Cloud dedicated line customer service application scenario identification method, device and computer readable storage medium

Abstract

The present disclosure relates to a method, device and computer-readable storage medium for identifying business application scenarios of cloud dedicated lines. The method comprises: a data acquisition step for acquiring object data for business application scenario identification; a feature appending step for appending flow-related features and time-related features to the object data acquired in the data acquisition step; a feature engineering step for selecting a feature group for the object data obtained by appending flow-related features and time-related features; and a strategy selection step for selecting a model for business application scenario identification based on the feature group selected in the feature engineering step.

Description

Technical Field

[0001] The present disclosure generally relates to a method, device and computer-readable storage medium for identifying application scenarios of cloud dedicated line customer services, and more specifically to a method, device and computer-readable storage medium for identifying application scenarios of cloud dedicated line customer services in the fields of network operations and big data technology. Background Art

[0002] With the development of social science and technology, people's business application scenarios have become rich and colorful, especially the advent of the 5G era has accelerated the pace of diversification of business application scenarios. When customers rent circuits, operators will provide various types of circuits. For example, customers will rent VPN (Virtual Private Network), IPRAN (IP Radio Access Network) and other circuits according to their needs. The business application scenarios of different types of circuits are different, such as video conferencing, data transactions, etc. However, customers will also modify the business application scenarios of the leased circuits according to their needs. When the operator is unaware of the customer's actual business, the bandwidth allocation will not meet the customer's needs, resulting in poor customer perception. The operator's customer engineers need to predict the customer's business application scenarios in advance, analyze the real needs of different customers, solve problems in a timely manner, and provide corresponding optimization suggestions.

[0003] There are many technologies for business application scenario identification in the prior art. For example, in patent document 1, a method and device for identifying instant messaging business operation behavior are recorded to solve the problems of low recognition accuracy and low efficiency of existing instant messaging business operation behavior identification methods. The method includes: collecting a binary code stream generated by the instant messaging business operation of the user terminal; when it is determined that the binary code stream contains a specific feature field, extracting the specific feature field content from the binary code stream; determining the instant messaging business operation behavior according to the correspondence between the extracted specific feature field content and the preset feature field content and the operation behavior. In patent document 2, a business scenario identification method, device and electronic device are disclosed, which can perform data anomaly detection on the transaction feature data of the extracted transaction data to obtain the data anomaly detection result, and then the correspondence between the business application scenario and the application system involved in the business application scenario, determine the business application scenario corresponding to the data anomaly detection result, and realize the identification of the business application scenario. In patent document 3, a network business scenario identification method and device are proposed, including: inputting feature data of a business data flow into a trained long short-term memory network model, and outputting a business type label corresponding to the business data flow; and obtaining the business type of the business data flow according to the business type label output by the long short-term memory network model.

[0004] Patent document 1: CN111314104A

[0005] Patent document 2: CN111191720A

[0006] Patent document 3: CN111245667A Summary of the invention

[0007] Then, the existing output results related to business application scenario identification have the following problems.

[0008] 1) The combination of multiple input features and the addition of new features are not considered.

[0009] 2) Failure to consider the time characteristics and the differences between different businesses in different time periods.

[0010] 3) Currently, only one model is used to identify business application scenarios for a single type of data, without considering the situation where multiple types of data are used to jointly select business application scenario identification strategies.

[0011] 4) The different characteristics of different types of data are not taken into account during the business application scenario identification process, and it is impossible to solve the optimal selection problem for business application scenario identification of multiple types of data (including the fusion of models in the business application scenario identification strategy).

[0012] 5) The business application scenario identification process and results are relatively independent and are not applied to the system for further analysis.

[0013] The present invention provides a method, device and computer-readable storage medium for identifying application scenarios of cloud dedicated line customer services.

[0014] According to one aspect of the present disclosure, a method for identifying a cloud dedicated line customer service application scenario is provided, comprising:

[0015] A data acquisition step, acquiring object data for business application scenario identification;

[0016] a feature adding step of adding flow-related features and time-related features to the object data acquired in the data acquiring step;

[0017] Feature engineering processing step: selecting a feature group from the object data obtained by adding features related to flow and features related to time; and

[0018] The strategy selection step selects a model for business application scenario identification based on the feature group selected in the feature engineering processing step.

[0019] According to another aspect of the present disclosure, a cloud dedicated line customer service application scenario identification device is provided, comprising:

[0020] A data acquirer, which acquires object data for business application scenario identification;

[0021] a feature appender, which appends flow-related features and time-related features to the object data acquired by the feature appender;

[0022] A feature engineering processor selects a feature group from the object data obtained by adding features related to flow and features related to time; and

[0023] A strategy selector selects a model for business application scenario identification based on the feature group selected by the feature engineering processor.

[0024] According to another aspect of the present disclosure, a computer-readable storage medium is provided, which includes computer-executable instructions. When the computer-executable instructions are executed by one or more processors, the one or more processors execute the method according to the above aspects of the present disclosure.

[0025] A brief overview of the disclosure is given below in order to provide a basic understanding of some aspects of the disclosure. However, it should be understood that this overview is not an exhaustive overview of the disclosure. It is not intended to identify the key or important parts of the disclosure, nor is it intended to limit the scope of the disclosure. Its purpose is simply to give some concepts of the disclosure in a simplified form as a prelude to a more detailed description given later. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings, which constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0027] The present disclosure may be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0028] Figure 1 The overall technical architecture diagram of the cloud dedicated line customer business application scenario identification disclosed in this disclosure is shown.

[0029] Figure 2 A functional block diagram of a cloud dedicated line customer service application scenario identification device is shown.

[0030] Figure 3 A flow chart of a method for identifying application scenarios of cloud dedicated line customer services is shown.

[0031] Figure 4 An internal flow chart of the strategy selector is shown.

[0032] Figure 5 The result graph of the Pearson correlation of each feature in the embodiment is shown.

[0033] Figure 6 An exemplary configuration of a computing device that can implement an embodiment according to the present disclosure is shown. DETAILED DESCRIPTION

[0034] The following detailed description is made with reference to the accompanying drawings, and the following detailed description is provided to assist in a comprehensive understanding of the various exemplary embodiments of the present disclosure. The following description includes various details to assist in understanding, but these details are considered to be examples only and not to limit the present disclosure, which is defined by the appended claims and their equivalents. The words and phrases used in the following description are only used to enable a clear and consistent understanding of the present disclosure. In addition, for the sake of clarity and brevity, descriptions of well-known structures, functions, and configurations may be omitted. Those of ordinary skill in the art will recognize that various changes and modifications may be made to the examples described herein without departing from the spirit and scope of the present disclosure.

[0035] Figure 1The overall technical architecture diagram of the cloud dedicated line customer service application scenario identification disclosed in the present invention is shown.

[0036] exist Figure 1 In the present disclosure, the overall technical architecture for identifying customer business application scenarios for cloud dedicated lines is roughly divided into four layers: application display layer, model design layer, data processing layer, and data layer. Of course, these four layers are just an example and are proposed for ease of understanding. Other hierarchical divisions can also be used. In the overall technical architecture for identifying customer business application scenarios for cloud dedicated lines disclosed in the present disclosure, data acquisition, data cleaning, feature data analysis, and feature engineering processing are performed, and models such as semi-supervised models and supervised models are trained and selected, and then the obtained models are applied to identify and analyze customer business application scenarios for cloud dedicated lines, and such identification results are applied to the CEN (Customer Experience Management: Customer Experience Management System) system.

[0037] The structure and characteristics of each layer are described in detail below.

[0038] In the data layer, for example, a customer information database, a circuit information database, a flow information database, etc. are stored. Regarding the data layer, the acquired customer information data, circuit information data, and flow information data are respectively stored in the above-mentioned customer information database, circuit information database, and flow information database. For example, the original data is collected by the network management of major customers and stored in the data warehouse Hive, and the object data for research and analysis is selected to form a CSV file. Among them, the flow data can collect the data with the largest daily flow. Of course, this is not limited to daily flow, and it can also be the data with the largest flow in a specific time period. The so-called Hive is a data warehouse infrastructure built on Hadoop (a distributed system infrastructure). Of course, the present invention is not limited to storage in the data warehouse Hive, but can also be other data warehouses.

[0039] In the data processing layer, the data cleaning, feature data analysis, addition of new features (sometimes referred to as new features for the sake of explanation below), and feature engineering processing are included. Data cleaning refers to data validity verification and invalid data processing, outliers, missing values, etc. After data cleaning here, unnecessary data that may cause interference is eliminated, which will make subsequent identification more accurate. Of course, if the acquired data itself is clean and meets the requirements, this data cleaning is not essential. Feature data analysis refers to analyzing customer business-related features, such as customer features, circuit features, and traffic features, and adding new features based on existing features to facilitate clustering and classification analysis. Feature engineering processing refers to nominal data processing and data standardization processing. The model design layer includes model library, model training, and model application. Among them, the so-called model library is a cloud dedicated line customer business application scenario recognition model, such as the Tri-training model for semi-supervised classification and the GMM model (Gaussian mixture model) for unsupervised clustering, etc., and the model needs to be selected according to the circuit type and label situation. The so-called model training is to train the model based on historical data and save it. The so-called model application refers to the application of the saved model to identify the data business application scenarios of the existing network.

[0040] In the application layer, for example, the output of the model is applied in the CEM system to display relevant information on multiple pages and conduct experience evaluation, etc. Multiple pages include, for example, a customer overview page, a single customer page, a single circuit page, etc.

[0041] The following uses Figure 2-4 The present invention describes the cloud dedicated line customer service application scenario identification method and device, and particularly describes the strategy formulation and process in detail.

[0042] Figure 2 It is a functional block diagram of a cloud dedicated line customer service application scenario identification device. Figure 3 The figure is a flow chart of a method for identifying business application scenarios of cloud dedicated line customers. Figure 4 is the internal flow chart of the strategy selector.

[0043] Figure 2The cloud dedicated line customer service application scenario identification device 1 in the embodiment includes a data acquirer 101, a feature appender 102, a feature engineering processor 103, a strategy selector 104 and a model applicator 105. The data acquirer 101 acquires object data. The object data is used for subsequent cloud dedicated line customer service application scenario identification. The object data can be acquired by performing feature data selection, data cleaning, association and other processing on the acquired data. Specifically, such as selecting and analyzing customer features, circuit features and traffic features. Associate with existing tag data according to circuit ID. The feature appender 102 appends new features to the object data acquired by the data acquirer 101. The newly added features here include the addition of features related to traffic and features related to time. The features related to traffic are, for example, traffic difference features, and the features related to time are, for example, dividing the data into 4 time periods on a daily basis, but are not limited to this, and can also be specifically divided as needed. The feature engineering processor 103 performs feature engineering processing on the data from the feature appender 102. The feature engineering process mainly includes feature normalization and feature group selection. For example, the normalized features may include inflow flow, outflow flow, flow difference, circuit rate, industry, inflow flow rate by time period, inflow flow rate by time period, non-zero number of inflow flow rate, outflow flow rate by time period, non-zero number of outflow flow rate by time period, etc. The feature group selection can be based on the Pearson correlation coefficient method to select the final feature group. The specific selection content will be described later. The present disclosure is not limited to the selection of the final feature group based on the Pearson correlation coefficient method, and other selection methods and bases can be selected according to needs. The strategy selector 104 selects a model for business application scenario identification, and then selects the corresponding business application scenario identification strategy for business application scenario identification. The model applicator 105 trains the model selected by the strategy selector 104 and applies it to business application scenario identification. Specifically, the strategy selector 104 determines which type of circuit the circuit belongs to based on the value of the circuit type attribute, selects a pre-established algorithm model based on the business label quantity, and selects the corresponding business application scenario identification strategy. The model applicator 105 applies the selected model to the business application scenario identification. In addition, the information of business application scenario identification is integrated into the CEM system, and the relevant feature information is displayed for each type of business application scenario, which is convenient for the customer to analyze customer needs. In addition, after application, the output label can be fed back to the strategy selector to increase the number of training sets output by the strategy selector and improve the accuracy of model training.

[0044] Use below Figure 3The process of the cloud dedicated line customer service application scenario identification method is described below. First, in step S01, the data acquirer 101 acquires data. Next, in step S02, the data acquirer 101 performs data preprocessing, including feature data selection and data cleaning, such as the selection and analysis of customer features, circuit features, and traffic features. And the circuit ID is associated with the existing tag data.

[0045] Next, in step S04, the feature appender 102 adds new features, including features related to flow and features related to time. Here, the flow difference feature is given as an example. In addition, the time-related feature is, for example, to divide the data into several time periods on a daily basis. The flow difference can be obtained by the following formula (1).

[0046] Flow difference = inflow flow - outflow flow (1)

[0047] In addition, other calculations may be performed on the flow-related features and the time-related features, for example, by referring to equations (2) and (3) to calculate parameters that relate the two features.

[0048]

[0049]

[0050] Next, in step S05, the feature engineering processor 103 performs feature engineering processing, which mainly includes feature normalization processing and feature group selection. Regarding the normalization processing, for example, the following formula (4) can be used to perform feature normalization.

[0051]

[0052] where x * is the normalized eigenvalue, x is the eigenvalue to be normalized, min is the minimum value of the historical data of the feature, and max is the maximum value of the historical data of the feature.

[0053] Features that need to be normalized include: inflow flow, outflow flow, flow difference, circuit rate, industry, inflow flow ratio by time period, inflow flow ratio by time period, outflow flow ratio by time period, outflow flow ratio by time period, outflow flow ratio by time period, but not limited to this, various normalizations can be performed as needed. This normalization process is not necessary, and can be performed as appropriate based on the situation when the above features are acquired. In addition, the final feature group needs to be selected based on the Pearson correlation coefficient method. For example, the formula used is the following formula (5).

[0054]

[0055] Where Cov(X,Y) is the covariance of X and Y, and σXσY is the product of their respective standard deviations.

[0056] Next, in step S06, the strategy selector 104 performs strategy selection. The strategy selector 104 needs to determine which type of circuit the circuit belongs to according to the value of the circuit type attribute, and select the corresponding business application scenario identification strategy based on the business label quantity and the pre-established algorithm model to perform business application scenario identification.

[0057] Then, in step S07, the information involved in the model selected by the strategy selector 104 can be input to the model applicator 105, and the model applicator 105 uses the information to train the selected model, so that the subsequent scene recognition is more accurate. In step S08, the model applicator 105 applies the trained model, and uses the business application scenario recognition strategy selected in step S06 to integrate the information of business application scenario recognition into the CEM system in step S09, and displays relevant feature information for each type of business application scenario, so as to facilitate customer workers to conduct customer demand analysis. In addition, the output label can also be fed back to the feature group data set of the cloud dedicated line customer business application scenario recognition device 1 to increase the number of training sets and improve the accuracy of model training. Of course, step S07 as an alternative step will make the cloud dedicated line customer business application scenario recognition strategy more perfect.

[0058] Regarding the strategy selection in step S06, refer to Figure 4To illustrate its specific steps. The strategy selection in step S06 is mainly implemented according to the following process. First, the strategy selector receives the input of the feature group. Then enter the circuit type identification step S301. In the circuit type identification step S301, the circuit type is determined according to the parameters related to the circuit type contained in the feature group. For example, it is determined whether the circuit type is VPN or IPRAN. Next, according to the judgment result of the circuit type, the step S302 of associating the label library is implemented. For example, when the circuit type is VPN, it corresponds to no corresponding service label, so the label is 0. When the circuit type is IPRAN, it corresponds to some service application scenario labels, so the label is part. Next, the label duty cycle is calculated in step S303. When the circuit type is VPN, there is no corresponding service label, so the label ratio is 0%, so the unsupervised clustering model can be selected. For example, when the circuit type is IPRAN, there are some service application scenario labels, so the label ratio is between 0%-100%, so the semi-supervised classification model can be selected. If the ratio of the scenario label corresponding to the judged circuit type is 100%, the supervised clustering model can be selected. Based on the selected model, the corresponding business application scenario identification strategy is selected for business application scenario identification. That is, the circuit type and label ratio in the feature group are used to select the model, and then the strategy corresponding to the selected model is selected to identify the business application scenario of the circuit corresponding to the type. Of course, this is just an example of a strategy selection, and different parameters can be used to make judgments in different ways according to specific needs.

[0059] Based on the above description, the present disclosure relates to a method and device for identifying business application scenarios for cloud dedicated line customers, which can formulate different business application scenario identification strategies based on the cloud dedicated line data characteristics of different types of circuits, newly added traffic and time characteristics, and the number of tags of business application scenarios. By judging the business application scenario tags of historical data, when new data is input, the business application scenario identification strategy can be automatically selected according to the number of tags and circuit types, so as to realize automatic identification of business application scenarios for cloud dedicated line customers. For example, for IPRAN cloud dedicated line data, which currently contains some business application scenario tags, a semi-supervised classification method can be used for business tag identification. When there is new input data, it is necessary to first determine the circuit type of the data. If it is IPRAN cloud dedicated line data, the semi-supervised classification method can be automatically selected, and preparations for adding new features can be made before training the model.

[0060] In addition, the method and device disclosed in the present invention can be integrated into the existing CEM system, and different plans can be formulated for all customers, single customers and single circuits respectively, and the identification results of the final business application scenarios can be applied to the above three pages to realize the information display of the corresponding business application scenarios.

[0061] An embodiment is briefly described below.

[0062] Here, the technical solution of the present disclosure is implemented for the circuit type of IPRAN cloud dedicated line data.

[0063] First, cloud dedicated line data is obtained from the VIP customer network management platform, including customer name, customer industry, circuit name, circuit rate, inbound flow, outbound flow, duplicate and null values ​​are deleted and the units are unified, customer, circuit and flow data tables are associated according to the circuit name, and data with both inbound flow and outbound flow of 0 are removed. The circuit ID is associated with the existing tag data. The steps so far can be implemented by the data acquirer 101.

[0064] Next, according to formulas (1), (2) and (3), additional features such as flow difference and time are added. This step can be implemented by the feature appender 102.

[0065] Then, use formula (4) to normalize the features. The time feature value is already between 0 and 1, so no normalization is required. The Pearson correlation between the features is calculated using formula (5). The result is: Figure 5 shown.

[0066] Among them, influx to out_zb represent, in order: inflow traffic, outflow traffic, circuit rate, customer industry, traffic difference, hour, inflow traffic ratio by time period, outflow traffic ratio by time period, inflow traffic number non-zero traffic ratio by time period, outflow traffic number non-zero traffic ratio by time period.

[0067] From the correlation results, it can be seen that outflux is highly correlated with d_value, in_value_per is highly correlated with out_value_per, and in_zb is highly correlated with out_zb. When identifying and analyzing business application scenarios, one feature is selected from each group.

[0068] Select the feature group: corr = ['influx', 'circrate', 'industry', 'd_value', 'hour', 'out_value_per', 'out_zb']

[0069] The steps up to this point can be implemented by the feature engineer 103 .

[0070] Next, the policy selector 104 determines the circuit type. Assuming that the present embodiment selects IPRAN cloud dedicated line data, the circuit type identified by the policy selector is IPRAN. Then, through the label library association calculation, it is obtained that the IPRAN label accounts for 50%, so the present embodiment applies the semi-supervised classification algorithm - Tri-training model.

[0071] Finally, the model applicator 105 applies the semi-supervised classification algorithm - Tri-training model in the CEN system to identify customer service application scenarios of IPRAN cloud dedicated line data.

[0072] The following is an evaluation of various results of applying the methods disclosed herein and methods not disclosed herein.

[0073] Solution evaluation 1 (evaluating the impact of selected features and new features on the identification of cloud dedicated line customer business application scenarios)

[0074] The following Table 1 compares the accuracy of different feature combinations.

[0075] Table 1

[0076] Algorithm / Accuracy corr corr1 corr2 Tri-training 99.65 98.7 96.8

[0077] Among them, corr is the feature group formulated in the present invention, corr1 is the feature group considering the combination of flow and circuit data: corr = ['influx', 'outflux', 'circrate']; corr2 is the feature group considering the combination of flow, circuit and customer data: corr = ['influx', 'outflux', 'circrate', 'industry']. It can be seen from the accuracy that feature selection and new features in the present invention are more effective.

[0078] Solution Evaluation 2 (Evaluating the Effectiveness of the Strategy Selector)

[0079] The present disclosure finally selected the Tri-training semi-supervised classification algorithm according to the label and circuit type. The accuracy rate in the above evaluation reached 99.65%, which also illustrates the effectiveness of the strategy selector device.

[0080] Based on the above description and embodiments, the present disclosure has the following advantages.

[0081] The present invention relies on the technical advantages of operators, and obtains customer work order data information from the key customer network management platform, including the customer's industry, circuit rate, circuit type, inflow and outflow traffic data, etc., and supports real-time updating of data.

[0082] In addition, the present disclosure can automatically select corresponding business application scenario identification strategies based on different types of circuits and business tags, supports automatic adjustment of parameters, and reduces manual participation.

[0083] In addition, the present disclosure directly applies the results of business application scenario identification to the existing CEM system without rebuilding the system, making the solution easy to promote.

[0084] Finally, the output results of the present disclosure are applied in the existing system, which helps the guest workers to further analyze customer needs and feedback business tags, improve the accuracy of the model and enhance customer perception, so that the operator can achieve greater economic benefits.

[0085] Figure 6 An exemplary configuration of a computing device 1200 capable of implementing embodiments according to the present disclosure is shown.

[0086] The computing device 1200 is an example of a hardware device to which the above aspects of the present disclosure can be applied. The computing device 1200 can be any machine configured to perform processing and / or computing. The computing device 1200 can be, but is not limited to, a workstation, a server, a desktop computer, a laptop computer, a tablet computer, a personal data assistant (PDA), a smart phone, a vehicle-mounted computer, or a combination thereof.

[0087] like Figure 6 As shown, the computing device 1200 may include one or more components that may be connected or communicated with the bus 1202 via one or more interfaces. The bus 2102 may include, but is not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, an Enhanced ISA (EISA) bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus. The computing device 1200 may include, for example, one or more processors 1204, one or more input devices 1206, and one or more output devices 1208. The one or more processors 1204 may be any type of processor and may include, but is not limited to, one or more general-purpose processors or special-purpose processors (such as special-purpose processing chips). The processor 1202 may correspond to, for example, Figure 2The cloud dedicated line customer service application scenario identification device 1 in the embodiment is configured to implement the cloud dedicated line customer service application scenario identification method. The input device 1206 can be any type of input device capable of inputting information to the computing device, and can include but is not limited to a mouse, a keyboard, a touch screen, a microphone and / or a remote controller. The output device 1208 can be any type of device capable of presenting information, and can include but is not limited to a display, a speaker, a video / audio output terminal, a vibrator and / or a printer.

[0088] The computing device 1200 may also include or be connected to a non-transitory storage device 1214, which may be any non-transitory storage device that can implement data storage, and may include, but is not limited to, a disk drive, an optical storage device, a solid-state memory, a floppy disk, a flexible disk, a hard disk, a tape, or any other magnetic medium, a compressed disk or any other optical medium, a cache memory, and / or any other storage chip or module, and / or any other medium from which a computer can read data, instructions, and / or code. The computing device 1200 may also include a random access memory (RAM) 1210 and a read-only memory (ROM) 1212. The ROM 1212 may store programs, utilities, or processes to be executed in a non-volatile manner. The RAM 1210 may provide volatile data storage and store instructions related to the operation of the computing device 1200. The computing device 1200 may also include a network / bus interface 1216 coupled to a data link 1218. The network / bus interface 1216 may be any type of device or system capable of enabling communication with an external device and / or a network, and may include but is not limited to a modem, a network card, an infrared communication device, a wireless communication device, and / or a chipset (such as a Bluetooth TM equipment, 802.11 equipment, WiFi equipment, WiMax equipment, cellular communication facilities, etc.).

[0089] The present disclosure may be implemented as any combination of an apparatus, method, system, integrated circuit, and a computer program on a non-transitory computer-readable medium. One or more processors may be implemented as an integrated circuit (IC), an application specific integrated circuit (ASIC), or a large-scale integrated circuit (LSI), a system LSI, a super LSI, or a super LSI component that performs some or all of the functions described in the present disclosure.

[0090] The present disclosure includes the use of software, applications, computer programs, or algorithms. The software, applications, computer programs, or algorithms may be stored on a non-transitory computer-readable medium to enable a computer such as one or more processors to perform the above steps and the steps described in the accompanying drawings. For example, one or more memories store the software or algorithm in executable instructions, and one or more processors may associate a set of instructions to execute the software or algorithm to provide various functions according to the embodiments described in the present disclosure.

[0091] Software and computer programs (which may also be referred to as programs, software applications, applications, components or code) include machine instructions for a programmable processor and may be implemented in a high-level procedural language, an object-oriented programming language, a functional programming language, a logic programming language, or an assembly language or a machine language. The term "computer-readable medium" refers to any computer program product, apparatus or device for providing machine instructions or data to a programmable data processor, such as magnetic disks, optical disks, solid-state storage devices, memories, and programmable logic devices (PLDs), including computer-readable media that receive machine instructions as computer-readable signals.

[0092] For example, computer readable media may include dynamic random access memory (DRAM), random access memory (RAM), read-only memory (ROM), electrically erasable read-only memory (EEPROM), compact disk read-only memory (CD-ROM) or other optical disk storage devices, magnetic disk storage devices or other magnetic storage devices, or any other medium that can be used to carry or store the desired computer readable program code in the form of instructions or data structures and can be accessed by a general or special purpose computer or a general or special purpose processor. As used herein, disk or disc includes compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, where disks usually reproduce data magnetically, while discs reproduce data optically by laser. Combinations of the above are also included in the scope of computer readable media.

[0093] The subject matter of the present disclosure is provided as an example of a device, method, program, and storage medium for performing the features described in the present disclosure. However, in addition to the above features, other features or variations may also be expected. It is expected that the implementation of the components and functions of the present disclosure may be accomplished with any emerging technology that may replace any of the above-mentioned implementation technologies.

[0094] In addition, the above description provides examples without limiting the scope, applicability or configuration set forth in the claims. Without departing from the spirit and scope of the present disclosure, the function and arrangement of the elements discussed may be changed. Various embodiments may appropriately omit, replace or add various processes or parts. For example, features described with respect to certain embodiments may be combined in other embodiments.

[0095] In addition, in the description of the present disclosure, the terms “first”, “second”, “third”, etc. are used for descriptive purposes only and cannot be understood as indicating or implying relative importance and order.

[0096] Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed to achieve desired results. In certain circumstances, multitasking and parallel processing can be advantageous.

Claims

1. A method for identifying a cloud dedicated line customer service application scenario, comprising: A data acquisition step, acquiring object data for business application scenario identification; a feature adding step of adding flow-related features and time-related features to the object data acquired in the data acquiring step; Feature engineering processing steps: selecting feature groups for object data obtained by adding features related to flow and features related to time; as well as The strategy selection step selects a model for business application scenario identification based on the feature group selected in the feature engineering processing step, including: judging the type of the circuit according to the parameters related to the circuit type contained in the feature group, judging which category of circuit the circuit belongs to according to the value of the circuit type attribute, and selecting the corresponding model based on the label information corresponding to the judged circuit.

2. The method for identifying application scenarios of cloud dedicated line customer services according to claim 1, wherein: Also includes: The model application step applies the model selected in the strategy selection step to the cloud dedicated line customer service application scenario identification.

3. The cloud dedicated line customer service application scenario identification method according to claim 2, wherein: In the model application step, the labels output after the cloud dedicated line customer service application scenario identification are fed back to the strategy selection step as training data for training the model.

4. The method for identifying the application scenario of a cloud dedicated line customer service according to any one of claims 1 to 3, wherein: The flow-related features may be any of the following features: flow difference, proportion of incoming flow in different time periods, proportion of non-zero number of incoming flow in different time periods, proportion of outgoing flow in different time periods, and proportion of non-zero number of outgoing flow in different time periods.

5. The method for identifying the application scenario of a cloud dedicated line customer service according to any one of claims 1 to 3, wherein: When acquiring object data for business application scenario identification in the data acquisition step, feature data selection, data cleaning and data association processing are performed, The feature data selection includes the selection of customer features, circuit features and traffic features. The data cleaning is the process of eliminating unnecessary data that may cause interference. The data association is the process of associating the circuit ID used for cloud dedicated line customer services with existing label data.

6. The method for identifying the application scenario of a cloud dedicated line customer service according to any one of claims 1 to 3, wherein: In the feature engineering processing step, the final feature group is selected based on the Pearson correlation coefficient method.

7. The method for identifying the application scenario of a cloud dedicated line customer service according to any one of claims 1 to 3, wherein: The tag information is a tag ratio value.

8. A cloud dedicated line customer service application scenario identification device, comprising: A data acquirer, which acquires object data for business application scenario identification; a feature appender, which appends flow-related features and time-related features to the object data acquired by the feature appender; A feature engineering processor selects a feature group from the object data obtained by appending features related to flow and features related to time; as well as A strategy selector selects a model for business application scenario identification based on the feature group selected by the feature engineering processor, including: The circuit type is determined based on the parameters related to the circuit type contained in the feature group. Which type? The circuit type to which the circuit belongs is determined according to the value of the circuit type attribute, and the corresponding model is selected based on the label information corresponding to the determined circuit.

9. The cloud dedicated line customer service application scenario identification device according to claim 8, wherein: Also includes: The model applicator applies the model selected by the policy selector to identify the application scenarios of cloud dedicated line customer services.

10. The cloud dedicated line customer service application scenario identification device according to claim 9, wherein: The model applicator feeds back the labels outputted after identifying the application scenarios of cloud dedicated line customer services to the strategy selector as training data for training the model.

11. The cloud dedicated line customer service application scenario identification device according to any one of claims 8 to 10, wherein: The flow-related features may be any of the following features: flow difference, proportion of incoming flow in different time periods, proportion of non-zero number of incoming flow in different time periods, proportion of outgoing flow in different time periods, and proportion of non-zero number of outgoing flow in different time periods.

12. The cloud dedicated line customer service application scenario identification device according to any one of claims 8 to 10, wherein: When the data acquirer acquires object data for business application scenario identification, it performs feature data selection, data cleaning, and data association processing. The feature data selection includes the selection of customer features, circuit features and traffic features. The data cleaning is the process of eliminating unnecessary data that may cause interference. The data association is the process of associating the circuit ID used for cloud dedicated line customer services with existing label data.

13. The cloud dedicated line customer service application scenario identification device according to any one of claims 8 to 10, wherein: The feature engineering processor selects the final feature group based on the Pearson correlation coefficient method.

14. The cloud dedicated line customer service application scenario identification device according to any one of claims 8 to 10, wherein: The tag information is a tag ratio value.

15. A computer-readable storage medium, comprising computer-executable instructions, which, when executed by one or more processors, enable the one or more processors to execute the cloud dedicated line customer service application scenario identification method according to any one of claims 1 to 7.

Images