Talk skill recommendation method and device, computer equipment and storage medium

Abstract

The embodiment of the invention belongs to the field of artificial intelligence, and relates to a verbal skill recommendation method, which comprises the following steps: obtaining a transaction record of a historical transaction customer and a non-transaction record of a historical non-transaction customer, the transaction record and the non-transaction record comprising customer information and a communication record; performing feature extraction on the transaction record according to a preset feature processing mode to obtain a transaction key feature, and performing feature extraction on the non-transaction record according to the feature processing mode to obtain a non-transaction key feature; training the initial verbal skill recommendation model according to the transaction key features and the non-transaction key features to obtain a verbal skill recommendation model; and obtaining a current record of the current customer, and inputting the current record into the verbal skill recommendation model to obtain a recommended verbal skill which is used for performing product recommendation on the current customer. The invention further provides a verbal skill recommendation device, computer equipment and a storage medium. The accuracy of verbal skill recommendation is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, computer equipment and storage medium for recommending speech skills. Background Art

[0002] In the field of finance and insurance, as people's insurance awareness increases, customers' demand for various insurances is growing, which increases the complexity of communication between salesmen and customers. Various new insurance products are emerging in the market, and the sales strategies are different, which leads to many challenges for salesmen in communication, affecting communication efficiency and sales success rate, and reducing the accuracy of product recommendations. Existing communication training methods mostly rely on experience transfer, lack of systematic and data-driven support, making it difficult for salesmen to quickly master effective communication skills. In view of these shortcomings, there is an urgent need for a system that can provide salesmen with speech recommendations to improve communication effectiveness and customer satisfaction. Summary of the invention

[0003] The purpose of the embodiments of the present application is to propose a method, device, computer equipment and storage medium for recommending speech techniques to solve the problem of low accuracy of speech techniques recommendation.

[0004] In order to solve the above technical problems, the present application embodiment provides a method for recommending speech skills, which adopts the following technical solution:

[0005] Obtain transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, wherein the transaction records and non-transaction records include customer information and communication records;

[0006] Extracting features from the transaction records according to a preset feature processing method to obtain key transaction features, and extracting features from the non-transaction records according to the feature processing method to obtain key non-transaction features;

[0007] The initial speech recommendation model is trained according to the key transaction features and the non-transaction key features to obtain a speech recommendation model;

[0008] The current record of the current customer is obtained, and the current record is input into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

[0009] In order to solve the above technical problems, the embodiment of the present application further provides a speech recommendation device, which adopts the following technical solution:

[0010] A record acquisition module, used to acquire transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, wherein the transaction records and non-transaction records include customer information and communication records;

[0011] A feature extraction module is used to extract features from the transaction records according to a preset feature processing method to obtain key transaction features, and to extract features from the non-transaction records according to the feature processing method to obtain key non-transaction features;

[0012] A model training module, used to train the initial speech recommendation model according to the key transaction features and the non-transaction key features to obtain a speech recommendation model;

[0013] The speech recommendation module is used to obtain the current record of the current customer and input the current record into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

[0014] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0015] Obtain transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, wherein the transaction records and non-transaction records include customer information and communication records;

[0016] Extracting features from the transaction records according to a preset feature processing method to obtain key transaction features, and extracting features from the non-transaction records according to the feature processing method to obtain key non-transaction features;

[0017] The initial speech recommendation model is trained according to the key transaction features and the non-transaction key features to obtain a speech recommendation model;

[0018] The current record of the current customer is obtained, and the current record is input into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

[0019] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0020] Obtain transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, wherein the transaction records and non-transaction records include customer information and communication records;

[0021] Extracting features from the transaction records according to a preset feature processing method to obtain key transaction features, and extracting features from the non-transaction records according to the feature processing method to obtain key non-transaction features;

[0022] The initial speech recommendation model is trained according to the key transaction features and the non-transaction key features to obtain a speech recommendation model;

[0023] The current record of the current customer is obtained, and the current record is input into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

[0024] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects: by systematically acquiring and analyzing the communication records of historical transaction customers and historical non-transaction customers, feature extraction is performed according to a preset feature processing method to obtain key transaction features and non-transaction key features, thereby achieving effective feature extraction; subsequently, the initial sales talk recommendation model is trained according to the key transaction features and non-transaction key features, and the key factors affecting the transaction are identified from a large amount of data, so as to provide accurate sales talk recommendations for the salesperson based on the current records of the current customers; through a data-driven approach, the communication strategy with customers is optimized, the success rate of product recommendations is improved, and personalized recommendation sales talk can also improve customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0026] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0027] Figure 2 is a flow chart of an embodiment of a method for recommending speech skills according to the present application;

[0028] Figure 3 is a structural schematic diagram of an embodiment of a speech recommendation device according to the present application;

[0029] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0030] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0031] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0032] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0033] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0034] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0035] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0036] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0037] It should be noted that the speech recommendation method provided in the embodiment of the present application is generally executed by a server, and accordingly, the speech recommendation device is generally set in the server.

[0038] It should be understood that Figure 1The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0039] Continue to refer Figure 2 , shows a flow chart of an embodiment of a speech recommendation method according to the present application. The speech recommendation method comprises the following steps:

[0040] Step S201, obtaining transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, where the transaction records and non-transaction records include customer information and communication records.

[0041] In this embodiment, the electronic device (eg Figure 1 The server shown in the figure) can communicate with the terminal device through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0042] Specifically, the transaction records of historical transaction customers and the non-transaction records of historical non-transaction customers are collected and sorted out. Among them, transaction customers refer to customers who have completed the product purchase, and non-transaction customers refer to customers who have not completed the product purchase.

[0043] Both transaction records and non-transaction records contain the customer's customer information and the communication records between the customer and the salesperson. Customer information includes the customer's basic information (such as name, contact information), purchased product information, transaction time, amount, etc.

[0044] Step S202: extracting features from transaction records according to a preset feature processing method to obtain key transaction features, and extracting features from non-transaction records according to the feature processing method to obtain key non-transaction features.

[0045] Specifically, the server extracts features from transaction records and non-transaction records based on a preset feature processing method. Feature extraction aims to identify and quantify key features related to customer transactions, including transaction key features extracted from transaction records and non-transaction key features extracted from non-transaction records. Key features may include keywords in communication, speech structure, customer reactions, and communication strategies during transactions.

[0046] It can be understood that the feature processing methods for transaction records and non-transaction records in this step are the same, and are also intended to extract the same type of information.

[0047] Furthermore, the step of extracting features from transaction records in accordance with a preset feature processing method to obtain key transaction features may include: extracting keywords from transaction records to obtain keyword features; performing speech structure analysis on transaction records to obtain speech structure features; performing customer response analysis on transaction records to obtain customer response information; obtaining customer tags from customer information in transaction records; and determining the obtained keyword features, speech structure features, customer response information and customer tags as key transaction features.

[0048] Specifically, this application performs feature processing on transaction records from multiple dimensions to achieve feature extraction and identify key features that affect customer transactions, including the following aspects:

[0049] 1. Keyword extraction: Through natural language processing technology, keywords are extracted from transaction records. For example, TF-IDF (term frequency-inverse document frequency) is used to identify frequently appearing and representative words in communication records. Keywords can reflect the core content that customers are concerned about and reveal the theme of effective communication.

[0050] 2. Speech structure analysis: Analyze the transaction records to extract the key components of the call. This process helps to identify the commonly used sentence patterns and speech structures in communication, thereby providing structured support for speech recommendations.

[0051] 3. Customer response analysis: By analyzing the customer's response information during the communication process, we can obtain the customer's feedback on the dialogue, such as whether they have questions, expressed interest or clearly refused. Customer response information provides an important basis for subsequent strategy adjustments.

[0052] 4. Customer tag extraction: Extract customer tags from transaction records, such as customer type, purchase intention, demand characteristics, etc. Customer tags provide a basis for personalized recommendations, making subsequent communication strategies more targeted.

[0053] The keyword features, speech structure features, customer response information and customer tags are aggregated into key transaction features to provide rich feature support for subsequent speech recommendations and model training.

[0054] In this embodiment, the accuracy and effectiveness of the speech recommendation system are improved by performing multi-dimensional feature extraction on transaction records; keyword extraction ensures the relevance of communication content, speech structure analysis reveals successful communication patterns, and customer response analysis provides real-time feedback for optimizing communication strategies; comprehensive feature extraction subsequently enables the model to fully learn and identify the key factors for high transaction rates, thereby helping sales staff to develop more scientific and personalized communication strategies.

[0055] Furthermore, the above-mentioned step of extracting keywords from the transaction records to obtain keyword features may include: extracting keywords from the transaction records based on TF-IDF information to obtain multiple keywords; determining each keyword and the TF-IDF information as keyword features; or; extracting keywords from the transaction records based on word frequency information to obtain multiple keywords; determining each keyword and the word frequency information as keyword features.

[0056] Specifically, effective keyword features are extracted from transaction records in a data-driven manner, and these features will be used for subsequent training and optimization of the recommendation model. The specific process includes the following two methods:

[0057] 1. Keyword extraction based on TF-IDF information: TF-IDF (term frequency-inverse document frequency) is a commonly used text mining technology that aims to measure the importance of a word in a document collection. Term frequency (TF) reflects the frequency of a word in a specific document, while inverse document frequency (IDF) reflects the rarity of the word in the entire document collection; by combining the two, important keywords in a specific context can be identified.

[0058] The extracted keywords represent the core content that customers care about in the transaction records. Keywords and the TF-IDF information corresponding to the keywords are determined as keyword features.

[0059] 2. Keyword extraction based on word frequency information: Analyze the word frequency in the transaction records and identify frequently occurring words. This method is relatively intuitive and is used to quickly identify high-frequency words in a specific context. Combine the extracted keywords with their corresponding word frequency information to form keyword features.

[0060] Through the above two methods, we can fully identify the key factors that affect customer transactions and provide strong support for sales talk recommendations.

[0061] In this embodiment, keyword extraction is performed on transaction records, which effectively enhances the pertinence and accuracy of recommended sales talk. The dual strategy of TF-IDF and word frequency information ensures that the extracted keywords reflect the needs and concerns of customers, and can also reveal common expressions that appear in the process of successful transactions. Keyword features provide a solid foundation for the training of subsequent models, enabling the recommendation system to better understand customer preferences and behaviors, and improve the relevance and effectiveness of recommended sales talk.

[0062] Furthermore, the above-mentioned step of performing speech structure analysis on the transaction records to obtain speech structure features may include: performing text analysis on the transaction records to obtain multiple key components; extracting keywords and common sentence patterns of each key component; obtaining the frequency of occurrence and speech of each key component; and determining the keywords, common sentence patterns, frequency of occurrence and speech of each key component as speech structure features.

[0063] Specifically, effective speech structure features are extracted from transaction records to help identify successful communication patterns. The specific process includes the following steps:

[0064] 1. Text analysis: Perform text analysis on transaction records to identify key components. For example, in one embodiment, key components may include greetings (greetings at the beginning of a call), demand guidance (the part of asking customers about their needs), product introduction (introducing product features and advantages), objection handling (responding to customer objections), and transaction promotion (the part of guiding customers to make purchase decisions).

[0065] 2. Extraction of keywords and common sentence patterns: Extract keywords and common sentence patterns from the identified key components. These keywords and sentence patterns are the core elements of sales talk, which can reflect the focus of customers in communication and the effective expressions used by salespeople.

[0066] 3. Frequency of occurrence and acquisition of sales pitch: Count the frequency of occurrence of each key component. For example, the frequency of occurrence of the “demand guidance” component among successful customers is 80%, while that among unsuccessful customers is 50%.

[0067] Script acquisition refers to comparing the specific scripts used in each key component of closed and unclosed customers. For example, the "objection handling" section of closed customers usually contains more positive responses and solutions, while unclosed customers may lack effective objection handling. This helps to understand which expressions are more attractive to customers during the closing process.

[0068] The extracted key components and their corresponding keywords, common sentence patterns, frequency of occurrence and specific speech are integrated to form systematic speech structure features, which will be used to build and optimize the speech recommendation model.

[0069] Through the above steps, you can fully grasp the structure of the sales talk and provide important basis for subsequent recommendations.

[0070] In this embodiment, the transaction records are analyzed for the structure of the speech, which improves the understanding and recognition of successful communication patterns; the extracted keywords and common sentence patterns constitute the core elements of the speech, so that the recommendation system can accurately identify the effective expressions commonly used in the transaction process; by counting the frequency of occurrence of each component, it is determined which speech is more popular among customers, so that these contents are given priority when recommending; this process improves the relevance of the speech and enhances the system's analysis of customer psychology, so that sales staff can more accurately match customer needs and preferences when communicating with customers, thereby improving the success rate of product recommendations.

[0071] It can be understood that the present application handles transaction records and non-transaction records in the same way.

[0072] Step S203, training the initial speech recommendation model according to the key features of transaction and the key features of non-transaction to obtain the speech recommendation model.

[0073] Specifically, the initial sales pitch recommendation model is trained based on the extracted key features and non-sales pitch features. The initial sales pitch recommendation model uses machine learning algorithms, such as decision trees, random forests, support vector machines, etc. During the training process, the model will learn and identify which features are significantly correlated with customer sales, thereby continuously optimizing its sales pitch recommendation capabilities. It can be understood that the training is supervised training, with key sales pitch features as positive samples and non-sales pitch features as negative samples. After the training is completed, the model can predict new communication records and determine their likelihood of success.

[0074] Step S204, obtain the current record of the current customer, and input the current record into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

[0075] Specifically, after the model training is completed and the model application is entered, the communication records of the current customer will be obtained and input into the trained speech recommendation model. Before inputting into the model, feature extraction can be performed according to the preset feature processing method to obtain key features, and then the key features can be input into the model.

[0076] The recommendation model will generate recommendation words suitable for current customers based on their specific information. These words will help sales staff recommend products more effectively when communicating with customers and increase the probability of closing deals.

[0077] In this embodiment, by systematically acquiring and analyzing the communication records between historical transaction customers and historical non-transaction customers, feature extraction is performed according to a preset feature processing method to obtain key transaction features and non-transaction key features, thereby achieving effective feature extraction; subsequently, the initial sales talk recommendation model is trained based on the key transaction features and non-transaction key features, and the key factors affecting the transaction are identified from a large amount of data, so that accurate sales talk recommendations can be provided to the salesperson based on the current records of the current customers; through a data-driven approach, the communication strategy with customers is optimized, the success rate of product recommendations is improved, and personalized recommendation sales talk can also improve customer satisfaction.

[0078] Furthermore, after the step of obtaining the key features of the transaction, the method may further include: storing the key features of the transaction in a database.

[0079] Specifically, after the key features of a transaction are extracted, they are stored in the database, which ensures the persistence and traceability of the data and provides a basis for subsequent analysis and model training.

[0080] The key features of the transaction can be stored in a structured database. Such storage can be in the form of a relational database or a non-relational database, depending on the complexity of the data and the query requirements. In one embodiment, it can be stored in an ES database (full name Elasticsearch database, an open source, distributed, RESTful search engine built on Lucene).

[0081] In the database, key transaction features can be organized and indexed according to dimensions such as customer ID and communication record timestamp. Such structured management improves the efficiency of data retrieval and facilitates subsequent data analysis and model updates.

[0082] By storing key transaction features in the database, the existing data set can be continuously updated and expanded, providing richer training data for the machine learning model and improving the accuracy of the model.

[0083] In this embodiment, the key features of the transaction are stored in the database, which realizes efficient data management and long-term data accumulation; provides rich historical data support for subsequent analysis, enables the model to be trained on a larger data set, thereby improving prediction capabilities and recommendation accuracy; and can also realize query services for key features of the transaction.

[0084] Furthermore, before the above-mentioned step of obtaining the current record of the current customer, it can also include: receiving a customer communication request; obtaining a customer tag of the current customer according to the customer communication request; querying the key transaction features associated with the customer tag from a database according to the customer tag; and sending the key transaction features to the terminal logged in by the target account.

[0085] Specifically, before formal online communication with a customer, it is necessary to first receive and process a customer communication request initiated by the customer. According to the received customer communication request, the system extracts the tag information of the current customer.

[0086] Customer tags can be generated based on historical transactions, communication records, customer preferences, and other dimensions to reflect customer needs and characteristics. For example, a customer may be labeled as a "high-value customer" or a "potential customer."

[0087] Based on the customer tag, the database is searched for key transaction features associated with the tag. Through precise matching, specific words and strategies used in historical successful cases are obtained to provide support for sales staff.

[0088] Finally, the acquired key transaction features are sent to the terminal logged in by the target account. The terminal logged in by the target account can be the terminal logged in by the salesperson's work account, ensuring that the salesperson can access relevant information in time and optimize the recommendation strategy and speech when there is no communication record before the communication.

[0089] In this embodiment, personalized responses to customer communication requests are achieved, ensuring that the information obtained by the salesperson matches the customer's needs; customer tags are used for feature queries, making the recommended key features of transactions more targeted and improving the effectiveness of the sales talk.

[0090] Furthermore, the above-mentioned speech recommendation method may also include: obtaining a transaction evaluation value output by the speech recommendation model according to the current record; generating reminder information according to the transaction evaluation value; and sending the reminder information to the terminal logged in by the target account.

[0091] Specifically, in addition to providing recommended sales talk for current customers, the system can also perform transaction evaluation by inputting the current customer's current record into the sales talk recommendation model. The model generates a transaction evaluation value by analyzing historical data and current records. This value reflects the possibility of a customer closing a deal and can help sales staff evaluate the effectiveness of communication.

[0092] According to the transaction evaluation value, a threshold can be set. If the evaluation value is higher than a certain standard, a corresponding reminder message will be generated to prompt the salesperson to strengthen the communication strategy. For example, if the evaluation value is low, the salesperson can be advised to adjust the words or increase the frequency of interaction.

[0093] The generated reminder information will be sent to the terminal where the target account is logged in, that is, the terminal where the salesperson's work account is logged in, to ensure that the salesperson can refer to these real-time suggestions and optimize the recommendation strategy when conducting the next communication.

[0094] In this embodiment, by integrating interactive transaction evaluation, real-time feedback is provided and actionable suggestions are provided to salesmen; the generation of transaction evaluation values ​​enables salesmen to more accurately judge customers' purchasing intentions, reduces the possibility of blind sales promotion, and improves the pertinence and effectiveness of customer communication.

[0095] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned transaction records and non-transaction records, the above-mentioned transaction records and non-transaction records can also be stored in a node of a blockchain.

[0096] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0097] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, 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 obtain the best results.

[0098] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0099] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0100] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0101] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a speech recommendation device, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0102] like Figure 3 As shown, the speech recommendation device 300 described in this embodiment includes: a record acquisition module 301, a feature extraction module 302, a model training module 303 and a speech recommendation module 304, wherein:

[0103] The record acquisition module 301 is used to acquire the transaction records of historical transaction customers and the non-transaction records of historical non-transaction customers. The transaction records and non-transaction records include customer information and communication records.

[0104] The feature extraction module 302 is used to extract features from transaction records according to a preset feature processing method to obtain key transaction features, and to extract features from non-transaction records according to the feature processing method to obtain key non-transaction features.

[0105] The model training module 303 is used to train the initial speech recommendation model according to the key features of transaction and the key features of non-transaction to obtain the speech recommendation model.

[0106] The speech recommendation module 304 is used to obtain the current record of the current customer and input the current record into the speech recommendation model to obtain the recommended speech, which is used to recommend products to the current customer.

[0107] In this embodiment, by systematically acquiring and analyzing the communication records between historical transaction customers and historical non-transaction customers, feature extraction is performed according to a preset feature processing method to obtain key transaction features and non-transaction key features, thereby achieving effective feature extraction; subsequently, the initial sales talk recommendation model is trained based on the key transaction features and non-transaction key features, and the key factors affecting the transaction are identified from a large amount of data, so that accurate sales talk recommendations can be provided to the salesperson based on the current records of the current customers; through a data-driven approach, the communication strategy with customers is optimized, the success rate of product recommendations is improved, and personalized recommendation sales talk can also improve customer satisfaction.

[0108] In some optional implementations of this embodiment, the feature extraction module 302 may also be used to:

[0109] Extract keywords from transaction records to obtain keyword features; analyze the speech structure of transaction records to obtain speech structure features; analyze customer reactions to transaction records to obtain customer reaction information; obtain customer tags from customer information in transaction records; determine the obtained keyword features, speech structure features, customer reaction information and customer tags as key transaction features.

[0110] In some optional implementations of this embodiment, the feature extraction module 302 may also be used to:

[0111] Extract keywords from the transaction records based on TF-IDF information to obtain multiple keywords; determine each keyword and the TF-IDF information as keyword features; or; extract keywords from the transaction records based on word frequency information to obtain multiple keywords; determine each keyword and the word frequency information as keyword features.

[0112] In some optional implementations of this embodiment, the feature extraction module 302 may also be used to:

[0113] Conduct text analysis on transaction records to obtain multiple key components; extract keywords and common sentence patterns of each key component; obtain the frequency of occurrence and the wording of each key component; determine the keywords, common sentence patterns, frequency of occurrence and wording of each key component as the wording structure features.

[0114] In some optional implementations of this embodiment, the speech technique recommendation device may further include:

[0115] The storage module is used to store the key features of the transaction in the database.

[0116] In some optional implementations of this embodiment, the speech technique recommendation device may further include:

[0117] The request receiving module is used to receive customer communication requests.

[0118] The tag acquisition module is used to obtain the customer tag of the current customer based on the customer communication request.

[0119] The feature query module is used to query the key transaction features associated with the customer tags from the database based on the customer tags.

[0120] The feature sending module is used to send the key features of the transaction to the terminal logged in by the target account.

[0121] In some optional implementations of this embodiment, the speech technique recommendation device may further include:

[0122] The evaluation value acquisition module is used to obtain the transaction evaluation value output by the speech recommendation model based on the current record.

[0123] The reminder generation module is used to generate reminder information according to the transaction evaluation value.

[0124] The reminder sending module is used to send reminder information to the terminal logged in by the target account.

[0125] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0126] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 4 with a memory 41, a processor 42, and a network interface 43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0127] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0128] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of the speech recommendation method, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0129] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions stored in the memory 41 or process data, such as computer-readable instructions for running the speech recommendation method.

[0130] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0131] The computer device provided in this embodiment can execute the above-mentioned speech recommendation method. Here, the speech recommendation method can be the speech recommendation method of each of the above-mentioned embodiments.

[0132] In this embodiment, by systematically acquiring and analyzing the communication records between historical transaction customers and historical non-transaction customers, feature extraction is performed according to a preset feature processing method to obtain key transaction features and non-transaction key features, thereby achieving effective feature extraction; subsequently, the initial sales talk recommendation model is trained based on the key transaction features and non-transaction key features, and the key factors affecting the transaction are identified from a large amount of data, so that accurate sales talk recommendations can be provided to the salesperson based on the current records of the current customers; through a data-driven approach, the communication strategy with customers is optimized, the success rate of product recommendations is improved, and personalized recommendation sales talk can also improve customer satisfaction.

[0133] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the above-mentioned speech recommendation method.

[0134] In this embodiment, by systematically acquiring and analyzing the communication records between historical transaction customers and historical non-transaction customers, feature extraction is performed according to a preset feature processing method to obtain key transaction features and non-transaction key features, thereby achieving effective feature extraction; subsequently, the initial sales talk recommendation model is trained based on the key transaction features and non-transaction key features, and the key factors affecting the transaction are identified from a large amount of data, so that accurate sales talk recommendations can be provided to the salesperson based on the current records of the current customers; through a data-driven approach, the communication strategy with customers is optimized, the success rate of product recommendations is improved, and personalized recommendation sales talk can also improve customer satisfaction.

[0135] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0136] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for recommending speech skills, characterized in that: The steps include: Obtain transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, wherein the transaction records and non-transaction records include customer information and communication records; Extracting features from the transaction records according to a preset feature processing method to obtain key transaction features, and extracting features from the non-transaction records according to the feature processing method to obtain key non-transaction features; The initial speech recommendation model is trained according to the key transaction features and the non-transaction key features to obtain a speech recommendation model; The current record of the current customer is obtained, and the current record is input into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

2. The method for recommending words according to claim 1, characterized in that: The step of extracting features from the transaction record according to a preset feature processing method to obtain key transaction features comprises: Extracting keywords from the transaction records to obtain keyword features; Performing a speech structure analysis on the transaction record to obtain speech structure features; Performing customer response analysis on the transaction record to obtain customer response information; Obtaining a customer tag from the customer information of the transaction record; The acquired keyword features, speech structure features, customer response information and customer tags are identified as key transaction features.

3. The method for recommending words according to claim 2, characterized in that: The step of extracting keywords from the transaction records to obtain keyword features comprises: Extract keywords from the transaction record based on TF-IDF information to obtain multiple keywords; Determine each keyword and the TF-IDF information as a keyword feature; or; Extract keywords from the transaction record based on word frequency information to obtain multiple keywords; Each keyword and the word frequency information are determined as keyword features.

4. The method for recommending words according to claim 2, characterized in that: The step of analyzing the transaction record for the speech structure to obtain the speech structure features comprises: Performing text analysis on the transaction records to obtain multiple key components; Extract keywords and common sentence patterns of each key component; Obtain the occurrence frequency and words of each key component; The keywords, common sentence patterns, occurrence frequencies and speech techniques of each key component are determined as speech structure features.

5. The method for recommending words according to claim 1, characterized in that: After the step of obtaining the key features of the transaction, the method further includes: The key transaction features are stored in a database.

6. The method for recommending words according to claim 5, characterized in that: Before the step of obtaining the current record of the current customer, the method further includes: Receive customer communication requests; According to the customer communication request, obtain a customer tag of the current customer; According to the customer tag, querying the database for key transaction features associated with the customer tag; The key transaction features are sent to the terminal logged in by the target account.

7. The method for recommending speech techniques according to claim 1, characterized in that: The method further comprises: Obtaining a transaction evaluation value output by the speech recommendation model according to the current record; Generate reminder information according to the transaction evaluation value; The reminder information is sent to the terminal logged in by the target account.

8. A speech recommendation device, characterized in that: include: A record acquisition module, used to acquire transaction records of historical transaction customers and non-transaction records of historical non-transaction customers, wherein the transaction records and non-transaction records include customer information and communication records; A feature extraction module is used to extract features from the transaction records according to a preset feature processing method to obtain key transaction features, and to extract features from the non-transaction records according to the feature processing method to obtain key non-transaction features; A model training module, used to train the initial speech recommendation model according to the key transaction features and the non-transaction key features to obtain a speech recommendation model; The speech recommendation module is used to obtain the current record of the current customer and input the current record into the speech recommendation model to obtain the recommended speech, and the recommended speech is used to recommend products to the current customer.

9. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the speech recommendation method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the speech recommendation method as described in any one of claims 1 to 7.

Images