interactive object through the target cloud office authentication data in the subsequent state collection step size of the single state collection step size; the set positive index interval comprises a first set positive index and a second set positive index, and the second set positive index is smaller than the second set positive index; further, the cloud office information processing server can determine a first requirement category corresponding to the first target interaction object state based on the first target interaction object state, the target cloud office identity verification data and the standby indication information matched with the target cloud office identity verification data, and when the first requirement category is matched with a target requirement category in the office guidance index, the cloud office interaction object is determined to meet the office guidance index; further, the cloud office information processing server can acquire an auxiliary project distribution network corresponding to the target office node, acquire, by the auxiliary project distribution network, office interaction assistance projects of N cloud office item categories matched with the interactive object state identification of the second category in the cloud office habit description, and transfer office guidance messages corresponding to the office interaction assistance projects of the N cloud office item categories to the target office node.

Similarly, in other examples, if a third type of interactive object state identifier exists in the cloud office habit description and indicates that the cloud office interactive object has not completed the second type of office cooperation application in the target office node through the target cloud office authentication data yet, the cloud office information processing server may take the interactive object state of reprocessing the second type of office cooperation application as the second target interactive object state when the cloud office interactive object reprocesses the second type of office cooperation application through the target cloud office authentication data; further, the cloud office information processing server can determine a second requirement category corresponding to the second target interaction object state based on the target interaction object state, the target cloud office identity verification data and the standby indication information matched with the target cloud office identity verification data, and when the second requirement category is matched with the target requirement category in the office guidance index, the cloud office interaction object is determined to meet the office guidance index; further, the cloud office information processing server can acquire an auxiliary project distribution network corresponding to the target office node, acquire N office interaction assistance projects of the cloud office item categories matched with the interactive object state identification of the third category in the cloud office habit description by the auxiliary project distribution network, and transfer office guidance messages corresponding to the office interaction assistance projects of the N cloud office item categories to the target office node.

In the embodiment of the application, when the cloud office interaction object corresponds to the cloud office knowledge graph of the target office node, for example, when the cloud office interaction object uses the target office node or executes cloud office cooperation application in the target office node, active and intelligent office guidance can be performed on the cloud office cooperation request device in the cloud office knowledge graph, and then office guidance messages (for example, click guidance, drag guidance, file transmission guidance and the like) of the office interaction assistance projects of the N cloud office item categories can be automatically dropped to the target office node, so that automatic triggering of the office guidance messages is realized. It can be understood that the value of N set forth in the embodiments of the present application may be 1 or multiple values. For example, when the value of N is greater than 1, a targeted multi-level office interaction assistance project may be provided for the cloud office interaction object, and the office interaction assistance project set forth in the embodiment of the present application may describe, by the cloud office information processing server, visual content configured for the cloud office interaction object in advance based on the cloud office habit of the cloud office interaction object. It can be understood that for different cloud office interaction objects, different cloud office habit descriptions can be obtained by mining based on the obtained status summary information of the corresponding cloud office interaction objects, and then when office guidance is performed based on the cloud office habit descriptions, differentiated demand handling can be achieved, so that accuracy of office guidance and business adaptability can be improved.

In some possible embodiments, the cloud office information processing server can be helped to obtain the interactive object state track of the cloud office interactive object by counting the cooperation state data between the cloud office interactive object and the target office node in the target office node, that is, the cloud office information processing server can reversely push the cloud office map unit in which the cloud office interactive object is located in the cloud office knowledge map by counting the cooperation state of the interactive object in the corresponding visual interaction unit, and can further recognize the cooperation tendency of the cloud office interactive object by the cloud office map obtained by reverse derivation, so that the cloud office interactive object can be helped to perform multiple times of globally-associated office session processing in the visual interaction unit, so that when the cloud office information processing server obtains the office interaction assistance application of the cloud office interactive object, the method can flexibly improve the office interaction assistance projects deployed in advance, and further can realize the intelligent improvement of the issued visual information so as to ensure that the improved visual information can be attached to the cooperation tendency of the interaction object as much as possible.

For the convenience of understanding, further provided is an embodiment of interaction between a cloud office information processing server and regional office equipment. When the cloud office information processing server executes step S11, the cloud office operation log of the cloud office interaction object (i.e., the status summary information) may be acquired, and then step S12 may be executed to select a key map unit keyword (i.e., the hot status content) that needs to be assisted and guided from the acquired cloud office operation log. At this time, the cloud office information processing server may perform step S13 to optimize the selected key map unit keywords into an index, and further may perform step S14 to configure a trigger index in the background according to the optimized index (the trigger index set forth in this embodiment may be the above-mentioned office guidance index). In this way, when the regional office equipment performs step S15, various pieces of status information generated in cloud office may be sent to the cloud office information processing server, so that the cloud office information processing server may perform step S14 according to the acquired pieces of status information to determine whether the current cloud office collaboration states of the cloud office interaction object meet the trigger index, and further, when the trigger index is met, perform step S17 to help the cloud office interaction object solve the problem in the corresponding cloud office graph unit in an automatic triggering manner.

In other examples, after the step S14 is executed, when it is determined that the current cloud office collaboration states of the cloud office interaction object do not meet the trigger index, the embodiment of the application may further jump to execute the step S11 to obtain the latest cloud office operation log.

Therefore, the cloud office information processing server can continuously obtain a new cloud office habit description for describing the cloud office interaction object based on the collaboration state information, so that when the cloud office interaction object corresponds to the cloud office knowledge graph of the target office node, for example, when the cloud office interaction object uses the target office node or executes a cloud office collaboration application in the target office node, the cloud office information processing server can perform active and intelligent office guidance for the cloud office collaboration request device in the cloud office knowledge graph, and further can automatically drop office guidance messages (for example, click guidance, drag guidance, file transmission guidance and the like) of the office interaction assistance projects of the N cloud office transaction categories to regional office guidance devices to realize automatic triggering of the office guidance messages, the cloud office interactive object can be provided with a targeted multi-layer office interactive assistance project, the office interactive assistance project set forth in the embodiment of the application can be used for configuring visual contents of different layers for the cloud office interactive object in advance based on new cloud office habit descriptions (namely current cloud office habit descriptions) of the cloud office interactive object for the cloud office information processing server, it can be understood that for different cloud office interactive objects, different cloud office habit descriptions can be obtained based on state summary information, and then when office guidance is performed based on the cloud office habit descriptions, differentiated demand can be met, and therefore the accuracy and business adaptability of office guidance can be improved.

When the technical scheme described in the embodiment of the application is applied, configuration data and reference data of a target office node are obtained, wherein the configuration data is used for representing office node configuration information of the target office node, and the reference data is used for representing reference keyword information of the target office node; the configuration data and the reference data are respectively analyzed to obtain configuration quantitative descriptions corresponding to the configuration data and reference quantitative descriptions corresponding to the reference data, wherein the configuration quantitative descriptions are used for representing key configuration contents corresponding to the target office node, and the reference quantitative descriptions are used for representing keyword description contents previously distributed by the target office node; and determining the office node task process to which the target office node belongs according to the configuration quantitative description, the reference quantitative description, the derivation quantitative description and the service quantitative description, wherein the office node task process comprises the target task process, so that the aim of comprehensively detecting the office node task process is fulfilled, and the effect of improving the richness of office node task process detection is realized.

< System embodiment >

On the basis of the above method embodiment, the embodiment of the present application further provides a system embodiment, that is, a cloud office information processing system combining an AI and big data, please refer to fig. 4, the cloud office information processing system 30 combining an AI and big data may include a cloud office information processing server 100 and an office node 200 that communicate with each other, and the cloud office information processing server 100 and the office node 200 communicate with each other to implement or partially implement the above method.

The above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.

Claims

1. A cloud office information processing method combining AI and big data is applied to a cloud office information processing server and comprises the following steps:

acquiring a cooperative office log of a cooperative office node connected with a target office node to be analyzed and a target office log generated by the target office node in a target service time period, wherein the target office log is an office record generated in the office process of the target office node;

analyzing the collaborative office logs and the target office logs respectively to obtain derived quantitative descriptions corresponding to the collaborative office logs and business quantitative descriptions corresponding to the target office logs, wherein the derived quantitative descriptions are used for representing attention-type contents between the target office nodes and the connected collaborative office nodes, and the business quantitative descriptions are used for representing operation preference descriptions presented by the target office nodes in the office process;

Determining a multi-modal behavior description of the target office node according to the derived quantitative description and the business quantitative description;

on the premise that the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than a set threshold value, determining that the target office node corresponds to a target task progress, wherein the target behavior description is the multi-modal behavior description of the office node corresponding to the target task progress.

2. The method according to claim 1, wherein the determining a multi-modal behavioral description of the collaborative office node to which the target office node is connected based on the derived quantitative description and the traffic quantitative description comprises:

integrating the derived quantitative description and the business quantitative description to obtain a target quantitative description;

determining the target quantization description as a multi-modal behavior description of the target office node;

or, determining the derived quantitative description as a first behavioral description of the target office node; determining the business quantification description as a second behavior description of the target office node, wherein the multi-modal behavior description of the target office node comprises the first behavior description and the second behavior description.

3. The method of claim 2, wherein the determining that the target office node corresponds to a target task process comprises:

on the premise that the target quantization description is determined to be a multi-modal behavior description of the target office node, and the difference degree between the multi-modal behavior description of the target office node and the target behavior description is not larger than the set threshold value, determining that the target office node corresponds to a target task process;

or, on the premise that the derived quantitative description is determined as the first behavior description and the difference between the first behavior description and a first target description is not greater than a first preset difference, determining that the target office node corresponds to a first target task process, wherein the first target description is the attention-type content between the target office node and the connected cooperative office node;

and on the premise that the business quantification description is determined to be the second behavior description, and the difference between the second behavior description and a second target description is not larger than a second preset difference, determining that the target office node corresponds to a second target task process, wherein the second target description is an operation preference description presented by the target office node in the office process.

4. The method according to claim 1, wherein the analyzing the collaborative office log and the target office log respectively to obtain a derived quantitative description corresponding to the collaborative office log and a business quantitative description corresponding to the target office log comprises:

on the premise that the target office log comprises a business interaction record, analyzing the business interaction record to obtain a quantitative description corresponding to the business interaction record, wherein the business interaction record is used for representing the interaction record generated by the target office node in the office process, and the business quantitative description comprises the quantitative description corresponding to the business interaction record;

analyzing the business item information on the premise that the target office log comprises business item information to obtain quantitative description corresponding to the business item information, wherein the business interaction record is used for representing the item information generated by the target office node in the office process, and the business quantitative description comprises the quantitative description corresponding to the business item information;

analyzing the office records of the matched type cooperative office node on the premise that the cooperative office log comprises the office records of the matched type cooperative office node connected with the target office node to obtain quantitative descriptions corresponding to the office records of the matched type cooperative office node, wherein the business quantitative descriptions comprise quantitative descriptions corresponding to the office records of the matched type cooperative office node;

On the premise that the cooperative office log includes the office record of the business cooperative office node connected to the target office node, analyzing the office record of the business cooperative office node to obtain a quantitative description corresponding to the office record of the business cooperative office node, where the business quantitative description includes a quantitative description corresponding to the office record of the business cooperative office node.

5. The method according to claim 4, wherein the parsing the service transaction information to obtain the quantitative description corresponding to the service transaction information comprises:

cleaning noise information in the service event information, wherein the noise information is visual information used for representing that the degree of association with service content does not meet a third preset degree of association in the service event information;

extracting significance information in the cleaned business item information, wherein the significance information is visual information used for representing that the association degree with the business content reaches a fourth preset association degree in the business item information;

and analyzing the significance information to obtain quantitative description corresponding to the business event information.

6. The method according to claim 5, wherein the parsing the significance information to obtain a quantitative description corresponding to the service event information comprises:

analyzing the significance information to obtain a quantitative description corresponding to the significance information;

determining the quantitative description corresponding to the significance information as the quantitative description corresponding to the service event information;

or pairing the significance information and target information to obtain a pairing result of the significance information and the target information, wherein the target information is visual information used for representing that the association degree with abnormal business content reaches a fifth set association degree;

analyzing the pairing result to obtain a quantitative description corresponding to the pairing result;

and determining the quantitative description corresponding to the pairing result as the quantitative description corresponding to the service item information.

7. The method according to claim 1, wherein after the obtaining of the collaborative office log of the collaborative office node to which the target office node to be analyzed is connected and the target office log generated by the target office node in the target service period, the method comprises:

Importing the collaborative office log and the target office log into an AI neural network, wherein the AI neural network is a machine learning model for analyzing data information, which is obtained by performing feature differentiation analysis on a plurality of training samples;

obtaining an analysis result output by the AI neural network, wherein the analysis result is a feature differentiation analysis result of the collaborative office log and the target office log;

and determining the office node task process to which the target office node belongs according to the analysis result.

8. The method of claim 7, prior to said importing the collaborative office log and the target office log into an AI neural network, comprising:

obtaining a plurality of training samples, wherein the training samples comprise sample cooperative office logs and sample target office logs;

circularly implementing the following operations until the AI neural network is obtained:

determining a current training sample from the plurality of training samples, and determining a current AI neural network, wherein the current training sample comprises a current sample collaborative office log and a current sample target office log;

Identifying a current feature differentiation analysis result of the current training sample through the current AI neural network;

on the premise that the current feature differentiation analysis result does not meet an analysis judgment index, acquiring a next training sample as the current training sample;

and determining the current AI neural network as the AI neural network on the premise that the current characteristic differentiation analysis result reaches an analysis judgment index.

9. The method of any of claims 1-8, further comprising, prior to the determining that the target office node corresponds to a target task process:

acquiring configuration data and reference data of the target office node, wherein the configuration data is used for representing office node configuration information of the target office node, and the reference data is used for representing reference keyword information of the target office node;

analyzing the configuration data and the reference data respectively to obtain a configuration quantitative description corresponding to the configuration data and a reference quantitative description corresponding to the reference data, wherein the configuration quantitative description is used for representing key configuration content corresponding to the target office node, and the reference quantitative description is used for representing keyword description content previously distributed by the target office node;

And determining an office node task process to which the target office node belongs according to the configuration quantitative description, the reference quantitative description, the derivation quantitative description and the business quantitative description, wherein the office node task process comprises the target task process.

10. A cloud office information processing server, comprising:

a memory for storing executable instructions;

a processor for operating the cloud office information processing server to perform the method of any one of claims 1-9, under control of the executable instructions.