TWI821666B - Service management system and adaption method of service information process - Google Patents

Abstract

A service management system and an adaption method of service information process are provided. In the method, the key feature of service information is obtained. The service information corresponds to a service on which a provision operation is performed. The rule is established by performing a machine learning algorithm on the key feature. The rule is related to a standard process of processing the service corresponding to the service information. An abnormality of the service corresponding to the service information is detected, and a corresponding rule is provided according to the abnormality. Accordingly, the abnormality could be removed, and the service process could be reduced in time.

Description

Adaptation methods for service management systems and service information processes

The present invention relates to service management, and in particular, to a service management system and an adjustment method of service information flow.

Traditional process management technologies mostly focus on processing standardized service information process integration and information synchronization. However, because telecommunications and value-added services have highly complex business rules and long operational processes for data transfer across many systems, various unexpected and abnormal errors are prone to occur, and even require human intervention to determine the situation and handle it. In addition, rapid changes in the company's business promotion or human operation problems in heterogeneous system processes may produce various process variations, which in turn may lead to the inability of the process to proceed normally.

In view of this, embodiments of the present invention provide a service management system and a service information process adjustment method, introduce machine learning, and develop an intelligent adjustment rule base based on this.

The adjustment method of the service information process in the embodiment of the present invention includes (but is not limited to) In) the following steps: Obtain one or more key characteristics of the service information. The service information corresponds to the service that performs the supply and installation operation. Establish rules for key features through one or more machine learning algorithms. This rule relates to the standardized process for processing service information corresponding to services. Detect and process exceptions in services corresponding to service information, and provide corresponding rules based on the exceptions.

The service management system of the embodiment of the present invention includes (but is not limited to) heterogeneous systems and intelligent adaptation modules. Heterogeneous systems are used for service provisioning operations. The intelligent adaptation module is used to obtain one or more key characteristics of the service information corresponding to the service, establish rules for the key characteristics through one or more machine learning algorithms, and detect and process anomalies in the service corresponding to the service information. , and provide corresponding rules based on exceptions. This rule relates to the standardized process for processing service information corresponding to services.

Based on the above, according to the service management system and the service information process adjustment method according to the embodiment of the present invention, machine learning is used to establish rules for processing services, and corresponding rules are provided in response to exceptions. In this way, exceptions can be effectively resolved and the processing process can be shortened.

In order to make the above-mentioned features and advantages of the present invention more obvious and easy to understand, embodiments are given below and described in detail with reference to the accompanying drawings.

U:User

1: Service management system

101:Service acceptance module

102: Process control module

103: Intelligent adjustment module

104:Service information database

105:Heterogeneous systems

106: Intelligent adjustment rule base

107: Intelligent process control report

SI: Service Information

PI: process information

KF: Key Features

SPI: Service Supply and Installation Information

S210~S290, S251~S253, S310~S350: steps

201:Random Forest Algorithm

202: Reinforcement Learning Method

FIG. 1 is a component block diagram of a service management system according to an embodiment of the present invention.

FIG. 2 is a flow chart of feature pre-processing according to an embodiment of the present invention.

Figure 3 is a flow chart of a service information flow adjustment method according to an embodiment of the present invention. Process map.

Figure 4 is a schematic diagram of an adjustment operation according to an embodiment of the present invention.

FIG. 1 is a component block diagram of a service management system 1 according to an embodiment of the present invention. Please refer to Figure 1. The service management system 1 includes (but is not limited to) a service acceptance module 101, a process control module 102, an intelligent adaptation module 103, a service information database 104, a heterogeneous system 105 and an intelligent adaptation rule database 106. .

In one embodiment, the service acceptance module 101, the process control module 102, the intelligent adaptation module 103, and the heterogeneous system 105 may be respectively composed of a computer, a smart phone, a server, a cloud system, a wearable device, and an intelligent assistant. Devices, chips, processors and other hardware components or device implementations. In one embodiment, some or all of the service acceptance module 101, the process control module 102, the intelligent adaptation module 103, and the heterogeneous system 105 may be integrated into a single device. In some embodiments, the functions of the service acceptance module 101, the process control module 102, the intelligent adaptation module 103, and the heterogeneous system 105 may be implemented by software.

In one embodiment, the service information database 104 and the intelligent adaptation rule database 106 can be implemented in a static or dynamic random access memory (Random Access Memory, RAM), read-only memory (Read-Only Memory, ROM), Flash Memory, various types of hard drives, cloud space or other storage media.

One or more purposes of one or more embodiments of the present invention is to provide a design method and system for intelligently adapting service information processes, introducing machine learning technology, And based on this, an intelligent adjustment rule base is automatically developed. The embodiment of the present invention continuously and automatically detects the service information process status. Once an unexpected abnormality or error occurs, intelligent adjustment operations are immediately carried out to automatically correct the abnormality. Embodiments of the present invention have the advantage of reinforcement learning, and find the best solution through long-term feedback automatic adjustment. In addition, embodiments of the present invention provide control reports to management and demand units, which can monitor service information process cycles, improve existing standard processes, and even perform process reengineering to achieve the best or better handling methods for establishing abnormal service information processes.

In the following, various components or modules in the service management system 1 will be used to describe the method described in the embodiment of the present invention. Each process of this method can be adjusted according to the implementation situation, and is not limited to this.

FIG. 2 is a flow chart of feature pre-processing according to an embodiment of the present invention. Referring to FIG. 2 , the user U can accept the service through the service acceptance module 101 (step S210 ) to generate service information SI. In one embodiment, the accepted services relate to telecommunications related services. For example, fiber optic network installation, mobile network bandwidth upgrade, or channel selection for multimedia content. In other embodiments, the accepted services may also be related to various types of resources (for example, water, electricity, or gas), engineering (for example, decoration, or construction), or other software and hardware services, and are not limited thereto. In addition, user U may request services through the Internet, phone, email, counter, etc. The service information SI is the content of the service to be applied for, and is one of the data sources for the subsequent construction of the key features KF of the intelligent adaptation rule base (to be described in detail later). Service information SI may include service items, service types (for example, add/adjust/delete) and/or service details (for example, fiber optic network installation location, network bandwidth, customer requirements, etc.), and may be applied in different fields And tune all.

The service acceptance module 101 may store the service information SI generated by the service acceptance module 101 into the service information database 104 (step S230).

The process control module 102 can receive the service information SI generated by the service acceptance module 110, and transmit the service information SI to the corresponding heterogeneous system 105 of the corresponding service station according to the rules of the standardized service process to perform the provisioning operation of this service ( Step S250). Specifically, the process control module 102 is responsible for process control of all service information SI. For example, process sequence, parameter transfer, job switching, etc. Among them, the process control module 102 sends the service information SI to the relevant interfaced heterogeneous systems 105 according to standardized rules according to the defined process. The values/parameters generated during the operation of the process control module 102 can be used as process information PI. For example, interface site, process result, process status, etc. Process information PI can also be used as one of the data sources for the subsequent construction of key features KF of the intelligent adjustment rule base (to be detailed later). It is worth noting that in some application scenarios, unexpected factors such as network instability may cause the service information SI to be unable to be sent to the heterogeneous system 105 normally, which may lead to abnormal process processing.

On the other hand, the heterogeneous system 105 is responsible for service provisioning operations, accounting operations, and/or providing maintenance and warranty services. The heterogeneous system 105 can return the processed service installation information SPI, and update the relevant content of the service accepted by the service acceptance module 101 accordingly, so that the user U can query to know the service installation progress. The service supply and installation information SPI includes, for example, supply and installation time, ordering time, supply and installation exchange office, etc. The service provision information SPI can also be used as one of the data sources for constructing the key features KF of the intelligent adaptation rule base (to be discussed in detail later). It is worth noting that in some application scenarios, the heterogeneous system 105 The operation process may not be able to return processing results normally due to unexpected factors, or abnormal and non-standardized data content may be returned, causing abnormal process processing.

In addition, the process information PI and/or the service installation information SPI can also be stored in the service information database 104 (steps S270 and S290) for access by the intelligent adaptation module 103. In some embodiments, the service information database 104 further stores historical data. That is, past service information SI, process information PI and/or service installation information SPI.

FIG. 3 is a flow chart of a method for adjusting a service information process according to an embodiment of the present invention. Referring to FIG. 3 , the intelligent adaptation module 103 can obtain one or more key features KF of the service information corresponding to the service (step S310 ). Specifically, the key feature KF is obtained through Feature Extraction. Feature extraction can be applied in fields such as machine learning, pattern recognition, image processing or other technical fields. Feature extraction is to construct informative and non-redundant derived values (or called feature values) from the data samples initially measured/collected/extracted. Feature extraction can assist the subsequent learning and rule induction process, and can provide a better interpretation of the initial data sample. In other words, feature extraction can simplify the input data into a feature set (which can be regarded as important or useful information), and directly use the feature set to perform subsequent tasks (for example, model training, component analysis, object detection, etc.).

The intelligent adaptation module 103 may establish rules for the key features KF through one or more machine learning algorithms (step S330). Specifically, the intelligent adaptation module 103 can obtain the key features KF of the service information SI, and use autonomous training and learning of machine learning technology to develop new rules (ie, create new rules), and add rules to the intelligent system accordingly. Adaptation rule base 106. This rule is related to processing service information SI correspondence Standardized processes for services. For example, the correct order, parameters, or objects. At the same time, the intelligent adaptation module 103 can continuously or periodically detect the status of the service process (for example, service provision, accounting operations and/or maintenance and warranty) by collecting the key features KF of the service information SI. , normal, abnormal, error, stop, etc.), and obtain key features based on the status of the service process. Once non-standardized anomalies (states different from standard processes) are found from key features, corresponding rules can be provided to perform intelligent adaptation operations (for example, repair or correct services and/or their service information based on rules) .

Figure 4 is a schematic diagram of an adjustment operation according to an embodiment of the present invention. Please refer to Figure 4. In one embodiment, the machine learning algorithm includes a random forest (Random Forest) algorithm 201 and a reinforcement learning (Reinforcement learning, RL) method 203.

The intelligent adaptation module 103 can determine the classification result of one or more key features KF through the random forest algorithm 201. Specifically, the random forest algorithm 201 can bring in the historical data corresponding to the key features KF of the service information SI (which may be accumulated for a specific period (for example, one year, three months or half a year)), and does not require human intervention to define features, Adopt autonomous sampling method, random classification, and through continuous training, intelligent adjustment rules (corresponding to a certain classification result) are automatically generated. This rule can be stored in the intelligent adaptation rule base 106.

In some application scenarios, if the number of key features KF of service information SI is less than the corresponding threshold, the number of intelligent adaptation rules that can be generated is limited, and it may take longer to accumulate intelligent adaptation rules.

The intelligent adaptation module 103 can assign rewards or penalties to new types of information through the reinforcement learning method 202, thereby generating rules for new types of information. Specifically, The reinforcement learning method 202 can be applied to new environments. The concept is that the robot learns to act based on the environment. Through the stimulation of rewards or punishments given by the environment after the action, the expectation of feedback is gradually formed, and then the rules of intelligent adjustment are formed. This new type of information is different from the service information that has been trained in the past. For example, different service items, types and/or details may be called new service information. Therefore, even for new value-added products, site expansion of heterogeneous systems 105, process changes, or other new types of information, its automated process learning robots can still develop new rules through continuous free training and learning and store them in Intelligently adjust the rule base 106 to improve the flexibility and scalability of learning.

What is more noteworthy is that in one embodiment, the intelligent adaptation module 103 can establish rules based on the classification results and the assignment results. Specifically, the intelligent adaptation module 103 can integrate the random forest algorithm 201 and the reinforcement learning method 202. For example, the random forest algorithm 201 is used for historical data, and the reinforcement learning method 202 is used for new types of information. Therefore, the embodiment of the present invention only needs to produce an intelligent adjustment rule base through automatic classification and continuous training, without the need for time-consuming manual construction, and the best solution of the process can be found through a long-term feedback adjustment process.

In step S350, the intelligent adaptation module 103 can detect anomalies in the service corresponding to the processing service information (step S251), and provide corresponding rules according to the anomaly (step S253). Specifically, exceptions include missing, missing or irregular service information, disordered processes, missing or missing return results, inconsistent content, or failed identity verification, and may vary depending on the application scenario. The intelligent adaptation module 103 continuously or periodically detects whether known anomalies or unknown anomalies occur. If an abnormality is detected, the intelligent adaptation module 103 can analyze the type of abnormality, provide corresponding rules, and make adjustments accordingly. Automatically correct incorrect data. During the aforementioned debugging and correction process, no time-consuming human intervention is required to solve the problem of lengthy and time-consuming processes, thereby reducing labor consumption and improving customer satisfaction, thereby increasing the speed and revenue of the supply and installation of telecommunications services and value-added products.

In addition, the intelligent adaptation module 103 can regularly export intelligent process control reports 107 for management and demand units to monitor abnormal service processes. Specifically, in the process of service information acceptance and installation across multiple systems, it is difficult to synchronize data to management and demand units in real time with existing technology. The embodiment of the present invention can export intelligent process control reports 107 regularly, in response to abnormal occurrences, or according to user U requirements, and provide it to the design unit of the service information process as a reference for improving the existing process, thereby effectively Realize the review of service information process design from demand sources, and seek to improve cross-system abnormal problem handling solutions, which will help enhance the sustainable development of the overall system. The intelligent process control report 107 may, for example, record the exception content, exception occurrence time, and/or the completed process.

To sum up, the service management system and the service information process adjustment method according to the embodiment of the present invention include the following features: Embodiments of the present invention provide intelligent anomaly correction and adjustment processes for service operations that have cross-system, cross-organization, and long operation process characteristics that rely on manpower to be executed at different operation checkpoints, and optimize the process accordingly.

Embodiments of the present invention do not require human intervention, and only need to develop a process intelligent adjustment rule base to intelligently correct process information. In addition, embodiments of the present invention can generate intelligent process control reports to achieve the purpose of handling abnormal processes.

The embodiment of the present invention uses the random forest algorithm to bring into the system historical data (which Including process station details as "key features of service information"), there is no need for human intervention to define features, adopt independent sampling methods, and randomly classify them. Through continuous training, an intelligent adjustment rule library is automatically generated.

Embodiments of the present invention improve the flexibility and scalability of learning through reinforcement learning methods. Even if there are new value-added products, heterogeneous system site expansion, process changes, or other new types of information, its automated process learning robot will undergo independent training and learning, and then develop new rules and add them to the intelligent scheduling rule library. In addition, taking advantage of the reinforcement learning method, the best solution to the rules can be found through a long-term feedback adjustment process.

Embodiments of the present invention will integrate the two methods of random forest and reinforcement learning, and only need to produce an intelligent adjustment rule base through automatic classification and continuous training, without the need for manual construction, thereby shortening the process.

Although the present invention has been disclosed above through embodiments, they are not intended to limit the present invention. Anyone with ordinary knowledge in the technical field may make some modifications and modifications without departing from the spirit and scope of the present invention. Therefore, The protection scope of the present invention shall be determined by the appended patent application scope.

S310~S350: steps

Claims (8)

A method for adapting a service information process, including: obtaining at least one key feature of a service information, wherein the service information corresponds to a service for performing an installation operation; and determining a classification of the at least one key feature through at least one machine learning algorithm As a result, rewards or penalties are assigned to a new type of information, where the new type of information is different from the service information; a rule is established based on the classification result and the assignment result, where the rule is related to the standardized process for processing the service information corresponding to the service ; and detect and process an exception in the service corresponding to the service information, provide the corresponding rule according to the exception, and automatically correct the error data corresponding to the exception according to the rule. The service information process adjustment method as described in claim 1, wherein the at least one machine learning algorithm includes a random forest (Random Forest) algorithm and a reinforcement learning (Reinforcement learning, RL) method, and through the at least one machine learning algorithm The steps of the learning algorithm to establish the rule include: determining the classification result of the at least one key feature through the random forest algorithm; and assigning rewards or penalties to the new type of information through the reinforcement learning method. The method for adapting a service information process as described in claim 1, wherein the step of obtaining at least one key characteristic of the service information includes: detecting the status of a service process, wherein the service process includes service provisioning, accounting service operation and maintenance warranty; and obtain the at least one key feature based on the status of the service process. The method for adapting a service information process as described in claim 3, wherein the status of the service process includes a service provisioning information, and the service provisioning information includes provisioning time, scheduling time, and at least one of the provisioning and installation exchange offices. One. A service management system, including: a heterogeneous system for performing installation operations of a service; and an intelligent adaptation module for: obtaining at least one key feature of a service information corresponding to the service; through at least one machine The learning algorithm determines a classification result of the at least one key feature, and assigns rewards or penalties to a new type of information, wherein the new type of information is different from the service information; establishes a rule based on the classification result and the assignment result, wherein the rule Relevant to the standardized process of processing the service information corresponding to the service; and detecting an exception in the service corresponding to the processing of the service information, providing the corresponding rule according to the exception, and automatically correcting the error corresponding to the exception according to the rule material. The service management system of claim 5, wherein the at least one machine learning algorithm includes a random forest algorithm and a reinforcement learning method, and the intelligent adaptation module is further used to: determine the random forest algorithm through the random forest algorithm. The classification result of at least one key feature; The new type of information is rewarded or punished through the reinforcement learning method. The service management system as described in claim 5, wherein the heterogeneous system is further used to: detect the status of a service process, wherein the service process includes service provision, accounting operations and maintenance and operation and maintenance; and based on the service process The state obtains the at least one key characteristic. The service management system of claim 7, wherein the status of the service process includes a service provisioning information, and the service provisioning information includes at least one of provisioning time, scheduling time, and provisioning exchange.

Images