A data processing method and apparatus

Abstract

The application discloses a kind of data processing method and device, it is related to computer technical field.The specific embodiment of the method includes obtaining the consultation request initiated by customer terminal, determines corresponding incoming line period, statistics the target interaction data of the customer terminal in the incoming line period, finds the target historical data corresponding to the customer terminal in database;According to the data processing rule of pre-set, the target interaction data and the target historical data are combined and handled, corresponding vector is obtained, to input to trained classification model, obtain the array of output, determine the serial number corresponding to target element in the array;Call the lookup table of pre-set to determine the customer service terminal corresponding to the serial number, to transfer the consultation request to the customer service terminal.Thereby, the embodiment of the application can solve the technical problem that existing customer online consultation waiting time is long.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a data processing method and apparatus. Background Technology

[0002] Currently, online customer service systems are widely used, providing convenient and professional consultation services for online customers and breaking through the limitations of time and location to respond to customer needs.

[0003] In the process of realizing this invention, the inventors discovered at least the following problems in the prior art:

[0004] While customers enjoy the convenience and professional consultation services of online customer service, long waiting times for online consultations are a common problem. This is because current technology requires matching the customer's sent inquiry description with the appropriate customer service representative, transferring the customer's inquiry page to that representative for a response. Therefore, after sending their question, customers must wait for the description to be parsed and for a representative to be matched, a process that is typically time-consuming, resulting in long online consultation wait times and a poor customer experience. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a data processing method and apparatus that can solve the technical problem of long waiting times for online customer consultations.

[0006] To achieve the above objectives, according to one aspect of the present invention, a data processing method is provided, comprising: acquiring a consultation request initiated by a customer terminal; determining a corresponding call-in period; statistically analyzing target interaction data of the customer terminal during the call-in period; searching for target historical data corresponding to the customer terminal in a database; merging the target interaction data and the target historical data according to preset data processing rules to obtain a corresponding vector, inputting it into a trained classification model to obtain an output array; determining the sequence number corresponding to the target element in the array; and retrieving a preset lookup table to determine the customer service terminal corresponding to the sequence number, so as to transfer the consultation request to the customer service terminal.

[0007] Optionally, the corresponding vector is obtained and input into the trained classification model, including:

[0008] Determine whether a preset target value exists among the vector elements included in the vector;

[0009] If so, the inquiry request will be transferred to the designated customer service terminal.

[0010] If not, the vector is then input into the trained classification model.

[0011] Optionally, before determining the index corresponding to the target element in the array, the process includes:

[0012] Determine whether there is a maximum value among the multiple array elements corresponding to the given array;

[0013] If so, then the element corresponding to the maximum value is determined as the target element;

[0014] Otherwise, the index corresponding to the target element is set to the default value.

[0015] Optionally, before retrieving the historical data corresponding to the client terminal from the database, the process includes:

[0016] Establish a database to store historical interaction data, historical transfer records, and historical consultation data for each client terminal up to a specified date.

[0017] Optionally, it includes:

[0018] The system receives a user terminal's request to cancel a consultation, uses the target interaction data corresponding to the consultation request as historical interaction data, uses the customer service terminal that the consultation request was finally matched as a historical transfer record, collects the corresponding consultation data as historical consultation data, and records it in the database.

[0019] Optionally, before being input into the trained classification model, the following are included:

[0020] Create a lookup table to determine the serial number corresponding to each customer service terminal, and record the mapping relationship between each customer service terminal and its corresponding serial number in the lookup table.

[0021] Optionally, training a classification model includes:

[0022] For each historical transfer record in the database

[0023] Retrieve the preset lookup table to determine the serial number of the matching customer service terminal;

[0024] Historical interaction data of the corresponding customer terminal during the corresponding call-in period is selected from the database, along with historical transfer records and historical consultation data within a specified period before the call-in period. The historical interaction data, historical transfer records, and historical consultation data are merged according to preset data processing rules to obtain the corresponding vector.

[0025] Associating the sequence number with the vector yields a training dataset, wherein the vector is the training data and the sequence number is the target data;

[0026] We obtained multiple training datasets corresponding to various historical transfer records in the database.

[0027] Optionally, after obtaining multiple training datasets corresponding to multiple historical transfer records in the database, the following are included:

[0028] A classification model is constructed by inputting the multiple training datasets into the classification model one by one, with the goal of making the index of the target element in the output array corresponding to each training dataset approach the target data in the same training dataset, and training the classification model to minimize the corresponding loss function.

[0029] Optionally, the target interaction data and the target historical data are merged according to preset data processing rules, including:

[0030] The data to be processed is determined to be the target interaction data and the target historical data;

[0031] Based on the incoming time period of the data to be processed, the target date corresponding to the data to be processed is determined, and date data in the data to be processed is selected. The time difference between each date data and the target date is calculated, and each date data is replaced with the corresponding time difference.

[0032] Select data with a limited range of values ​​from the data to be processed, retrieve a preset dictionary to determine the mapping value corresponding to the data, and use the mapping value to replace the corresponding data.

[0033] According to a preset order, the replaced data to be processed are merged into a vector.

[0034] In addition, the present invention also provides a data processing device, including a collection module for acquiring consultation requests initiated by customer terminals, determining the corresponding call-in period, statistically analyzing the target interaction data of the customer terminals during the call-in period, and searching for the target historical data corresponding to the customer terminals in a database; a classification module for merging the target interaction data and the target historical data according to preset data processing rules to obtain a corresponding vector, which is then input into a trained classification model to obtain an output array and determine the sequence number corresponding to the target element in the array; and a transfer module for retrieving a preset lookup table to determine the customer service terminal corresponding to the sequence number, so as to transfer the consultation request to the customer service terminal.

[0035] One embodiment of the above invention has the following advantages or beneficial effects: The present invention determines the corresponding call-in period by acquiring the consultation request initiated by the customer terminal, thus completing the determination of the corresponding customer terminal information collection period and laying the foundation for targeted analysis of the customer's current status in subsequent operations. Furthermore, the present invention obtains the target data of the corresponding customer terminal by statistically analyzing the target interaction data of the customer terminal during the call-in period and searching for the corresponding target historical data of the customer terminal in the database, providing data support for accurately matching customer service terminals in subsequent operations. Simultaneously, the present invention merges the target interaction data and the target historical data according to preset data processing rules to obtain corresponding vectors, realizing a unified data organization process for the collected customer terminal target data, achieving the transformation of chaotic data into vectors with a fixed structure. The input method provides a standardized model input for subsequent operations. Furthermore, by inputting the vector into a trained classification model to obtain an output array, and determining the index of the target element in the array, this invention completes the process of predicting the customer service terminal corresponding to the aforementioned customer terminal using a preset classification model. It also completes the initial processing of the corresponding model output. This process only requires partial target data of the corresponding customer terminal without the consultation description sent by the customer terminal, thus achieving the effect of matching the corresponding customer service terminal before the customer initiates a consultation description, thereby minimizing the customer's online consultation waiting time. Finally, this invention retrieves a preset lookup table to determine the customer service terminal corresponding to the index, transferring the consultation request to that customer service terminal, thus completing the association between the customer terminal and the matched customer service terminal and responding to the consultation request initiated by the customer terminal.

[0036] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0037] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein:

[0038] Figure 1 This is a schematic diagram of the main flow of the data processing method according to the first embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of the main flow of the data processing method according to the second embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the main flow of the data processing method according to the third embodiment of the present invention;

[0041] Figure 4 This is a schematic diagram of the main modules of a data processing apparatus according to a first embodiment of the present invention;

[0042] Figure 5 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied;

[0043] Figure 6 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Detailed Implementation

[0044] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the acquisition, storage, and application of user personal information involved in the technical solutions of this disclosure comply with relevant laws and regulations and do not violate public order and good morals.

[0045] Figure 1 This is a schematic diagram of the main flow of the data processing method according to the first embodiment of the present invention, as shown below. Figure 1 As shown, the data processing method includes:

[0046] Step S101: Obtain the consultation request initiated by the client terminal, determine the corresponding call time period, count the target interaction data of the client terminal during the call time period, and search for the target historical data corresponding to the client terminal in the database.

[0047] In this embodiment, acquiring a consultation request initiated by a customer terminal can mean listening for the customer terminal opening the corresponding online customer service consultation page. That is, responding immediately upon detecting the customer terminal's intention to conduct online consultation, rather than waiting until the customer completes their consultation description. This achieves the effect of matching a customer service terminal with a suitable service provider before the customer terminal sends their consultation description, thereby minimizing customer waiting time and effectively enhancing the customer experience. Furthermore, by determining the corresponding call-in time period of the customer terminal, and statistically analyzing the target data corresponding to the customer terminal based on that time period, it helps to accurately analyze the current state of the customer terminal, thus allocating a suitable customer service terminal to it. In some embodiments, the target data may include target interaction data of the customer terminal during the call-in period, such as call-in method data (e.g., service phone, WeChat, application entry, etc.), shopping cart data, order data, logistics data, payment / refund data, complaint handling data, etc. In addition, the target data may also include target historical data corresponding to the customer terminal, such as historical transfer records (i.e., the customer service group ultimately transferred in the most recent inquiries) and historical consultation data (e.g., the time of the most recent inquiries, call-in method, consultation content, product number / order number / logistics tracking number, etc.). In some embodiments, the target data may also include basic information data corresponding to the customer terminal, such as username, gender, age, region, customer level, etc. In some embodiments, to better adapt to business needs, the target data may also include other customer status / recent activities / external system data, depending on the actual situation and development changes of the business system. This step provides data support for predicting customer needs and matching customer service terminals in subsequent operations.

[0048] In some embodiments, in order to accurately and comprehensively record the data corresponding to each customer terminal, a database can be established to store the historical interaction data, historical transfer records and historical consultation data corresponding to each customer terminal before a specified date, thereby laying a sufficient data foundation for predicting the preferences and needs of each customer terminal in subsequent business.

[0049] Step S102: Merge the target interaction data and the target historical data according to the preset data processing rules to obtain a corresponding vector, which is then input into the trained classification model to obtain an output array, and determine the index corresponding to the target element in the array.

[0050] In this embodiment, by processing the target data corresponding to the customer terminal obtained from statistics into corresponding vectors, the data cleaning and data format conversion process is realized. At the same time, the input format of the classification model is unified, which helps the classification model to focus more on the differences between different data values ​​corresponding to different customer terminals when parsing the data, rather than the differences caused by different data formats.

[0051] In some embodiments, to generate numeric type tags corresponding to different customer terminals, a lookup table can be established to determine the serial number corresponding to each customer service terminal, so as to record the mapping relationship between each customer service terminal and the corresponding serial number in the lookup table. In large enterprises, due to the complexity of business, customer service often tends to be specialized, that is, each customer service representative is only good at answering business questions in a certain area. Therefore, it is necessary to establish different numeric type tags corresponding to different customer service representatives to distinguish them. The setting of numeric tags enables the customer service matching system of the present invention to more intuitively and quickly query the corresponding customer service terminal based on the output of the classification model, and the form of the lookup table is also easy to maintain and modify to adapt to different business needs. It is worth noting that the serial number corresponding to each customer service terminal in the present invention can also be understood as the serial number corresponding to each type of customer service terminal. That is, after matching the customer service category corresponding to the customer terminal, a customer service representative is randomly selected from multiple customer service representatives of the same category to respond to the customer terminal's inquiry request.

[0052] In some embodiments, to build a sufficient training dataset before training the classification model, data support from a database can be utilized. Specific operations include: for each historical transfer record in the database, retrieving a preset lookup table to determine the sequence number of the matching customer service terminal; selecting historical interaction data of the corresponding customer terminal during the corresponding call-in period, and historical transfer records and historical consultation data within a specified period before the call-in period from the database; merging the historical interaction data, historical transfer records, and historical consultation data according to preset data processing rules to obtain a corresponding vector; associating the sequence number with the vector to obtain a set of training datasets, wherein the vector is the training data and the sequence number is the target data; and obtaining multiple sets of training datasets corresponding to multiple historical transfer records in the database. In both the training and application phases of the classification model, the collection and processing of input data are similar. First, the incoming call time slot corresponding to each set of input data is determined. Then, user terminal interaction data within the incoming call time slot and historical user terminal data (including historical transfer records and historical consultation data) within a specified period prior to the incoming call time slot are located. In addition, when constructing the training data, a corresponding target value needs to be determined, i.e., the direction of the training model. In this invention, this corresponds to the transfer records of the incoming call time slot, i.e., the ultimately matched customer service terminal. Through this process, multiple training datasets are obtained based on the database, thereby maximizing the value of the stored data assets.

[0053] In some embodiments, to obtain an accurate classification model, a classification model can be constructed by sequentially inputting multiple training datasets into the model. The goal is to train the model so that the index of the target element in the output array corresponding to each training dataset approximates the target data in the same training dataset, thereby minimizing the corresponding loss function. In practice, the above classification model can be constructed based on algorithms such as XGBoost, logistic regression, LightGBM machine learning, or DNN deep learning. When the number of training datasets reaches a certain threshold, or the accuracy of the classification model reaches a certain threshold, the model can be considered a well-trained model and put into application. This step yields a classification model capable of predicting customer needs. It can predict customer needs and match customer service representatives in advance without requiring a description of the customer's inquiry, helping to reduce the waiting time for customer service matching and thus enhancing the customer experience.

[0054] In some embodiments, in order to organize and convert the collected customer data into a standard and unified format, the data can be processed using preset data processing rules. Specific operations include: merging the target interaction data and the target historical data according to the preset data processing rules, including: determining the data to be processed as the target interaction data and the target historical data; determining the target date corresponding to the data to be processed based on the inbound time period of the data to be processed, selecting date data from the data to be processed, calculating the time difference between each date data and the target date, and replacing each date data with the corresponding time difference; selecting data with a limited value range from the data to be processed, retrieving a preset dictionary to determine the mapping value corresponding to the data, and using the mapping value to replace the corresponding data; and merging the replaced data to be processed into a vector according to a preset order. The aforementioned date data, i.e., data of date / time type (e.g., date of birth, order date), can be converted into the number of days / years remaining from the target date / time for measurement and calculation. The aforementioned data with a limited value range can correspond to Boolean type data, i.e., the value range is only yes or no (e.g., whether a member is a member, whether an item is overdue), and the data will be converted through mapping (e.g., yes -> 1, no -> 0). The aforementioned data with a limited value range can also correspond to data of types such as gender and ethnicity, and similarly, its values ​​can be mapped to a numerical value through a dictionary (e.g., male -> 1, female -> 2). In addition, other commonly used data processing rules are also applicable to this invention, including: for numeric data (e.g., amount, weight), the original data is used directly; for identifier data, such as mobile phone numbers, order numbers, ID card numbers, and text fields such as addresses and remarks, the original format is preserved.

[0055] In some embodiments, to determine the corresponding customer service terminal based on the output of the classification model, it can be determined whether there is a maximum value among the multiple array elements corresponding to the array; if so, the element corresponding to the maximum value is determined as the target element; otherwise, the index corresponding to the target element is determined as the default value. In this invention, the value output by the classification model can be understood as the probability of matching the customer terminal to each customer service terminal. Therefore, the maximum value is the customer service terminal that best matches the customer terminal predicted by the classification model. However, there may be a situation where there is no maximum value, i.e., the classification model fails to match normally. In this case, the customer terminal is transferred to the default customer service terminal, which then transfers the customer again after obtaining the customer's inquiry description. This step improves the customer inquiry transfer mechanism, ensuring that each customer terminal that initiates an inquiry request can be transferred to a customer service terminal in a timely manner, avoiding the situation of ignoring or forgetting customers.

[0056] In some embodiments, to provide targeted customer service to relatively specific customer groups, when a corresponding vector is obtained and input into a trained classification model, it can be determined whether a preset target value exists among the vector elements. If so, the consultation request is transferred to a designated customer service terminal; otherwise, the vector is input into the trained classification model. In this step, the presence or absence of a target value in the vector can identify whether the corresponding customer terminal belongs to a special group, such as unregistered customers or customers who access the service through a complaint hotline. For special customers, classification model prediction is no longer performed; instead, the customer is directly transferred to a specific customer service terminal, such as new user reception or complaint handling customer service. This is because new customers lack sufficient historical data to build prediction data for the classification model, while complaint customers have clear needs and require targeted referrals. This step comprehensively caters to customers in special groups, and the targeted referral mechanism further improves the efficiency of online customer consultation.

[0057] Step S103: Retrieve a preset lookup table to determine the customer service terminal corresponding to the serial number, so as to transfer the inquiry request to the customer service terminal.

[0058] In some embodiments, to prevent inaccurate predictions by the classification model, the consultation request can be transferred to the customer service terminal, and the customer service terminal can then determine whether the consultation request has been transferred correctly; otherwise, the consultation request can be transferred to another matching customer service terminal through the customer service terminal.

[0059] In some embodiments, in order to continuously update customer data in the database, the user terminal can receive a cancellation request for consultation, use the target interaction data corresponding to the consultation request as historical interaction data, use the customer service terminal that the consultation request is finally matched as a historical transfer record, collect the corresponding consultation data as historical consultation data, and record it in the database accordingly.

[0060] In some embodiments, to further train the classification model and continuously improve its accuracy, a training dataset can be constructed based on the latest historical transfer records in the database. This training dataset is then input into the classification model to further train it. It is worth noting that the aforementioned historical transfer records should correspond to the customer service terminals that ultimately transferred the client terminal, to ensure the accuracy of the target values ​​in the constructed training dataset.

[0061] Figure 2 This is a schematic diagram of the main flow of a data processing method according to a second embodiment of the present invention, the data processing method including:

[0062] Step S201: Establish a database to store historical interaction data, historical transfer records, and historical consultation data for each customer terminal before a specified date.

[0063] Step S202: Establish a lookup table to determine the serial number corresponding to each customer service terminal, so as to record the mapping relationship between each customer service terminal and the corresponding serial number in the lookup table.

[0064] Step S203: Construct a training dataset based on the database.

[0065] Step S204: Build a classification model.

[0066] Step S205: Train the classification model using the training dataset.

[0067] Ideally, this minimizes the loss function of the classification model.

[0068] Step S206 yields the trained classification model.

[0069] Step S207: Obtain the consultation request initiated by the client terminal.

[0070] Step S208: Determine the corresponding incoming call time period, and then collect the target behavior data corresponding to the customer terminal based on the incoming call time period.

[0071] Step S209: Process the target behavior data according to preset data processing rules to obtain the corresponding vector.

[0072] Step S210: Determine whether the target value is included among the multiple vector elements corresponding to the vector. If yes, proceed to step S211; otherwise, proceed to step S212.

[0073] Step S211: Transfer the consultation request to the designated customer service terminal.

[0074] Step S212: Input the vector into the trained classification model to obtain the output array.

[0075] Step S213: Determine whether there is a maximum value among the multiple array elements corresponding to the array. If yes, proceed to step S214; otherwise, proceed to step S215.

[0076] Step S214: Determine the element corresponding to the maximum value as the target element.

[0077] Step S215: Determine the sequence number corresponding to the target element as the default value.

[0078] Step S216: Retrieve a preset lookup table to determine the customer service terminal corresponding to the serial number, so as to transfer the inquiry request to the customer service terminal.

[0079] Figure 3 This is a schematic diagram of the main flow of a data processing method according to a third embodiment of the present invention, the data processing method including:

[0080] Step S301: Establish a database to store historical interaction data, historical transfer records, and historical consultation data for each customer terminal before a specified date.

[0081] Step S302: Establish a lookup table to determine the serial number corresponding to each customer service terminal, so as to record the mapping relationship between each customer service terminal and the corresponding serial number in the lookup table.

[0082] Step S303: Construct multiple training datasets based on the database.

[0083] Preferably, for each historical transfer record in the database, a preset lookup table is retrieved to determine the sequence number of the matching customer service terminal; historical interaction data of the corresponding customer terminal during the corresponding call-in period is selected from the database, along with historical transfer records and historical consultation data within a specified period before the call-in period; the historical interaction data, historical transfer records, and historical consultation data are merged according to preset data processing rules to obtain a corresponding vector; the sequence number and the vector are associated to obtain a set of training datasets, wherein the vector is the training data and the sequence number is the target data; thus, multiple sets of training datasets corresponding to multiple historical transfer records in the database are obtained.

[0084] Step S304: Build a classification model.

[0085] Preferably, the classification model may be based on the XGboost algorithm, logistic regression algorithm, LightGBM machine learning algorithm, or DNN deep learning algorithm.

[0086] Step S305: Input the multiple training datasets into the classification model one by one, and train the classification model with the goal of the index of the target element in the output array corresponding to each training dataset being close to the target data in the same training dataset.

[0087] Ideally, this minimizes the loss function of the classification model.

[0088] Step S306: The trained classification model is obtained.

[0089] Step S307: Obtain the consultation request initiated by the client terminal.

[0090] Step S308: Determine the corresponding incoming call time period, and then collect the target behavior data corresponding to the customer terminal based on the incoming call time period.

[0091] Preferably, the target interaction data of the client terminal during the incoming call period is statistically analyzed, and the historical transfer records and historical consultation data of the client terminal within a specified period before the incoming call period are searched in the database, so as to use the target interaction data, historical transfer records and historical consultation data as target behavior data.

[0092] Step S309: Process the target behavior data according to preset data processing rules to obtain the corresponding vector.

[0093] Preferably, the data processing rules include: determining the corresponding data to be processed and the incoming time period; determining the target date corresponding to the incoming time period, selecting date data from the data to be processed, calculating the time difference between each date data and the target date, and replacing each date data with the corresponding time difference; selecting data with a limited value range from the data to be processed, retrieving a preset dictionary to determine the mapping value corresponding to the data, and using the mapping value to replace the corresponding data; and merging the replaced data to be processed into a vector according to a preset order.

[0094] Step S310: Input the vector into the trained classification model to obtain the output array, and determine the index corresponding to the target element in the array.

[0095] Step S311: Retrieve a preset lookup table to determine the customer service terminal corresponding to the serial number, so as to transfer the inquiry request to the customer service terminal.

[0096] Step S312: Determine whether the inquiry request has been correctly transferred through the customer service terminal. If not, proceed to step S313; if yes, proceed to step S314.

[0097] Step S313: Transfer the inquiry request to another matching customer service terminal through the customer service terminal.

[0098] Step S314: Receive the user terminal's request to cancel consultation, use the target interaction data corresponding to the consultation request as historical interaction data, use the customer service terminal that was finally matched by the consultation request as a historical transfer record, collect the corresponding consultation data as historical consultation data, and record it in the database accordingly.

[0099] Step S315: Construct a training dataset based on the latest historical transfer records in the database, and input the training dataset into the classification model to complete further training of the classification model.

[0100] Figure 4 This is a schematic diagram of the main modules of a data processing apparatus according to an embodiment of the present invention, such as... Figure 4As shown, the data processing device 400 includes a collection module 401, a classification module 402, and a transfer module 403. The collection module 401 acquires inquiries initiated by customer terminals, determines the corresponding call-in time period, statistically analyzes the target interaction data of the customer terminal during the call-in time period, and searches for the corresponding target historical data of the customer terminal in a database. The classification module 402 merges the target interaction data and the target historical data according to preset data processing rules to obtain a corresponding vector, which is then input into a trained classification model to obtain an output array, and determines the sequence number corresponding to the target element in the array. The transfer module 403 retrieves a preset lookup table to determine the customer service terminal corresponding to the sequence number, and transfers the inquiry request to that customer service terminal.

[0101] In some embodiments, the transfer module 403 is further configured to: when obtaining a corresponding vector to input into a trained classification model, determine whether there is a preset target value among the vector elements included in the vector; if yes, transfer the consultation request to a designated customer service terminal; if no, input the vector into the trained classification model.

[0102] In some embodiments, the classification module 402 is further configured to: determine whether there is a maximum value among the multiple array elements corresponding to the array before determining the index of the target element in the array; if so, determine the element corresponding to the maximum value as the target element; otherwise, determine the index of the target element as the default value.

[0103] In some embodiments, the collection module 401 is further configured to: establish a database to store historical interaction data, historical transfer records and historical consultation data corresponding to each customer terminal before querying the database to obtain the historical data corresponding to the customer terminal before a specified date.

[0104] In some embodiments, the collection module 401 is further configured to: receive the user terminal's request to cancel the consultation, use the target interaction data corresponding to the consultation request as historical interaction data, use the customer service terminal that was finally matched by the consultation request as a historical transfer record, collect the corresponding consultation data as historical consultation data, and record it in the database accordingly.

[0105] In some embodiments, the classification model 402 is further configured to: before inputting into the trained classification model, for each historical transfer record in the database, retrieve a preset lookup table to determine the sequence number of the matching customer service terminal; select historical interaction data of the corresponding customer terminal in the corresponding call-in period, and historical transfer records and historical consultation data within a specified period before the call-in period in the database, and merge the historical interaction data, historical transfer records, and historical consultation data according to preset data processing rules to obtain a corresponding vector; associate the sequence number and the vector to obtain a set of training datasets, wherein the vector is training data and the sequence number is target data; and obtain multiple sets of training datasets corresponding to multiple historical transfer records in the database.

[0106] In some embodiments, classification model 402 is further configured to: after obtaining multiple sets of training datasets corresponding to multiple historical transfer records in the database, construct a classification model, input the multiple sets of training datasets into the classification model one by one, and train the classification model with the goal of the index of the target element in the output array corresponding to each set of training datasets being close to the target data in the same set of training datasets, so that the corresponding loss function reaches the minimum.

[0107] In some embodiments, the classification module 402 is further configured to: when merging the target interactive data and the target historical data according to preset data processing rules, determine the data to be processed as the target interactive data and the target historical data; based on the incoming time period of the data to be processed, determine the target date corresponding to the data to be processed, select date data in the data to be processed, calculate the time difference between each date data and the target date respectively, and replace each date data with the corresponding time difference; select data with a limited value range in the data to be processed, retrieve a preset dictionary to determine the mapping value corresponding to the data, and use the mapping value to replace the corresponding data; and merge the replaced data to be processed into a vector according to a preset order.

[0108] It should be noted that the data processing method and the data processing device described in this invention have a corresponding relationship in their specific implementation, so repeated content will not be described again.

[0109] Figure 5 An exemplary system architecture 500 is shown that can be applied to the data processing method or data processing apparatus of the present invention.

[0110] like Figure 5As shown, system architecture 500 may include terminal devices 501, 502, and 503, a network 504, and a server 505. Network 504 serves as the medium for providing communication links between terminal devices 501, 502, and 503 and server 505. Network 504 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0111] Users can use terminal devices 501, 502, and 503 to interact with server 505 via network 504 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 501, 502, and 503.

[0112] Terminal devices 501, 502, and 503 can be various electronic devices with a page display processing screen that supports web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0113] Server 505 can be a server that provides various services, such as a backend management server that supports users using terminal devices 501, 502, and 503 (for example only). The backend management server can analyze and process data such as received product information query requests, and feed back the processing results (such as target push information, product information - for example only) to the terminal device.

[0114] It should be noted that the data processing method provided in the embodiments of the present invention is generally executed by server 505, and correspondingly, the computing device is generally located in server 505.

[0115] It should be understood that Figure 5 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0116] The following is for reference. Figure 6 It shows a schematic diagram of the structure of a computer system 600 suitable for implementing a terminal device of the present invention. Figure 6 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0117] like Figure 6As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 602 or programs loaded from storage section 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the computer system 600. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0118] The following components are connected to I / O interface 605: an input section 605 including a keyboard, mouse, etc.; an output section 606 including a cathode ray tube (CRT), liquid crystal display processor (LCD), and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card and a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed into storage section 608 as needed.

[0119] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611. When the computer program is executed by central processing unit (CPU) 601, it performs the functions defined above in the system of this invention.

[0120] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0122] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be housed in a processor; for example, a processor may be described as including a collection module, a sorting module, and a transfer module. The names of these modules do not necessarily limit the functionality of the module itself.

[0123] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to: acquire a consultation request initiated by a customer terminal; determine the corresponding call-in period; statistically analyze the target interaction data of the customer terminal during the call-in period; search for the target historical data corresponding to the customer terminal in a database; merge the target interaction data and the target historical data according to preset data processing rules to obtain a corresponding vector, input it into a trained classification model to obtain an output array, determine the sequence number corresponding to the target element in the array; and retrieve a preset lookup table to determine the customer service terminal corresponding to the sequence number, so as to transfer the consultation request to the customer service terminal.

[0124] The technical solution according to the embodiments of the present invention can solve the technical problem of long waiting time for online customer consultation.

[0125] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A data processing method, characterized in that, The method for matching a customer terminal with a customer service terminal before the customer terminal sends an inquiry description includes: The system acquires consultation requests initiated by a client terminal, determines the corresponding call-in period, statistically analyzes the target interaction data of the client terminal during the call-in period, and searches the database for the target historical data corresponding to the client terminal; wherein, acquiring consultation requests initiated by a client terminal refers to listening to the client terminal opening the corresponding online customer service consultation page. According to preset data processing rules, the target interaction data and the target historical data are merged to obtain a corresponding vector, which is then input into a trained classification model to obtain an output array. The index of the target element in the array is determined. Multiple training datasets are input into the classification model group by group, with the goal of training the model to ensure that the index of the target element in the output array of each training dataset approximates the target data in the same training dataset. The array includes the probability of matching the customer terminal to each customer service terminal. A preset lookup table is retrieved to determine the customer service terminal corresponding to the serial number, so as to transfer the inquiry request to the customer service terminal; The step of merging the target interactive data and the target historical data according to preset data processing rules includes: determining the data to be processed as the target interactive data and the target historical data; determining the target date corresponding to the data to be processed based on the incoming time period of the data to be processed, selecting date data in the data to be processed, calculating the time difference between each date data and the target date respectively, and replacing each date data with the corresponding time difference; selecting data with a limited value range in the data to be processed, retrieving a preset dictionary to determine the mapping value corresponding to the data, and using the mapping value to replace the corresponding data; and merging the replaced data to be processed into a vector according to a preset order.

2. The method according to claim 1, characterized in that, The corresponding vectors are obtained and input into the trained classification model, including: Determine whether a preset target value exists among the vector elements included in the vector; If so, the inquiry request will be transferred to the designated customer service terminal. If not, the vector is then input into the trained classification model.

3. The method according to claim 1, characterized in that, Before determining the index of the target element in the array, the following steps are included: Determine whether there is a maximum value among the multiple array elements corresponding to the given array; If so, then the element corresponding to the maximum value is determined as the target element; Otherwise, the index corresponding to the target element is set to the default value.

4. The method according to claim 1, characterized in that, Before retrieving the historical data corresponding to the client terminal from the database, the process includes: Establish a database to store historical interaction data, historical transfer records, and historical consultation data for each client terminal up to a specified date.

5. The method according to claim 4, characterized in that, After transferring the inquiry request to the customer service terminal, the process includes: The system receives user terminal cancellation requests, uses the target interaction data corresponding to the cancellation request as historical interaction data, uses the customer service terminal matched with the cancellation request as historical transfer records, collects the corresponding consultation data as historical consultation data, and records it in the database.

6. The method according to claim 1, characterized in that, The process of constructing the multiple training datasets includes: For each historical transfer record in the database: Retrieve the preset lookup table to determine the serial number of the matching customer service terminal. Historical interaction data of the corresponding customer terminal during the corresponding call-in period is selected from the database, along with historical transfer records and historical consultation data within a specified period prior to the call-in period. These historical interaction data, historical transfer records, and historical consultation data are then merged according to preset data processing rules to obtain a corresponding vector. Associating the sequence number with the vector yields a training dataset, wherein the vector is the training data and the sequence number is the target data; This allows us to obtain multiple training datasets corresponding to various historical transfer records in the database.

7. The method according to claim 1 or 6, characterized in that, The method further includes: A classification model is constructed with the goal of making the index of the target element in the output array corresponding to each training dataset approach the target data in the same training dataset. The classification model is then trained to minimize the corresponding loss function.

8. A data processing apparatus, characterized in that, The device is used to match a customer terminal with a customer service terminal before the customer terminal sends an inquiry description, the device comprising: The collection module is used to acquire consultation requests initiated by customer terminals, determine the corresponding call-in period, collect target interaction data of the customer terminals during the call-in period, and search for the target historical data corresponding to the customer terminals in the database; wherein, acquiring consultation requests initiated by customer terminals refers to listening to the operation of the customer terminals opening the corresponding online customer service consultation page; The classification module is used to merge the target interaction data and the target historical data according to preset data processing rules to obtain a corresponding vector, which is then input into a trained classification model to obtain an output array and determine the index of the target element in the array. Multiple training datasets are input into the classification model group by group, with the goal of training the model to ensure that the index of the target element in the output array of each training dataset approximates the target data in the same training dataset. The array includes the probability of matching the customer terminal to each customer service terminal. The step of merging the target interactive data and the target historical data according to preset data processing rules includes: determining the data to be processed as the target interactive data and the target historical data; determining the target date corresponding to the data to be processed based on the incoming time period of the data to be processed, selecting date data in the data to be processed, calculating the time difference between each date data and the target date respectively, and replacing each date data with the corresponding time difference; selecting data with a limited value range in the data to be processed, retrieving a preset dictionary to determine the mapping value corresponding to the data, and using the mapping value to replace the corresponding data; and merging the replaced data to be processed into a vector according to a preset order. The transfer module is used to retrieve a preset lookup table to determine the customer service terminal corresponding to the serial number, so as to transfer the inquiry request to the customer service terminal.

9. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images