Outbound list generation method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides an outbound list generation method and device, electronic equipment and a storage medium, belongs to the technical field of artificial intelligence, and is suitable for financial science and technology scenes. The method comprises the following steps: acquiring candidate outbound objects from a preset user resource pool, and acquiring original object information of the candidate outbound objects; carrying out feature extraction on the original object information to obtain candidate behavior features of candidate outbound objects; performing conversion rate prediction on the candidate behavior characteristics through a preset target conversion rate prediction model to obtain a target prediction conversion rate of the candidate outbound object; obtaining candidate seat objects from a preset seat resource pool; performing model construction based on the candidate outbound object, the candidate seat object and the target prediction conversion rate to obtain an outbound list planning model; and performing list planning through the outbound list planning model to obtain an outbound user list of the candidate seat object. According to the embodiment of the invention, the call-out efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology and is applicable to financial technology scenarios, and in particular to a method and device for generating an outbound call list, an electronic device, and a storage medium. Background Art

[0002] Outbound calling is an operational method that proactively establishes contact with target customers through telephone channels and conducts interactive activities around business processes such as product promotion, service consultation, customer return visits, and sales follow-up. Outbound calling can be applied in various scenarios. For example, in FinTech scenarios, outbound calling can be used to promote, sell, and return customers of insurance and financial products to improve the efficiency of financial services and the customer experience.

[0003] Currently, outbound calls primarily rely on human agents, who independently select and make calls based on a user list. However, in real-world scenarios, agents spend a significant amount of time evaluating the potential value of users on the list to select target users for outbound calls, which impacts outbound call efficiency.

[0004] Therefore, how to improve the efficiency of outbound calls has become a technical problem that needs to be solved urgently. Summary of the Invention

[0005] The main purpose of the embodiments of the present application is to provide a method and device for generating an outbound call list, an electronic device and a storage medium, aiming to improve the efficiency of outbound calls.

[0006] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application provides a method for generating an outbound call list, the method comprising:

[0007] Obtaining a candidate outbound call object from a preset user resource pool, and obtaining original object information of the candidate outbound call object;

[0008] Extracting features from the original object information to obtain candidate behavioral features of the candidate outbound call object;

[0009] Perform conversion rate prediction on the candidate behavior characteristics using a preset target conversion rate prediction model to obtain a target predicted conversion rate of the candidate outbound call object;

[0010] Get candidate agent objects from the preset agent resource pool;

[0011] A model is constructed based on the candidate outbound call objects, the candidate agent objects, and the target predicted conversion rate to obtain an outbound call list planning model;

[0012] The outbound call list planning model is used to plan the list to obtain the outbound call user list of the candidate agent object.

[0013] In some embodiments, constructing a model based on the candidate outbound call objects, the candidate agent objects, and the target predicted conversion rate to obtain an outbound call list planning model includes:

[0014] An outbound call decision feature is constructed based on the candidate outbound call object and the candidate agent object; wherein the outbound call decision feature is used to indicate whether the candidate agent object makes an outbound call to the candidate outbound call object;

[0015] A model is constructed based on the outbound call decision characteristics and the target predicted conversion rate to obtain the outbound call list planning model.

[0016] In some embodiments, performing list planning using the outbound call list planning model to obtain the outbound call user list of the candidate agent object includes:

[0017] Obtain the outbound call count constraint of the candidate agent object;

[0018] Obtaining the object time constraint and answering number constraint of the candidate outbound call object;

[0019] Obtaining a preset agent outbound call time window, and generating an outbound call time constraint based on the agent outbound call time window;

[0020] Performing linear programming on the outbound call list planning model based on the outbound call number constraint, the object time constraint, the answering number constraint, and the outbound call time constraint to obtain an original outbound call list of the candidate agent object;

[0021] The original outbound call list is sorted to obtain the outbound call user list of the candidate agent object.

[0022] In some embodiments, the original outbound call list includes candidate outbound call recipients, where the candidate outbound call recipients are used to indicate recipients to whom the candidate agent may make outbound calls; and sorting the original outbound call list to obtain the outbound call user list of the candidate agent may include:

[0023] Obtaining the object time constraint corresponding to the candidate outbound call object;

[0024] The candidate outbound call objects are grouped based on the agent outbound call time window and the object time constraint to obtain multiple initial outbound call lists; wherein each of the initial outbound call lists corresponds to one of the agent outbound call time windows;

[0025] For each of the initial outbound call lists, sorting the initial outbound call lists based on the target predicted conversion rate to obtain a plurality of candidate outbound call lists;

[0026] The outbound call user list of the candidate agent object is generated based on a plurality of the candidate outbound call lists.

[0027] In some embodiments, extracting features from the original object information to obtain candidate behavioral features of the candidate outbound call object includes:

[0028] For each candidate outbound call object, performing data cleaning on the original object information to obtain initial object information;

[0029] Performing data encoding on the initial object information to obtain encoded object information;

[0030] Standardizing the encoded object information to obtain standardized object information;

[0031] Performing feature screening on the standardized object information to obtain original object features;

[0032] Feature construction is performed on the original object features to obtain the candidate behavior features of the candidate outbound call object.

[0033] In some embodiments, performing data cleaning on the original object information to obtain initial object information includes:

[0034] Screening the original object information for missing information to obtain missing object information;

[0035] Performing abnormal information detection on the original object information to obtain abnormal object information;

[0036] Information repair is performed on the missing object information and the abnormal object information in the original object information to obtain the initial object information.

[0037] In some embodiments, before performing conversion rate prediction on the candidate behavior features using a preset target conversion rate prediction model to obtain the target predicted conversion rate of the candidate outbound call recipient, the method further includes: training the target conversion rate prediction model;

[0038] The target conversion rate prediction model is trained in the following way:

[0039] Acquire sample object data, wherein the sample object data includes sample object information and sample object tags of a plurality of sample objects, wherein the sample object tags are used to indicate whether the sample objects are converted after the outbound call;

[0040] Performing feature extraction on the sample object information to obtain sample object features;

[0041] Performing conversion rate prediction on the sample object features using a preset original conversion rate prediction model to obtain a sample predicted conversion rate for each sample object;

[0042] Determining a sample predicted conversion label according to the sample predicted conversion rate;

[0043] Perform loss calculation based on the sample object label and the sample prediction conversion label to obtain a sample prediction loss function;

[0044] The original conversion rate prediction model is optimized based on the sample prediction loss function to obtain the target conversion rate prediction model.

[0045] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides an outbound call list generation device, the device comprising:

[0046] An outbound call object acquisition module is used to obtain candidate outbound call objects from a preset user resource pool and obtain original object information of the candidate outbound call objects;

[0047] An object feature extraction module is used to extract features from the original object information to obtain candidate behavioral features of the candidate outbound call object;

[0048] A conversion rate prediction module is used to predict the conversion rate of the candidate behavior characteristics using a preset target conversion rate prediction model to obtain a target predicted conversion rate of the candidate outbound call object;

[0049] The seat object acquisition module is used to obtain candidate seat objects from the preset seat resource pool;

[0050] A model building module, configured to build a model based on the candidate outbound call objects, the candidate agent objects, and the target predicted conversion rate to obtain an outbound call list planning model;

[0051] The list planning module is used to plan the list through the outbound call list planning model to obtain the outbound call user list of the candidate seat object.

[0052] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0053] To achieve the above-mentioned purpose, the fourth aspect of the embodiments of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the method described in the first aspect.

[0054] The outbound call list generation method and device, electronic device and storage medium proposed in this application obtain candidate outbound call objects from a preset user resource pool, obtain candidate behavioral characteristics of the candidate outbound call objects, and then use the candidate behavioral characteristics to predict the conversion rate. It can accurately evaluate the potential conversion possibility of each candidate outbound call object and obtain the target predicted conversion rate of each candidate outbound call object, providing a basis for subsequent list generation. Then, after obtaining the candidate agent object from the preset agent resource pool, a model is constructed based on the candidate outbound call object, the candidate agent object and the target predicted conversion rate to obtain an outbound call list planning model. Finally, the list is planned through the outbound call list planning model to obtain an outbound user list of the candidate agent object. The list is more reasonable and targeted, which is conducive to improving the efficiency of outbound calls, reducing invalid outbound calls, and at the same time improving the success rate of outbound calls, increasing business conversion volume, and thus improving overall business benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flowchart of the method for generating an outbound call list provided in an embodiment of the present application;

[0056] Figure 2 yes Figure 1 Flowchart of step S102 in FIG.

[0057] Figure 3 yes Figure 2 Flowchart of step S201 in FIG.

[0058] Figure 4 This is a flowchart of a method for generating an outbound call list provided in another embodiment of the present application;

[0059] Figure 5 yes Figure 1 Flowchart of step S105 in FIG.

[0060] Figure 6 yes Figure 1 Flowchart of step S106 in FIG.

[0061] Figure 7 yes Figure 6 Flowchart of step S605 in FIG.

[0062] Figure 8 This is a structural diagram of the outbound call list generation device provided in an embodiment of the present application;

[0063] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0065] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0067] First, let’s analyze some of the terms used in this application:

[0068] Artificial intelligence (AI) is a new technical discipline that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence. A branch of computer science, AI seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thinking. It also encompasses the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0069] Outbound calling is an operational method that proactively establishes contact with target customers through the telephone channel and conducts interactive activities around business processes such as product promotion, service consultation, customer return visits, and sales follow-up. The core purpose of outbound calling is to achieve sales conversion, service follow-up, or data collection through telephone communication. It is commonly used in fields such as telesales and customer service centers. Outbound calling relies on dialing systems (such as predictive dialing and automatic dialing) and list management (such as customer prioritization), and is subject to compliance requirements (such as avoiding harassment and avoiding off-hours).

[0070] Linear programming is a mathematical optimization method used to find the maximum or minimum value of a linear objective function under a set of linear constraints. Typical applications of linear programming include resource allocation, production planning, and transportation problems. Common solvers (such as CPLEX) are based on the simplex method or the interior point method.

[0071] Constraints are mathematical expressions that restrict the values of decision variables in linear programming, typically expressed as linear equations or inequalities. Constraints can be used for various purposes, such as resource limits and logical constraints. In dial list optimization, constraints may include dialing time windows and minimum agent capacity limits, ensuring that the optimal solution meets actual business rules.

[0072] Feature engineering is a key step in machine learning and data mining. It involves extracting, constructing, or selecting features (variables) from raw data that significantly impact the model's prediction target through domain knowledge, statistical methods, and data transformation techniques. This improves model performance, generalization, and interpretability. The core purpose of feature engineering is to transform raw data into a set of features that better represent the essence of the problem.

[0073] Outbound calls can be applied to multiple application scenarios. For example, in the FinTech scenario, outbound calls can be used to promote and sell insurance and financial products to users, conduct customer return visits, and improve financial service efficiency and customer experience.

[0074] Currently, outbound calls primarily rely on human agents, who independently select and make calls based on a user list. However, in real-world scenarios, user lists often contain information on tens of thousands of potential users. However, due to resource and time constraints, agents cannot follow up with every user on the list with equal intensity. Instead, they can only allocate dialing resources based on historical contact data and their own experience. This consumes a significant amount of time assessing the potential value of users on the list to select target users for outbound calls. This can result in missing out on users with high potential value and overinvesting in users with lower potential value, impacting outbound call efficiency.

[0075] Based on this, the embodiments of the present application provide a method and device for generating an outbound call list, an electronic device, and a storage medium, aiming to improve the efficiency of outbound calls.

[0076] The outbound call list generation method and device, electronic device and storage medium provided in the embodiments of the present application are specifically explained through the following embodiments. First, the outbound call list generation method in the embodiments of the present application is described.

[0077] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0078] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0079] The outbound call list generation method provided in the embodiment of the present application relates to the field of artificial intelligence technology. The outbound call list generation method provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the outbound call list generation method, etc., but is not limited to the above forms.

[0080] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0081] It should be noted that in each specific embodiment of the present application, when it comes to the need to perform relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of such data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0082] Figure 1 This is an optional flowchart of the method for generating an outbound call list provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S106.

[0083] Step S101: obtaining a candidate outbound call object from a preset user resource pool, and obtaining original object information of the candidate outbound call object;

[0084] Step S102: extracting features from the original object information to obtain candidate behavioral features of the candidate outbound call object;

[0085] Step S103, performing conversion rate prediction on the candidate behavior characteristics using a preset target conversion rate prediction model to obtain a target predicted conversion rate of the candidate outbound call recipient;

[0086] Step S104, obtaining candidate seat objects from a preset seat resource pool;

[0087] Step S105: constructing a model based on the candidate outbound call targets, candidate agent targets, and target predicted conversion rate to obtain an outbound call list planning model;

[0088] Step S106: Perform list planning using the outbound call list planning model to obtain an outbound call user list of candidate agents.

[0089] Steps S101 to S106 shown in the embodiment of the present application, by obtaining candidate outbound call objects from a preset user resource pool, and obtaining candidate behavioral characteristics of the candidate outbound call objects, and then using the candidate behavioral characteristics to predict the conversion rate, can accurately evaluate the potential conversion possibility of each candidate outbound call object, obtain the target predicted conversion rate of each candidate outbound call object, and provide a basis for subsequent list generation. Then, after obtaining the candidate agent object from the preset agent resource pool, a model is constructed based on the candidate outbound call object, the candidate agent object and the target predicted conversion rate to obtain an outbound call list planning model. Finally, the list is planned through the outbound call list planning model to obtain an outbound user list of the candidate agent object. The list is more reasonable and targeted, which is conducive to improving the efficiency of outbound calls, reducing invalid outbound calls, and at the same time improving the success rate of outbound calls, increasing business conversion volume, and thus improving overall business benefits.

[0090] In step S101 of some embodiments, the preset user resource pool is a pre-built collection of multiple candidate outbound call recipients. This preset user resource pool can be stored in a database, data warehouse, or other system and is used to provide a customer list for outbound calling services. The candidate outbound call recipients are customers in a specific business scenario. For example, in an insurance application scenario, the candidate outbound call recipients may be users who have registered on an insurance platform or purchased insurance products. The preset user resource pool includes users who have registered on the insurance platform or purchased insurance products.

[0091] Candidate outbound call targets are customers in specific business scenarios. For example, in the financial field, candidate outbound call targets can be users registered on the financial platform, users who have purchased insurance products, users who have purchased financial products, etc.

[0092] After determining the candidate outbound call recipients, the original recipient information corresponding to each candidate can be retrieved from the user information database. The user information database and the preset user resource pool can be stored in different data tables within the same database / repository. The original recipient information is the relevant information stored in the user information database for the candidate outbound call recipient, including but not limited to basic information, historical interaction data, and behavioral data. This data has not been processed or manipulated specifically for outbound call services and comprehensively reflects the basic situation and relevant characteristics of the candidate outbound call recipient.

[0093] For example, in an insurance application scenario, the original object information includes the basic information of the candidate outbound call object (such as age, gender, occupation, income level, family situation, etc.), historical interaction data (such as historical insurance product purchase records, complaint records, claims records, communication history, etc.), behavioral data (such as website browsing behavior, insurance APP usage behavior, etc.) and customer value (such as loyalty, price sensitivity, customer lifetime value), etc.

[0094] See also Figure 2 In some embodiments, step S102 may include but is not limited to steps S201 to S205:

[0095] Step S201: For each candidate outbound call target, clean the original target information to obtain initial target information;

[0096] Step S202, performing data encoding on the initial object information to obtain encoded object information;

[0097] Step S203, normalizing the encoding object information to obtain normalized object information;

[0098] Step S204, performing feature screening on the standardized object information to obtain original object features;

[0099] Step S205: constructing features of the original object to obtain candidate behavior features of the candidate outbound call object.

[0100] In steps S201 to S205 shown in the embodiment of the present application, for each candidate outbound call object, the original object information is cleaned to remove missing values, outliers, etc. in the original object information to obtain more accurate initial object information. Then, the initial object information is converted into coded object information by means of data encoding, which can unify the data format. Furthermore, by performing feature screening on the standardized object information, the original object features are obtained to ensure that the feature data is on the same scale. Finally, by performing feature screening and feature construction on the standardized object information, the candidate behavioral features of the candidate outbound call object are obtained, which helps to more accurately understand the characteristics of the candidate outbound call object, provide data support for outbound call decision-making, and thus improve the effect of outbound calls.

[0101] See also Figure 3 In some embodiments, step S201 may include but is not limited to steps S301 to S304:

[0102] Step S301, screening the original object information for missing information to obtain missing object information;

[0103] Step S302, performing abnormal information detection on the original object information to obtain abnormal object information;

[0104] Step S303: Repair the missing object information and abnormal object information in the original object information to obtain initial object information.

[0105] In steps S301 to S303 shown in the embodiment of the present application, missing object information and abnormal object information are obtained by screening the original object information for missing information and detecting abnormal information; and information repair is performed on the missing object information and abnormal object information in the original object information, which can fill data gaps and correct abnormal data, making the original object information more complete and accurate. The initial object information finally obtained is of higher quality, which can effectively reduce decision-making errors caused by data problems and help improve the accuracy of business decisions.

[0106] In step S301 of some embodiments, the original object information is screened for missing information based on preset missing rules to obtain missing object information, for example: the field value is empty, or the field value does not conform to the preset format (such as a date field, the field value is abcdefg).

[0107] In step S302 of some embodiments, all fields in the original object information are traversed and checked to find out the existing abnormal information and obtain abnormal object information.

[0108] For example, for a specific field, by counting all field values within that field, we obtain the first quartile (Q1) and the third quartile (Q3). Then, we calculate the interquartile range (IQR) based on the first and third quartiles, where IQR = Q3 - Q1. Next, we calculate the upper and lower bounds based on the first and third quartiles, where the upper bound = (Q3 + 1.5 * IQR) and the lower bound = (Q1 - 1.5 * IQR). Finally, for each field value within that field, any value below the lower bound or above the upper bound is considered an anomaly.

[0109] In step S303 of some embodiments, first, the missing object information and abnormal object information are deduplicated and integrated to obtain the object information to be processed; then, the object information to be processed is replaced based on the average value / median / mode corresponding to the field to which each object information to be processed belongs, thereby obtaining the repaired initial object information.

[0110] In another embodiment, a custom value may be used to replace the information of the object to be processed.

[0111] In step S202 of some embodiments, one-hot encoding or label encoding can be used to encode the discrete data in the initial object information and convert it into a form that can be processed by the model. For example, if the gender data is male or female, 0 can be used to represent male and 1 can be used to represent female.

[0112] In step S203 of some embodiments, a standardization method (such as Z-score standardization) or a normalization method (such as Min-Max normalization) is used to standardize features with large data magnitude differences in the encoded object information to ensure that these features are on the same scale.

[0113] In step S204 of some embodiments, a preset filter field is determined based on the business scenario, and then the standardized object information is subjected to feature filtering based on the preset filter field to obtain original object features. For example, in the insurance application scenario in the field of FinTech, original object features may include, but are not limited to: age, gender, occupation, income level, price sensitivity, customer lifetime value, customer attitude towards insurance, cumulative number of complaints, insurance product category, number of insurance products, total number of years of auto insurance coverage, maximum number of days in advance of insurance coverage, historical interaction behavior (number of inquiries / calls / call duration in the previous year / in the past six months), etc.

[0114] It should be noted that the preset filter fields are pre-set by business personnel based on business scenarios and have a significant impact on conversion rates. Specific settings need to be based on actual application scenarios and are not limited to these.

[0115] In step S205 of some embodiments, by constructing features for multiple fields in the original object features to form new features, the number of features can be simplified and the expressiveness of the features can be improved.

[0116] For example, multiple fields in the original object features can be combined to form new features, such as the best contact time (based on historical interaction behavior analysis), customer response tendency (based on historical behavior analysis), etc.

[0117] In addition, you can also cross-combine multiple fields in the original object features to form new features. For example, cross "gender" and "insurance product category" to generate the discrete feature "gender_insurance product category".

[0118] It can be understood that the embodiment illustrated by the above steps S204 to S205 is to perform feature engineering on the standardized object information to obtain candidate behavioral features that can characterize the characteristics of the candidate outbound call objects, which helps to understand the characteristics of the candidate outbound call objects more accurately, provide data support for outbound call decisions, and thus improve the effectiveness of outbound calls.

[0119] Before step S103 in some embodiments, the outbound call list generation method further includes: training a target conversion rate prediction model.

[0120] See also Figure 4 In some embodiments, the training process of the target conversion rate prediction model may include but is not limited to steps S401 to S406:

[0121] Step S401: Acquire sample object data, wherein the sample object data includes sample object information and sample object tags of multiple sample objects, and the sample object tags are used to indicate whether the sample objects are converted after the outbound call;

[0122] Step S402: extracting features from the sample object information to obtain sample object features;

[0123] Step S403, performing conversion rate prediction on the sample object features using a preset original conversion rate prediction model to obtain a sample predicted conversion rate for each sample object;

[0124] Step S404, determining a sample predicted conversion label according to the sample predicted conversion rate;

[0125] Step S405: Calculate the loss based on the sample object label and the sample prediction conversion label to obtain a sample prediction loss function;

[0126] Step S406 , optimizing the original conversion rate prediction model based on the sample prediction loss function to obtain a target conversion rate prediction model.

[0127] In steps S401 to S406 shown in the embodiment of the present application, sample object data including sample object information and sample object labels indicating whether the outbound call is converted is obtained, and the sample object data is used to train the original conversion rate prediction model. Next, feature extraction is performed on the sample object information to obtain the sample object features of each sample object, which helps the model focus on key information that can predict the conversion rate. The original conversion rate prediction model is used to predict the conversion rate of the sample object features to obtain the sample predicted conversion rate of each sample object, and the sample predicted conversion label is determined based on the sample predicted conversion rate, which provides a basis for the subsequent evaluation of the training effect of the model. By calculating the loss based on the sample object label and the sample predicted conversion label, a sample prediction loss function is obtained, which can quantify the deviation between the model prediction and the actual result. Finally, the original conversion rate prediction model is optimized based on the sample prediction loss function, which enables the model to continuously adjust parameters, reduce prediction errors, and obtain a target conversion rate prediction model, which can more accurately predict the conversion of the object after the outbound call and provide support for business decision-making.

[0128] In step S401 of some embodiments, the sample object data is pre-collected and desensitized data based on business scenarios, and is used to train the original conversion rate prediction model. The sample object data includes sample object information and sample object labels for multiple sample objects. The sample object labels are used to indicate whether the corresponding sample object converts in actual scenarios after an outbound call. Conversion indicates whether the sample object purchases the relevant product or subscribes to the relevant service.

[0129] In step S402 of some embodiments, for each sample object, feature extraction is performed on the sample object information of the sample object to obtain the sample object features of each sample object. Specifically, the implementation of feature extraction is basically the same as the specific implementation of steps S201 to S205 above, and will not be repeated here.

[0130] In step S403 of some embodiments, a conversion rate prediction is performed on the sample object features of each sample object using the original conversion rate prediction model to obtain a sample predicted conversion rate for each sample object. Specifically, the original conversion rate prediction model can be an XGBoost model, a LightGBM model, or a CatBoost model, etc., but is not limited thereto. The selection should be based on the actual application scenario.

[0131] It's understandable that since the sample predicted conversion rate is numerical data, it's necessary to determine the sample predicted conversion label based on the sample predicted conversion rate. For example, if the sample predicted conversion rate ranges from [0, 1], then a sample predicted conversion rate with a value in the range of [0, 0.5] can be set, and the corresponding sample predicted conversion label indicates that the sample object did not convert after the outbound call; a sample predicted conversion rate with a value in the range of (0.5, 1] can be set, and the corresponding sample predicted conversion label indicates that the sample object converted after the outbound call.

[0132] In step S405 of some embodiments, a corresponding loss function is selected according to the original conversion rate prediction model, and the loss is calculated for the sample object label and the sample predicted conversion label to obtain a sample prediction loss function; wherein, the loss function can select a mean square error loss function, a cross entropy loss function, etc., which is not limited in the embodiments of the present application.

[0133] In step S406 of some embodiments, the model parameters of the original conversion rate prediction model can be adjusted based on the sample prediction loss function using methods such as backpropagation, gradient descent, and momentum update, thereby optimizing the learning ability and generalization ability of the model and obtaining the target conversion rate prediction model.

[0134] In one embodiment, if the original conversion rate prediction model is an XGBoost model, a grid search method may be selected to perform model optimization on the original conversion rate prediction model.

[0135] In step S103 of some embodiments, the target conversion rate prediction model obtained through training is used to predict the target conversion rate of each candidate outbound call object to obtain the target predicted conversion rate of the candidate outbound call object.

[0136] It should be noted that, since the conversion label does not need to be considered in the subsequent outbound call list generation process, there is no need to convert the target predicted conversion rate of the candidate outbound call object into a predicted conversion label.

[0137] In step S104 of some embodiments, the preset agent resource pool is a pre-built collection of multiple candidate agent objects. This preset agent resource pool can be stored in a database, data warehouse, or other system and used to provide an agent list for outbound call services. For example, in an insurance application scenario, the candidate agent objects are insurance business personnel or agents who contact and interact with the candidate outbound call objects.

[0138] In some embodiments, the candidate agent object has an agent capability attribute, which is used to characterize the individual capability of the candidate agent object, including but not limited to the lower and upper limits of the number of outbound calls on the day, the outbound call conversion rate, etc.

[0139] See also Figure 5 In some embodiments, step S105 may also include but is not limited to steps S501 to S502:

[0140] Step S501: constructing an outbound call decision feature based on the candidate outbound call object and the candidate agent object; wherein the outbound call decision feature is used to indicate whether the candidate agent object makes an outbound call to the candidate outbound call object;

[0141] Step S502: construct a model based on the outbound call decision features and the target predicted conversion rate to obtain an outbound call list planning model.

[0142] Steps S501 and S502, as illustrated in the embodiment of this application, construct outbound call decision features based on candidate outbound call recipients and candidate agent recipients. This comprehensively considers key information from both parties involved in the outbound call process and accurately characterizes whether the candidate agent should place an outbound call to the candidate outbound call recipient. Subsequently, a model is constructed based on the outbound call decision features and the target predicted conversion rate to produce an outbound call list planning model. This model learns the potential relationship between outbound call decisions and conversion rates, thereby accurately planning the outbound call list, improving the relevance and efficiency of outbound calls, and ultimately enhancing the overall conversion effect of outbound calls.

[0143] In some embodiments, assuming that the candidate agent object is i and the candidate outbound call object is j, the outbound call decision feature constructed based on the candidate outbound call object and the candidate agent object can be recorded as x(i,j), where the outbound call decision feature x(i,j)=1 indicates that the candidate agent object i should make an outbound call to the candidate outbound call object j; the outbound call decision feature x(i,j)=0 indicates that the candidate agent object i should not make an outbound call to the candidate outbound call object j.

[0144] In step S502 of some embodiments, assuming that the target predicted conversion rate corresponding to the candidate outbound call object j is recorded as y(j), the objective function is defined as:

[0145] Objective function = ∑x(i,j)y(j)(1);

[0146] It can be understood that the outbound call decision features are constructed based on the candidate outbound call objects and candidate agent objects. The outbound call decision features are the variables that need to be solved. Based on the outbound call decision features and the target predicted conversion rate, a model is constructed to obtain the outbound call list planning model. The outbound call list planning model is to maximize the objective function, that is:

[0147] Outbound call list planning model = Max(objective function) = Max∑x(i,j)y(j)(2);

[0148] By planning and solving the outbound call list planning model, the optimal solution of the outbound call decision characteristics is obtained, and the outbound call list is generated based on the optimal solution of the outbound call decision characteristics.

[0149] See also Figure 6 In some embodiments, step S106 includes but is not limited to steps S601 to S605:

[0150] Step S601: Obtain the outbound call count constraint of the candidate agent.

[0151] Step S602: Obtain the object time constraint and answering number constraint of the candidate outbound call object;

[0152] Step S603: Obtain a preset agent outbound call time window, and generate an outbound call time constraint based on the agent outbound call time window;

[0153] Step S604: performing a linear programming solution on the outbound call list planning model based on the outbound call number constraint, the object time constraint, the answering number constraint, and the outbound call time constraint to obtain an original outbound call list of candidate agents;

[0154] Step S605: sort the original outbound call list to obtain an outbound call user list of candidate agents.

[0155] Steps S601 to S605 shown in the embodiment of the present application, by obtaining the outbound call count constraint of the candidate agent object, can prevent the candidate agent object from affecting the service quality due to excessive outbound calls, or affecting the business benefits due to too few outbound calls; obtaining the object time constraint and the answering number constraint of the candidate outbound call object can avoid making outbound calls when the candidate outbound call object is inconvenient, and frequently making outbound calls to the same candidate outbound call object, thereby improving user experience and answering rate. Obtaining the preset agent outbound call time window and generating the outbound call time constraint can reasonably arrange the agent's outbound call time period and improve work efficiency. Based on the outbound call count constraint, object time constraint, answering number constraint and outbound call time constraint, a linear programming solution is performed on the outbound call list planning model to obtain the original outbound call list, which is more in line with actual business rules and restrictions. Finally, the original outbound call list is sorted to obtain the final outbound call user list, which can make the outbound call list more accurate and reasonable, and help improve the efficiency of outbound calls.

[0156] In step S601 of some embodiments, an outbound call count constraint corresponding to each candidate agent is determined based on the agent capability attributes of each candidate agent. The outbound call count constraint represents the minimum and maximum daily outbound calls for the candidate agent. For example, the minimum daily outbound call count for candidate agent 1 is 50, and the maximum daily outbound call count is 70. The outbound call count constraint prevents the candidate agent from affecting service quality due to excessive outbound calls, or from affecting business profitability due to insufficient outbound calls, thereby improving the conversion rate of outbound calls.

[0157] In step S602 of some embodiments, the original object information may be analyzed to obtain the object time constraint and the number of answering constraints of the candidate outbound call object. For example, a candidate outbound call object can only make outbound calls between 10:00 and 12:00, and can only answer one outbound call per day.

[0158] If the original caller information cannot be analyzed to determine the caller time and answer count constraints, you can use the preset caller time and answer count constraints. Setting these constraints can help prevent outbound calls to a candidate caller when they are unavailable, as well as prevent frequent outbound calls to the same candidate caller, improving user experience and answer rate.

[0159] For example:

[0160] Object time constraint: The outbound calling period is all day, which corresponds to the agent's outbound calling time window;

[0161] Limit on the number of outbound calls: Each candidate outbound call recipient can answer a maximum of 2 outbound calls per day.

[0162] During the time period in which outbound calls are allowed, outbound calls can be made to the candidate outbound call recipients. During the time period outside the time period in which outbound calls are allowed, outbound calls cannot be made.

[0163] In step S603 of some embodiments, the agent outbound call time window is the working time period during which the candidate agent can perform outbound call tasks, such as 8:00-12:00 or 13:00-18:00. Furthermore, an outbound call time constraint is generated based on the agent outbound call time window to constrain the candidate agent to complete outbound calls to the candidate outbound call recipient within the agent outbound call time window, thereby ensuring that outbound call tasks are allocated within the agent outbound call time window.

[0164] It should be noted that the time period and granularity of the agent's outbound call time window need to be set according to the business scenario and business work arrangements, but are not limited to this.

[0165] In some embodiments, the agent outbound call time window can be divided into hourly units, for example: 8:00-9:00, 9:00-10:00, 10:00-11:00, 11:00-12:00, 13:00-14:00, 14:00-15:00, 15:00-16:00, 16:00-17:00, 17:00-18:00.

[0166] In step S604 of some embodiments, a preset linear programming solver is used to solve the outbound call list planning model based on the outbound call count constraint, the target time constraint, the number of answering calls constraint, and the outbound call time constraint. Specifically, the objective function is maximized to obtain an optimal solution for the outbound call decision feature x(i, j). This optimal solution is used to indicate whether candidate agent i should place an outbound call to candidate outbound caller j. Next, for each candidate agent i, candidate outbound caller j corresponding to the outbound call decision feature x(i, j) = 1 is selected to obtain the original outbound call list for each candidate agent i.

[0167] For example, for candidate agent 1, the corresponding outbound call decision features are:

[0168] x(1,1)=0, x(1,2)=1, x(1,3)=0,…, x(1,n)=1;

[0169] This means that candidate agent 1 should make outbound calls to candidate outbound caller 2, ..., candidate outbound caller n;

[0170] Then, for the candidate agent object 1, the candidate outbound call object j corresponding to the outbound call decision feature x(1,j)=1 is screened out to form an original outbound call list.

[0171] It should be noted that the linear programming solver can be CPLEX, Gurobi and other solvers, and the specific selection needs to be based on the actual application scenario, but is not limited to this.

[0172] In some embodiments, the candidate outbound call object j corresponding to x(i,j)=1 is defined as the candidate outbound call object, that is, the object to be called by the candidate agent object i. Therefore, the original outbound call list includes multiple candidate outbound call objects.

[0173] In step S605 of some embodiments, the original outbound call list of each candidate agent is sorted to obtain the outbound call user list of the candidate agent.

[0174] See also Figure 7 In some embodiments, step S605 may include but is not limited to steps S701 to S704:

[0175] Step S701, obtaining the object time constraint corresponding to the candidate outbound call object;

[0176] Step S702: Grouping candidate outbound call targets based on the agent outbound call time window and the target time constraint to obtain multiple initial outbound call lists; wherein each initial outbound call list corresponds to an agent outbound call time window;

[0177] Step S703: for each initial outbound call list, sort the initial outbound call list based on the target predicted conversion rate to obtain multiple candidate outbound call lists;

[0178] Step S704: Generate an outbound call user list of candidate seats based on the multiple candidate outbound call lists.

[0179] Steps S701 to S704 shown in the embodiment of the present application, by obtaining the object time constraints of the alternative outbound call objects, can accurately grasp the time range in which the alternative outbound call objects can accept outbound calls, avoid disturbing the alternative outbound call objects when they are inconvenient, improve customer experience, and at the same time improve the outbound call connection rate. Next, the alternative outbound call objects are grouped based on the agent outbound call time window and the object time constraint to obtain multiple initial outbound call lists. Then, for each initial outbound call list, it is sorted according to the target predicted conversion rate to obtain an alternative outbound call list, and the alternative outbound call objects with high potential conversion rates are contacted first to increase the business conversion volume. Finally, based on the combination of multiple alternative outbound call lists, an outbound user list is generated, which helps to improve the quality and efficiency of the overall outbound call business.

[0180] It can be understood that the alternative outbound call object is the candidate outbound call object j corresponding to x(i,j)=1. Therefore, the object time constraint corresponding to the alternative outbound call object is the object time constraint corresponding to the candidate outbound call object, and the target predicted conversion rate corresponding to the alternative outbound call object is the target predicted conversion rate of the candidate outbound call object.

[0181] In step S702 of some embodiments, candidate outbound callers are grouped based on the agent's outbound calling time window to obtain multiple initial outbound call lists, each of which corresponds to an agent's outbound calling time window. For candidate outbound callers with specific object time constraints, the candidate outbound caller is assigned to the initial outbound call list corresponding to the agent's outbound calling time window that matches the candidate.

[0182] As for the alternative outbound call objects without specific object time constraints, they are dynamically allocated to each initial outbound call list based on the principle of object quantity balance.

[0183] For example, if the agent's outbound calling time windows are 8:00-12:00 and 13:00-18:00, the initial outbound calling list will have two entries, one for the agent's outbound calling time window 8:00-12:00 and the other for the agent's outbound calling time window 13:00-18:00. Candidates with a time constraint of 8:00-12:00 will be assigned to the initial outbound calling list corresponding to the agent's outbound calling time window 8:00-12:00.

[0184] In step S703 of some embodiments, for each initial outbound call list, the candidates are sorted based on their target predicted conversion rates to obtain a candidate outbound call list. Specifically, the candidate list is sorted in descending order of target predicted conversion rates, with the candidate with the higher target predicted conversion rate being called first.

[0185] In step S704 of some embodiments, a plurality of candidate outbound call lists are sorted and combined according to the order of the agent outbound call time windows to generate an outbound call user list of candidate agent objects.

[0186] In some embodiments, after obtaining the outbound user list of the candidate agent object, the outbound user list is sent to each candidate agent object, so that each candidate agent object can make business outbound calls based on the outbound user list, without spending time and manpower to screen users from a preset user resource pool for outbound calls, thereby improving the outbound call efficiency of the candidate agent object.

[0187] The outbound call list generation method provided in the embodiments of this application can determine the outbound call list for each agent from a preset user resource pool, identify the most likely potential outbound call recipients, and prioritize their calls. This improves the agent's outbound calling efficiency and outbound call conversion rate. It also ensures the effective utilization of agent outbound call resources, avoids excessive outbound calls or waste of agent outbound call resources, and enables more accurate and efficient outbound calls.

[0188] See also Figure 8 The embodiment of the present application further provides an outbound call list generation device that can implement the above-mentioned outbound call list generation method, and the device includes:

[0189] The outbound call object acquisition module 801 is used to obtain candidate outbound call objects from a preset user resource pool and obtain original object information of the candidate outbound call objects;

[0190] The object feature extraction module 802 is used to extract features from the original object information to obtain candidate behavioral features of the candidate outbound call object;

[0191] The conversion rate prediction module 803 is used to predict the conversion rate of the candidate behavior characteristics using a preset target conversion rate prediction model to obtain the target predicted conversion rate of the candidate outbound call object;

[0192] The seat object acquisition module 804 is used to obtain candidate seat objects from a preset seat resource pool;

[0193] Model building module 805, for building a model based on candidate outbound call targets, candidate agent targets, and target predicted conversion rate to obtain an outbound call list planning model;

[0194] The list planning module 806 is used to plan the list using the outbound call list planning model to obtain an outbound call user list of candidate seats.

[0195] The specific implementation of the outbound call list generating device is basically the same as the specific embodiment of the above-mentioned outbound call list generating method, and will not be repeated here.

[0196] The present application also provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-mentioned outbound call list generation method. The electronic device can be any smart terminal including a tablet computer, an in-vehicle computer, or the like.

[0197] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0198] The processor 901 can be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0199] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902, and the processor 901 calls and executes the outbound call list generation method of the embodiment of the present application;

[0200] Input / output interface 903, used to implement information input and output;

[0201] Communication interface 904, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WiFi, Bluetooth, etc.);

[0202] Bus 905 , which transmits information between various components of the device (e.g., processor 901 , memory 902 , input / output interface 903 , and communication interface 904 );

[0203] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0204] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned outbound call list generation method is implemented.

[0205] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0206] The outbound call list generation method and device, electronic device and storage medium provided in the embodiments of the present application obtain candidate outbound call objects from a preset user resource pool, obtain candidate behavioral characteristics of the candidate outbound call objects, and then use the candidate behavioral characteristics to predict the conversion rate, which can accurately evaluate the potential conversion possibility of each candidate outbound call object, obtain the target predicted conversion rate of each candidate outbound call object, and provide a basis for subsequent list generation. Then, after obtaining the candidate agent object from the preset agent resource pool, a model is constructed based on the candidate outbound call object, the candidate agent object and the target predicted conversion rate to obtain an outbound call list planning model. Finally, the list is planned through the outbound call list planning model to obtain an outbound user list of the candidate agent object. The list is more reasonable and targeted, which is conducive to improving the efficiency of outbound calls, reducing invalid outbound calls, and at the same time improving the success rate of outbound calls, increasing business conversion volume, and thus improving overall business benefits.

[0207] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0208] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0209] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0210] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0211] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0212] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0213] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0214] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0215] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0216] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0217] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A method for generating an outbound call list, characterized in that: The method comprises: Obtaining a candidate outbound call object from a preset user resource pool, and obtaining original object information of the candidate outbound call object; Extracting features from the original object information to obtain candidate behavioral features of the candidate outbound call object; Perform conversion rate prediction on the candidate behavior characteristics using a preset target conversion rate prediction model to obtain a target predicted conversion rate of the candidate outbound call object; Get candidate agent objects from the preset agent resource pool; A model is constructed based on the candidate outbound call objects, the candidate agent objects, and the target predicted conversion rate to obtain an outbound call list planning model; The outbound call list planning model is used to plan the list to obtain the outbound call user list of the candidate agent object.

2. The method according to claim 1, characterized in that The model is constructed based on the candidate outbound call objects, the candidate agent objects, and the target predicted conversion rate to obtain an outbound call list planning model, including: An outbound call decision feature is constructed based on the candidate outbound call object and the candidate agent object; wherein the outbound call decision feature is used to indicate whether the candidate agent object makes an outbound call to the candidate outbound call object; A model is constructed based on the outbound call decision characteristics and the target predicted conversion rate to obtain the outbound call list planning model.

3. The method according to claim 2, characterized in that The outbound call list planning model is used to plan the outbound call list to obtain the outbound call user list of the candidate agent object, including: Obtain the outbound call count constraint of the candidate agent object; Obtaining the object time constraint and answering number constraint of the candidate outbound call object; Obtaining a preset agent outbound call time window, and generating an outbound call time constraint based on the agent outbound call time window; Performing linear programming on the outbound call list planning model based on the outbound call number constraint, the object time constraint, the answering number constraint, and the outbound call time constraint to obtain an original outbound call list of the candidate agent object; The original outbound call list is sorted to obtain the outbound call user list of the candidate agent object.

4. The method according to claim 3, characterized in that The original outbound call list includes candidate outbound call targets, and the candidate outbound call targets are used to indicate targets to be called by the candidate agent target. The process of sorting the original outbound call list to obtain the outbound call user list of the candidate agent target includes: Obtaining the object time constraint corresponding to the candidate outbound call object; The candidate outbound call objects are grouped based on the agent outbound call time window and the object time constraint to obtain multiple initial outbound call lists; wherein each of the initial outbound call lists corresponds to one of the agent outbound call time windows; For each of the initial outbound call lists, sorting the initial outbound call lists based on the target predicted conversion rate to obtain a plurality of candidate outbound call lists; The outbound call user list of the candidate agent object is generated based on a plurality of the candidate outbound call lists.

5. The method according to claim 1, characterized in that The extracting features of the original object information to obtain the candidate behavior features of the candidate outbound call object includes: For each candidate outbound call object, performing data cleaning on the original object information to obtain initial object information; Performing data encoding on the initial object information to obtain encoded object information; Standardizing the encoded object information to obtain standardized object information; Performing feature screening on the standardized object information to obtain original object features; Feature construction is performed on the original object features to obtain the candidate behavior features of the candidate outbound call object.

6. The method according to claim 5, characterized in that The step of performing data cleaning on the original object information to obtain initial object information includes: Screening the original object information for missing information to obtain missing object information; Performing abnormal information detection on the original object information to obtain abnormal object information; Information repair is performed on the missing object information and the abnormal object information in the original object information to obtain the initial object information.

7. The method according to claim 1, characterized in that Before performing conversion rate prediction on the candidate behavior features using a preset target conversion rate prediction model to obtain the target predicted conversion rate of the candidate outbound call recipient, the method further includes: training the target conversion rate prediction model; The target conversion rate prediction model is trained in the following way: Acquire sample object data, wherein the sample object data includes sample object information and sample object tags of a plurality of sample objects, wherein the sample object tags are used to indicate whether the sample objects are converted after the outbound call; Performing feature extraction on the sample object information to obtain sample object features; Performing conversion rate prediction on the sample object features using a preset original conversion rate prediction model to obtain a sample predicted conversion rate for each sample object; Determining a sample predicted conversion label according to the sample predicted conversion rate; Perform loss calculation based on the sample object label and the sample prediction conversion label to obtain a sample prediction loss function; The original conversion rate prediction model is optimized based on the sample prediction loss function to obtain the target conversion rate prediction model.

8. An outbound call list generating device, characterized in that: The device comprises: An outbound call object acquisition module is used to obtain candidate outbound call objects from a preset user resource pool and obtain original object information of the candidate outbound call objects; An object feature extraction module is used to extract features from the original object information to obtain candidate behavioral features of the candidate outbound call object; A conversion rate prediction module is used to predict the conversion rate of the candidate behavior characteristics using a preset target conversion rate prediction model to obtain a target predicted conversion rate of the candidate outbound call object; The seat object acquisition module is used to obtain candidate seat objects from the preset seat resource pool; A model building module, configured to build a model based on the candidate outbound call objects, the candidate agent objects, and the target predicted conversion rate to obtain an outbound call list planning model; The list planning module is used to plan the list through the outbound call list planning model to obtain the outbound call user list of the candidate seat object.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images