WO2023050669A1 - Neural network-based information pushing method and system, device, and medium - Google Patents

Abstract

The present application relates to the field of artificial intelligence, and provides a neural network-based information pushing method. The method comprises: obtaining first dialogue information of a target user, and second dialogue information replied by a target object according to the first dialogue information; obtaining first product information according to the first dialogue information, and determining, according to the first product information, whether second product information corresponding to the first product information exists in the second dialogue information; if the second product information exists in the second dialogue information, determining target product information according to the second product information, and returning target push information corresponding to the target product information; if the second product information does not exist in the second dialogue information, determining whether the second dialogue information is objection information; and if the second dialogue information is objection information, returning objection push information corresponding to the second dialogue information. According to the present application, = the efficiency and accuracy of information pushing are improved.

Description

基于神经网络的信息推送方法、***、设备及介质Information pushing method, system, device and medium based on neural network

本申请申明2021年09月30日递交的申请号为202111156853.0、名称为“基于神经网络的信息推送方法、***、设备及介质”的中国专利申请的优先权，该中国专利申请的整体内容以参考的方式结合在本申请中。This application declares the priority of the Chinese patent application with application number 202111156853.0 and titled "Neural Network-Based Information Push Method, System, Equipment and Medium" submitted on September 30, 2021. The entire content of the Chinese patent application is referred to way is incorporated in this application.

技术领域technical field

本申请实施例涉及人工智能领域，尤其涉及一种基于神经网络的信息推送方法、***、设备及介质。The embodiments of the present application relate to the field of artificial intelligence, and in particular to a neural network-based information push method, system, device, and medium.

背景技术Background technique

在线下或线上销售过程中，往往需要销售人员向顾客沟通介绍商品，但是一些销售人员往往会因为经验不足或对产品相关信息、产品相关的竞争产品的相关信息了解不够全面等问题，导致销售人员无法快速准确的应对客户提出的问题，从而导致客户满意度降低失去销售契机。发明人发现，目前，相关技术只能向推销者提供预先收集的话术以解决销售人员经验不足的问题，但是，无法解决销售人员对产品相关信息了解不够全面，无法向销售人员推送产品相关的竞争产品的相关信息等问题。因此，如何解决现有技术无法实现向销售人员推送竞品的相关信息，且信息推送准确率低、效果差的问题，成为了当前亟需解决的技术问题。In the process of offline or online sales, sales staff are often required to communicate and introduce products to customers, but some sales staff often have problems such as lack of experience or insufficient understanding of product-related information and product-related competitive products. Personnel cannot quickly and accurately respond to questions raised by customers, resulting in reduced customer satisfaction and lost sales opportunities. The inventor found that, at present, the related technology can only provide pre-collected words to the salesperson to solve the problem of insufficient experience of the salesperson, but it cannot solve the problem that the salesperson does not have a comprehensive understanding of product-related information and cannot push the competition related to the product to the salesperson Product related information and other issues. Therefore, how to solve the problem that the existing technology cannot push the relevant information of competing products to the salesperson, and the accuracy of information push is low and the effect is poor has become a technical problem that needs to be solved urgently.

发明内容Contents of the invention

有鉴于此，有必要提供一种基于神经网络的信息推送方法、***、设备及可读存储介质，以解决现有技术无法实现向销售人员推送竞品的相关信息，且信息推送准确率低、效果差的问题。In view of this, it is necessary to provide a neural network-based information push method, system, device, and readable storage medium to solve the problem that existing technologies cannot push information about competing products to salespersons, and the accuracy of information push is low, The problem of poor performance.

为实现上述目的，本申请实施例提供了一种基于神经网络的信息推送方法，所述方法步骤包括：In order to achieve the above purpose, an embodiment of the present application provides a method for pushing information based on a neural network, and the steps of the method include:

获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为 异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.

为实现上述目的，本申请实施例还提供了一种基于神经网络的信息推送***，包括：In order to achieve the above purpose, the embodiment of the present application also provides a neural network-based information push system, including:

信息获取模块，用于获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；An information acquisition module, configured to acquire the first dialogue information of the target user and the second dialogue information replied by the target object according to the first dialogue information;

第一判断模块，用于根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；A first judging module, configured to acquire first product information according to the first dialog information, and judge whether there is second product information corresponding to the first product information in the second dialog information according to the first product information ;

第一发送模块，用于如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；The first sending module is configured to determine target product information according to the second product information if the second product information exists in the second dialog information, and send target push information corresponding to the target product information to a user terminal associated with said target user;

第二判断模块，用于如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及A second judging module, configured to judge whether the second dialog information is objection information if the second product information does not exist in the second dialog information; and

第二发送模块，用于如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。The second sending module is configured to send objection push information corresponding to the second dialog information to the user terminal if the second dialog information is objection information.

为实现上述目的，本申请实施例还提供了一种计算机设备，所述计算机设备包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令，所述计算机可读指令被处理器执行时实现以下步骤：To achieve the above object, an embodiment of the present application further provides a computer device, the computer device includes a memory, a processor, and computer-readable instructions stored in the memory and operable on the processor, the The computer readable instructions implement the following steps when executed by the processor:

获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.

为实现上述目的，本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质内存储有计算机可读指令，所述计算机可读指令可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：In order to achieve the above purpose, the embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor, causing the at least one processor to perform the following steps:

获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.

本申请实施例提供的基于神经网络的信息推送方法、***、计算机设备及计算机可读存储介质，本实施例通过判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息，若存在第二产品信息则返回对应的推送信息，解决了现有技术无法实现向销售人员推送竞品的相关信息的问题，通过判断所述第二对话信息是否为异议信息，若存在异议信息则返回所述第二对话信息对应的异议推送信息，提高了信息推送的效率和准确性。In the neural network-based information push method, system, computer equipment, and computer-readable storage medium provided in the embodiments of the present application, this embodiment judges whether there is a second product corresponding to the first product information in the second dialogue information. information, if there is second product information, the corresponding push information will be returned, which solves the problem that the existing technology cannot push the relevant information of competing products to the salesperson. By judging whether the second dialogue information is objection information, if there is objection The information returns the objection push information corresponding to the second dialogue information, which improves the efficiency and accuracy of information push.

附图说明Description of drawings

图1为本申请实施例基于神经网络的信息推送方法的流程示意图；FIG. 1 is a schematic flow diagram of an information push method based on a neural network in an embodiment of the present application;

图2为本申请基于神经网络的信息推送***实施例二的程序模块示意图；Fig. 2 is the schematic diagram of the program module of the second embodiment of the neural network-based information push system of the present application;

图3为本申请计算机设备实施例三的硬件结构示意图。FIG. 3 is a schematic diagram of the hardware structure of Embodiment 3 of the computer equipment of the present application.

具体实施方式Detailed ways

为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, not to limit the present application. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.

需要说明的是，在本申请中涉及“第一”、“第二”等的描述仅用于描述目的，而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。另外，各个实施例之间的技术方案可以相互结合，但是必须是以本领域普通技术人员能够实现为基础，当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在，也不在本申请要求的保护范围之内。It should be noted that the descriptions in this application involving "first", "second" and so on are for descriptive purposes only, and should not be understood as indicating or implying their relative importance or implicitly indicating the number of indicated technical features . Thus, the features defined as "first" and "second" may explicitly or implicitly include at least one of these features. In addition, the technical solutions of the various embodiments can be combined with each other, but it must be based on the realization of those skilled in the art. When the combination of technical solutions is contradictory or cannot be realized, it should be considered that the combination of technical solutions does not exist , nor within the scope of protection required by the present application.

实施例一Embodiment one

参阅图1，示出了本申请实施例之基于神经网络的信息推送方法的步骤流程图。可以理解，本方法实施例中的流程图不用于对执行步骤的顺序进行限定。本实施例中的基于神经网络的信息推送***可以被执行在计算机设备2中，下面以计算机设备2为执行主体进 行示例性描述。具体如下。Referring to FIG. 1 , it shows a flow chart of the steps of the neural network-based information pushing method according to the embodiment of the present application. It can be understood that the flowchart in this method embodiment is not used to limit the sequence of execution steps. The neural network-based information push system in this embodiment can be implemented in the computer device 2, and the computer device 2 is used as the execution subject for an exemplary description below. details as follows.

步骤S100，获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息。Step S100, acquiring the first dialogue information of the target user and the second dialogue information replied by the target object according to the first dialogue information.

所述目标用户可以是产品销售过程中的销售人员，所述目标对象可以是所述销售人员的顾客；在线下或线上销售过程中，往往需要销售人员(目标用户)向顾客(目标对象)沟通介绍商品，但是一些销售人员往往会因为经验不足或对产品相关信息了解不够全面等问题，导致销售人员无法快速准确的应对客户提出的问题，从而导致客户满意度降低失去销售契机。鉴于此，本实施例提出一种基于神经网络的数据推送方法，可以应用于目标产品的销售场景，以帮助销售人员快速准确的回答客户提出的问题，所述目标产品可以是各类保险产品。The target user can be a salesperson in the product sales process, and the target object can be the customer of the salesperson; in the offline or online sales process, it is often necessary for the salesperson (target user) to send a message to the customer (target object) Communicate and introduce products, but some salespersons are often unable to quickly and accurately respond to questions raised by customers due to lack of experience or insufficient understanding of product-related information, resulting in reduced customer satisfaction and lost sales opportunities. In view of this, this embodiment proposes a neural network-based data push method, which can be applied to the sales scene of a target product to help salespersons quickly and accurately answer questions raised by customers. The target product can be various types of insurance products.

所述目标对象与所述目标用户的对话场景可以是面对面真人对话场景、电话语音对话场景、线上文字对话场景等。所述第一对话信息可以是所述销售人员的对话信息，即，销售人员说的话，如销售人员(目标用户)向顾客(目标对象)沟通介绍商品时说的话；所述第二对话信息为顾客说的话。The dialogue scene between the target object and the target user may be a face-to-face real dialogue scene, a telephone voice dialogue scene, an online text dialogue scene, and the like. The first dialogue information may be the dialogue information of the salesperson, that is, what the salesperson said, such as when the salesperson (target user) communicated and introduced the product to the customer (target object); the second dialogue information is What customers say.

步骤S102，根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息。Step S102, obtaining first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information.

需要说明的是，一般销售人员在对话刚开始时就会向顾客(目标对象)介绍商品，本实施例可以从所述销售人员(目标用户)向顾客(目标对象)介绍产品中获取所述第一产品信息，所述第一产品信息为所述销售人员(目标用户)向顾客(目标对象)介绍的产品信息。It should be noted that the general salesperson will introduce the product to the customer (target object) at the beginning of the conversation, and this embodiment can obtain the first Product information, the first product information is product information introduced by the salesperson (target user) to customers (target object).

当得到所述第一产品信息后，计算机设备2可以根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息。其中，所述第二产品信息可以是第一产品的竞品信息；例如，所述第一产品信息对应的产品为A保险，若所述第二对话信息为“我昨天看了B公司的X保险，觉得那个比较好”，那么，“X保险”及“A保险”的一个竞品，即“X保险”可以是所述第一产品信息对应的第二产品信息。After obtaining the first product information, the computer device 2 may determine whether there is second product information corresponding to the first product information in the second dialog information according to the first product information. Wherein, the second product information may be the competing product information of the first product; for example, the product corresponding to the first product information is A insurance, if the second dialogue information is "I saw the X Insurance, I think that is better", then, a competing product of "X Insurance" and "A Insurance", that is, "X Insurance" may be the second product information corresponding to the first product information.

在示例性的实施例中，步骤S102还可以包括步骤S200～步骤S202，其中：步骤S200，判断所述第一对话信息和所述第二对话信息是否为文本信息；步骤S202，如果所述第一对话信息和所述第二对话信息不是文本信息，则分别对所述第一对话信息和所述第二对话信息进行转写操作，以得到第一文本信息和第二文本信息；及步骤S204，根据所述第一文本信息获取第一产品信息，并根据所述第一产品信息判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息。本实施例通过判断所述第一对话信息和所述第二对话信息是否为文本信息，以使本实施例可以应用于各种对话场景，提高了应用范围。In an exemplary embodiment, step S102 may further include steps S200 to S202, wherein: step S200, judging whether the first dialogue information and the second dialogue information are text information; step S202, if the first dialogue information If the first dialogue information and the second dialogue information are not text information, then respectively perform transcribing operations on the first dialogue information and the second dialogue information to obtain the first text information and the second text information; and step S204 , acquiring first product information according to the first text information, and judging whether there is second product information corresponding to the first product information in the second text information according to the first product information. In this embodiment, by judging whether the first dialog information and the second dialog information are text information, this embodiment can be applied to various dialog scenarios, and the scope of application is improved.

需要说明的是，所述第一对话信息和所述第二对话信息可以文字信息，所述第一对话信息和所述第二对话信息也可以不是文字信息(如可以语音信息)。It should be noted that the first dialogue information and the second dialogue information may be text information, and the first dialogue information and the second dialogue information may not be text information (eg, voice information).

当得到所述第一对话信息和所述第二对话信息时，计算机设备2还可以检测所述第一对话信息和所述第二对话信息是否为文本信息，如果不是文本信息则需要先将所述第一对话信息和所述第二对话信息转换为文本信息。When obtaining the first dialog information and the second dialog information, the computer device 2 can also detect whether the first dialog information and the second dialog information are text information, if not text information, it needs to first Converting the first dialogue information and the second dialogue information into text information.

示例性的，当所述对话场景为面对面真人对话场景或电话语音对话场景等语音对话场景时，则获取到的所述第一对话信息和所述第二对话信息为语音信息，此时需要先将对话时的语音转化为文字。例如，计算机设备2可以通过转写的方式所述第一对话信息和所述第二对话信息转为所述第一文本信息和所述第二文本信息。Exemplarily, when the dialogue scene is a voice dialogue scene such as a face-to-face human dialogue scene or a telephone voice dialogue scene, the acquired first dialogue information and the second dialogue information are voice information. Convert speech to text during conversations. For example, the computer device 2 may convert the first dialog information and the second dialog information into the first text information and the second text information by transcribing.

在本实施例中，计算机设备2可以对所述语音形式的对话信息(所述第一对话信息或所述第二对话信息)进行转写操作，以得到多个语音文本；然后计算每个语音文本与所述语音形式的对话信息之间匹配度；并将所述多个语音文本中与所述语音形式的对话信息匹配度最高的语音文本作为所述文字形式的对话信息，以得到所述第一文本信息或所述第二文本信息。其中，计算每个语音文本w和所述语音形式的对话信息之间的匹配度，具体可以将与所述语音形式的对话信息匹配度最高的语音文本作为所述文字形式的对话信息

In this embodiment, the computer device 2 can transcribe the dialogue information in the voice form (the first dialogue information or the second dialogue information) to obtain a plurality of voice texts; then calculate each voice text The matching degree between the text and the dialogue information in the voice form; and the voice text with the highest matching degree with the dialogue information in the voice form among the plurality of voice texts is used as the dialogue information in the text form, so as to obtain the The first text information or the second text information. Wherein, calculate the matching degree between each speech text w and the dialog information in the speech form, specifically, the speech text with the highest matching degree with the dialog information in the speech form can be used as the dialog information in the text form

通过贝叶斯公式可以将p(w|x)转换为p(x|w)p(w)进行计算

其中，p(x|w)表示声学模型；p(w)表示语言模型；p(x)表示声学特征概率，对于不同的w该量是常值，在计算

时可以忽略。本实施例通过从所述多个语音文本中获取与所述语音形式的对话信息匹配度最高的文字形式的对话信息，提高了语音识别的转写正确率。 Through the Bayesian formula, p(w|x) can be converted into p(x|w)p(w) for calculation

Among them, p(x|w) represents the acoustic model; p(w) represents the language model; p(x) represents the probability of the acoustic feature, which is a constant value for different w.

can be ignored. In this embodiment, the dialogue information in text form that has the highest matching degree with the dialogue information in speech form is obtained from the plurality of voice texts, so as to improve the accuracy rate of transcription in speech recognition.

当所述对话场景为线上文字对话场景等文字对话场景时，则计算机设备2可以直接获取文字形式的对话信息(所述第一文本信息和所述第二文本信息)。When the dialogue scene is a text dialogue scene such as an online text dialogue scene, the computer device 2 may directly acquire dialogue information in text form (the first text information and the second text information).

在示例性的实施例中，步骤S204还可以包括步骤S300～步骤S310，其中：步骤S300，从所述第一文本信息中提取多个第一关键词；步骤S302，根据所述多个第一关键词从预设映射表中确定所述第一产品信息；步骤S304，获取所述第一产品信息关联的多个关联关键词，其中，每个关联词对应的一个第二产品信息；步骤S306，从所述第二文本信息中提取多个第二关键词；及步骤S308，分别计算每个第二关键词与各个关联关键词的关键词相似度，得到每个日志关键词对应的多个关键词相似度；步骤S310，根据各个第二关键词所对应的多个关键词相似度和预设关键词相似度判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息，其中，当所述多个关键词相似度中存在至少一个关键词相似度大于所述预设关键词相似度时，则所述第二文本信息中存在所述第一产品信息对应的第二产品信息。本实施例通过提取多个第一关键词、多个第二关键词以及多个关联关键词， 并根据所述多个第二关键词和是多个关联关键词判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息，提高了所述第二文本信息的识别准确性。需要说明的是，为了提高匹配范围本实施例还可以通过余弦相识度算法计算每个第二关键词与各个关联关键词的关键词相似度，并根据各个第二关键词所对应的多个关键词相似度和预设关键词相似度判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息。In an exemplary embodiment, step S204 may further include steps S300 to S310, wherein: step S300, extracting a plurality of first keywords from the first text information; step S302, extracting a plurality of first keywords according to the plurality of first keywords Keywords determine the first product information from the preset mapping table; step S304, obtain a plurality of associated keywords associated with the first product information, wherein each associated word corresponds to a second product information; step S306, Extract a plurality of second keywords from the second text information; and step S308, respectively calculate the keyword similarity between each second keyword and each associated keyword, and obtain a plurality of keywords corresponding to each log keyword word similarity; step S310, judging whether there is a second product corresponding to the first product information in the second text information according to a plurality of keyword similarities corresponding to each second keyword and a preset keyword similarity information, wherein, when there is at least one keyword similarity among the multiple keyword similarities that is greater than the preset keyword similarity, then there is the first product information corresponding to the first product information in the second text information. 2. Product information. In this embodiment, by extracting a plurality of first keywords, a plurality of second keywords, and a plurality of associated keywords, and judging whether the plurality of second keywords and the plurality of associated keywords are included in the second text information Whether there is second product information corresponding to the first product information improves the recognition accuracy of the second text information. It should be noted that, in order to improve the matching range, this embodiment can also use the cosine acquaintance algorithm to calculate the keyword similarity between each second keyword and each associated keyword, and according to the multiple keywords corresponding to each second keyword Word similarity and preset keyword similarity determine whether there is second product information corresponding to the first product information in the second text information.

在本实施例中，计算机设备2可以将所述第一文本信息输入到预先训练好的神经网络模型中，以得到所述神经网络模型输出的第一关键词；计算机设备2还可以将所述第二文本信息输入到所述神经网络模型中，以得到所述神经网络模型输出的第二关键词；其中，所述神经网络模型可以是双向长短期记忆神经网络，所述双向长短期记忆神经网络用于从所述第一文本信息和第二文本信息中提取主要信息，以得到所述多个第一关键词和所述多个第二关键词。In this embodiment, the computer device 2 can input the first text information into the pre-trained neural network model to obtain the first keyword output by the neural network model; the computer device 2 can also input the The second text information is input into the neural network model to obtain the second keyword output by the neural network model; wherein, the neural network model can be a two-way long-short-term memory neural network, and the two-way long-short-term memory neural network The network is used to extract main information from the first text information and the second text information, so as to obtain the plurality of first keywords and the plurality of second keywords.

所述预设映射表可以是预先根据多个产品的基础信息预先构建的信息映射表，例如，计算机设备2可以预先获取多个产品的基础信息，例如，属性信息，每个产品可以包括多个属性，计算机设备2可以于每个产品的多个属性提取该产品的多个标签，以根据多个标签建立该产品的信息映射表，当所述多个第一关键词和所述预设映射表其中一个产品的多个标签匹配时，则该产品的信息为第一文本信息对应的第一产品信息。The preset mapping table may be an information mapping table pre-built according to the basic information of multiple products, for example, the computer device 2 may pre-acquire the basic information of multiple products, such as attribute information, and each product may include multiple attribute, the computer device 2 can extract multiple tags of the product from multiple attributes of each product, so as to establish an information mapping table of the product according to the multiple tags, when the multiple first keywords and the preset mapping When multiple tags of a product in the table match, the product information is the first product information corresponding to the first text information.

所述预设映射表中预先为每个产品配置了多个竞品，并根据每个产品的多个竞品的基础信息为该产品提取了多个关联关键词。当得到所述第一产品信息时，计算机设备2可以根据所述第一产品信息从所述预设映射表中获取所述第一产品信息的多个关联关键词。Multiple competing products are pre-configured for each product in the preset mapping table, and multiple associated keywords are extracted for the product according to the basic information of the multiple competing products of each product. When the first product information is obtained, the computer device 2 may obtain multiple associated keywords of the first product information from the preset mapping table according to the first product information.

当得到所述多个关联关键词和多个第二关键词后，计算机设备2可以根据所述多个关联关键词判断所述预设映射表中是否存在对应的产品，如果存在对应的产品则根据所述多个第二关键词判断该产品是否为所述第二产品信息，即判断该产品是否为所述第一产品的竞品。After obtaining the plurality of associated keywords and the plurality of second keywords, the computer device 2 can judge whether there is a corresponding product in the preset mapping table according to the plurality of associated keywords, and if there is a corresponding product, then Judging whether the product is the second product information according to the plurality of second keywords, that is, judging whether the product is a competing product of the first product.

步骤S104，如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端。Step S104, if the second product information exists in the second dialogue information, determine target product information according to the second product information, and send target push information corresponding to the target product information to the The user terminal associated with the target user.

如果所述多个第二关键词对应的产品信息与所述多个关联关键词中任意一个匹配，则所述第二文本信息中存在所述第一产品信息对应的第二产品信息。If the product information corresponding to the multiple second keywords matches any one of the multiple associated keywords, then the second product information corresponding to the first product information exists in the second text information.

在示例性的实施例中，步骤S104还可以包括步骤S400～步骤S404，其中：步骤S400，根据预先配置的多个属性分类模型，获取所述第二产品信息对应的多个属性和各个属性对应的预测属性值；步骤S402，根据所述多个属性和所述多个预测属性值，提取所述第二产品信息的属性预测向量；及步骤S404，根据所述属性预测向量，从数据库中获取与所述属性预测向量相似度最高的目标向量，并根据所述目标向量确定所述目标产品信息。在销售过程中，销售人员(目标用户)向顾客(目标对象)沟通介绍商品时，销售人员(目标用 户)提到的产品(第一产品信息)往往是产品的专有名称，计算机设备2可以直接识别出对应的产品，但是，顾客(目标对象)提到的可能不是产品的专有名称，所以，本实施例在识别顾客(目标对象)提到的产品(第二产品信息)时，需要通过预先训练的多个属性分类模型，确定第二产品信息对应的所述目标产品信息，以提高所述第二产品信息识别准确性。In an exemplary embodiment, step S104 may further include steps S400 to S404, wherein: step S400, according to a plurality of pre-configured attribute classification models, obtain multiple attributes corresponding to the second product information and corresponding attributes of each attribute the predicted attribute value; step S402, according to the multiple attributes and the multiple predicted attribute values, extract the attribute prediction vector of the second product information; and step S404, according to the attribute prediction vector, obtain from the database A target vector with the highest similarity to the attribute prediction vector is determined, and the target product information is determined according to the target vector. In the sales process, when the salesperson (target user) communicates and introduces the product to the customer (target object), the product (first product information) mentioned by the salesperson (target user) is often the proper name of the product, and the computer device 2 can Directly identify the corresponding product, but what the customer (target object) mentions may not be the proper name of the product, so this embodiment needs to identify the product (second product information) mentioned by the customer (target object) The target product information corresponding to the second product information is determined through a plurality of pre-trained attribute classification models, so as to improve the recognition accuracy of the second product information.

需要说明的是，每个产品可以包括多个属性，如，产品：X保险的属性可以包括保险类型(如，重疾险)、销售地区(如，北京市和天津市)、年龄段(如，70岁)等。It should be noted that each product can include multiple attributes, for example, the attributes of product: X insurance can include insurance type (such as critical illness insurance), sales area (such as Beijing and Tianjin), age group (such as , 70 years old) and so on.

示例性的，所述多个属性分类模型包括保险类型分类模型、销售地区分类模型、年龄段分类模型等，其中，所述第二产品信息的每个属性的概率值(预测属性值)可以通过一个对应的属性分类模型获取。例如，将“x保险”输入到保险类型分类模型中，所述保险类型分类模型的softmax层所输出的输出值可以用于表示“x保险”在不同保险类型上的概率分布；计算机设备2可以选取softmax层所输出的输出值中的最大值最为作为“x保险”的保险类型判断结果；例如，若输出值中的最大值对应的类型是“重疾险”，则“x保险”的保险类型为“重疾险”，其中，“重疾险”对应的预测属性值为“重疾险”的概率。其中，各个属性分类模型的softmax层所输出的输出值的范围为0～1，各个属性分类模型可以通过多个标注数据进行训练。Exemplarily, the multiple attribute classification models include an insurance type classification model, a sales area classification model, an age group classification model, etc., wherein the probability value (predicted attribute value) of each attribute of the second product information can be obtained by A corresponding attribute classification model is obtained. For example, "x insurance" is input into the insurance type classification model, and the output value output by the softmax layer of the insurance type classification model can be used to represent the probability distribution of "x insurance" on different insurance types; the computer device 2 can Select the maximum value of the output value output by the softmax layer as the judgment result of the insurance type of "x insurance"; for example, if the type corresponding to the maximum value of the output value is "critical illness insurance", then the insurance of "x insurance" The type is "Critical Illness Insurance", where the predicted attribute value corresponding to "Critical Illness Insurance" is the probability of "Critical Illness Insurance". Wherein, the output value output by the softmax layer of each attribute classification model ranges from 0 to 1, and each attribute classification model can be trained through multiple labeled data.

当获取到所述第二产品信息对应的多个属性和各个属性对应的预测属性值后，计算机设备2可以根据所述多个属性和所述多个预测属性值，提取所述第二产品信息的属性预测向量，其中，所述属性预测向量可以用于从数据库中匹配与所述属性预测向量相似度最高的目标向量，并根据所述目标向量确定所述目标产品信息。其中，计算机设备2可以通过余弦相似度算法计算数据库中各个预先配置的向量与所述属性预测向量的相似度，并从各个相识度中选择与所述属性预测向量相似度最高的向量作为所述目标向量。After obtaining the multiple attributes corresponding to the second product information and the predicted attribute values corresponding to each attribute, the computer device 2 may extract the second product information according to the multiple attributes and the multiple predicted attribute values The attribute prediction vector, wherein the attribute prediction vector can be used to match the target vector with the highest similarity with the attribute prediction vector from the database, and determine the target product information according to the target vector. Wherein, the computer device 2 can calculate the similarity between each pre-configured vector in the database and the attribute prediction vector through the cosine similarity algorithm, and select the vector with the highest similarity with the attribute prediction vector from each degree of acquaintance as the target vector.

步骤S106，如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息。Step S106, if the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information.

如果所述第二对话信息中不存在对应的所述第二产品信息，即目标对象没有在对话中提及与所述第一产品相关的产品，本实施例还可以判断所述第二对话信息是否出自异议信息，所述异议信息可以对所述对话信息不认同、反驳等信息，其中，所述异议信息的识别可以通过异议意图识别模型识别，以确定所述第二对话信息是否为异议信息。If there is no corresponding second product information in the second dialogue information, that is, the target object does not mention a product related to the first product in the dialogue, this embodiment can also judge the second dialogue information Whether it comes from objection information, the objection information can be information such as disapproval or refutation of the dialogue information, wherein the identification of the objection information can be identified through an objection intention recognition model to determine whether the second dialogue information is objection information .

示例性的，为了提高异议信息的识别效率，本实施例还可以通过预先训练的异议分类模型对所述第二对话信息进行异议信息的识别。其中，所述异议分类模型可以是一个二分类的分类器，所述异议分类模型可以通过标注的异议和非异议数据作为训练样本进行训练。本实施例还可以通过MLP为分类器，识别出“我没空”为“异议”。Exemplarily, in order to improve the identification efficiency of objection information, in this embodiment, objection information may be identified on the second dialogue information by using a pre-trained objection classification model. Wherein, the objection classification model may be a binary classification classifier, and the objection classification model may be trained by using marked objection and non-objection data as training samples. In this embodiment, the MLP may also be used as a classifier to identify "I am not available" as "objection".

步骤S108，如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。Step S108, if the second dialog information is objection information, send objection push information corresponding to the second dialog information to the user terminal.

在示例性的实施例中，步骤S106还可以包括步骤S500～步骤S502，其中：步骤S500，提取所述异议信息对应的异议向量，并根据所述异议向量从所述数据库中获取与所述异议向量相识度最高的目标异议向量；及步骤S502，根据所述目标异议向量确定目标异议，并根据所述目标异议从数据库中获取所述异议推送信息。其中，获取与所述异议向量相识度最高的目标异议向量可以通过余弦相似度算法计算数据库中各个预先配置的向量与所述异议向量的相似度，并从各个相识度中选择与所述异议向量相似度最高的向量作为所述目标异议向量。本实施例通过提取所述异议信息对应的异议向量，从数据库中获取所述异议推送信息，提高了异议信息匹配效率和准确率。In an exemplary embodiment, step S106 may further include steps S500 to S502, wherein: step S500, extracting the objection vector corresponding to the objection information, and obtaining the objection vector corresponding to the objection information from the database according to the objection vector The target objection vector with the highest degree of vector acquaintance; and step S502, determining the target objection according to the target objection vector, and obtaining the push information of the objection from the database according to the target objection. Wherein, obtaining the target objection vector with the highest degree of acquaintance with the objection vector can calculate the similarity between each pre-configured vector in the database and the objection vector through the cosine similarity algorithm, and select the target objection vector from each degree of acquaintance with the objection vector The vector with the highest similarity is used as the target objection vector. In this embodiment, the objection push information is obtained from the database by extracting the objection vector corresponding to the objection information, which improves the matching efficiency and accuracy of objection information.

在示例性的实施例中，所述基于神经网络的信息推送方法还包括步骤S600～步骤S602，其中：步骤S600，如果所述第二对话信息中不存在异议信息，则将所述第二对话信息输入到预先训练好的环节分类模型，以通过所述环节分类模型输出所述第二对话信息的当前对话环节类型；及步骤S602，根据所述当前对话环节类型和所述第二对话信息从数据库中获取所述第二对话信息对应的对话推送信息，并将所述对话推送信息发送到所述用户终端。在本实施例中，如果所述第二对话信息中不存在所述异议信息，即目标对象没有对所述第一产品提出异议，这时计算机设备2可以根据所述当前对话环节从数据库中获取所述第二对话信息对应的对话推送信息，该对话推送信息可以是针对当前环节的通用话术等，本实施例通过获取所述第二对话信息的当前对话环节；并根据所述当前对话环节从数据库中获取所述第二对话信息对应的对话推送信息，提高了信息推送的准确性。本实施例可以将所述第二对话信息输入到预先训练好的环节分类模型，以通过所述环节分类模型输出所述第二对话信息的当前对话环节类型，其中，所述环节分类模型可以是通过大量的训练样本对一个softmax多分类的分类器训练得到，所述环节分类模型的训练样本可以是预先标注有环节标签的对话文本，例如，对话文本：“你好”可以属于“开场白”环节，即，对话文本：“你好”的环节标签可以是“开场白”(即，将对话文本：“你好”标注为“开场白”)。具体的，计算机设备2可以先通过GRU(Gated Recurrent Unit门控循环单元)神经网络对所述第二对话信息进行编码，并将所述第二对话信息的编码输入到所述预先训练好的环节分类模型，以通过所述预先训练好的环节分类模型输出所述第二对话信息的环节类型。In an exemplary embodiment, the neural network-based information push method further includes steps S600 to S602, wherein: step S600, if there is no objection information in the second dialogue information, send the second dialogue information to Information is input into a pre-trained link classification model, so as to output the current dialogue link type of the second dialogue information through the link classification model; and step S602, according to the current dialogue link type and the second dialogue information from The dialog push information corresponding to the second dialog information is obtained from the database, and the dialog push information is sent to the user terminal. In this embodiment, if the objection information does not exist in the second dialog information, that is, the target object has not raised an objection to the first product, then the computer device 2 can obtain the objection information from the database according to the current dialog link. Dialogue push information corresponding to the second dialogue information, the dialogue push information may be general speech for the current session, etc., this embodiment obtains the current dialogue session of the second dialogue information; and according to the current dialogue session The dialog pushing information corresponding to the second dialog information is acquired from the database, which improves the accuracy of information pushing. In this embodiment, the second dialogue information may be input into a pre-trained link classification model, so as to output the current dialogue link type of the second dialogue information through the link classification model, wherein the link classification model may be A softmax multi-classification classifier is trained through a large number of training samples. The training samples of the link classification model can be dialogue texts that are pre-marked with link labels. For example, the dialogue text: "Hello" can belong to the "opening remarks" link , that is, the dialogue text: "Hello" link label can be "opening remarks" (that is, the dialogue text: "Hello" is marked as "opening remarks"). Specifically, the computer device 2 may first encode the second dialogue information through a GRU (Gated Recurrent Unit) neural network, and input the encoding of the second dialogue information into the pre-trained link A classification model, for outputting the segment type of the second dialogue information through the pre-trained segment classification model.

在一些实施例中，如果所述对话场景为面对面真人对话场景或电话语音对话场景等语音对话场景，则将文字信息转换为语音返回到销售人员的用户终端关联的耳机中，或用户终端关联的显示屏幕上。若所述对话场景为线上文字对话场景等文字对话场景，则将文字信息返回到用户终端关联的显示屏幕上。本实施例还可以大幅提高保险销售人员的成功率，降低销售人员培训练习的时间，即使是初级人员，也可以通过该助手获取实时的高质量销售话术提示。In some embodiments, if the dialogue scene is a voice dialogue scene such as a face-to-face real person dialogue scene or a telephone voice dialogue scene, the text information is converted into voice and returned to the headset associated with the user terminal of the salesperson, or the user terminal associated with the user terminal. displayed on the screen. If the dialogue scene is a text dialogue scene such as an online text dialogue scene, the text information is returned to a display screen associated with the user terminal. This embodiment can also greatly increase the success rate of insurance salespersons and reduce the training time for salespersons. Even junior personnel can obtain real-time high-quality sales speech prompts through the assistant.

在示例性的实施例中，所述基于神经网络的信息推送方法还包括步骤S700：将所述目标推送信息或所述异议推送信息上传到区块链。In an exemplary embodiment, the neural network-based information push method further includes step S700: uploading the target push information or the objection push information to a block chain.

示例性的，将所述目标推送信息或所述异议推送信息上传至区块链可保证其安全性和公正透明性。本示例所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain)，本质上是一个去中心化的数据库，是一串使用密码学方法相关联产生的数据块，每一个数据块中包含了一批次网络交易的信息，用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。Exemplarily, uploading the target push information or the objection push information to the block chain can ensure its security, fairness and transparency. The blockchain referred to in this example is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain (Blockchain), essentially a decentralized database, is a series of data blocks associated with each other using cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

实施例二Embodiment two

图2为本申请基于神经网络的信息推送***实施例二的程序模块示意图。基于神经网络的信息推送***20可以包括或被分割成一个或多个程序模块，一个或者多个程序模块被存储于存储介质中，并由一个或多个处理器所执行，以完成本申请，并可实现上述基于神经网络的信息推送方法。本申请实施例所称的程序模块是指能够完成特定功能的一系列计算机程序指令段，比程序本身更适合于描述基于神经网络的信息推送***20在存储介质中的执行过程。以下描述将具体介绍本实施例各程序模块的功能：FIG. 2 is a schematic diagram of the program modules of Embodiment 2 of the neural network-based information push system of the present application. The neural network-based information push system 20 may include or be divided into one or more program modules, one or more program modules are stored in a storage medium, and are executed by one or more processors to complete the present application, And the above-mentioned information pushing method based on the neural network can be realized. The program module referred to in the embodiment of this application refers to a series of computer program instruction segments capable of completing specific functions, which is more suitable than the program itself to describe the execution process of the neural network-based information push system 20 in the storage medium. The following description will specifically introduce the functions of each program module of the present embodiment:

信息获取模块200，用于获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息。The information obtaining module 200 is configured to obtain the first dialogue information of the target user and the second dialogue information replied by the target object according to the first dialogue information.

第一判断模块202，用于根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息。The first judging module 202 is configured to acquire first product information according to the first dialog information, and judge whether there is a second product corresponding to the first product information in the second dialog information according to the first product information information.

第一发送模块204，用于如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端。The first sending module 204 is configured to determine the target product information according to the second product information if the second product information exists in the second dialog information, and push the target product information corresponding to the target product information sent to the user terminal associated with the target user.

第二判断模块206，用于如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息。The second judging module 206 is configured to judge whether the second dialog information is objection information if the second product information does not exist in the second dialog information.

第二发送模块208，用于如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。The second sending module 208 is configured to send objection push information corresponding to the second dialog information to the user terminal if the second dialog information is objection information.

示例性的，所述第一判断模块202，还用于：判断所述第一对话信息和所述第二对话信息是否为文本信息；如果所述第一对话信息和所述第二对话信息不是文本信息，则分别对所述第一对话信息和所述第二对话信息进行转写操作，以得到第一文本信息和第二文本信息；及根据所述第一文本信息获取第一产品信息，并根据所述第一产品信息判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息。Exemplarily, the first judging module 202 is further configured to: judge whether the first dialog information and the second dialog information are text information; if the first dialog information and the second dialog information are not text information, respectively performing a transcribing operation on the first dialogue information and the second dialogue information to obtain the first text information and the second text information; and obtaining the first product information according to the first text information, And judge whether there is second product information corresponding to the first product information in the second text information according to the first product information.

示例性的，所述第一判断模块202，还用于：从所述第一文本信息中提取多个第一关键词；根据所述多个第一关键词从预设映射表中确定所述第一产品信息；获取所述第一产品信息关联的多个关联关键词，其中，每个关联词对应的一个第二产品信息；从所述第二文本信息中提取多个第二关键词；分别计算每个第二关键词与各个关联关键词的关键词相 似度，得到每个日志关键词对应的多个关键词相似度；及根据各个第二关键词所对应的多个关键词相似度和预设关键词相似度判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息，其中，当所述多个关键词相似度中存在至少一个关键词相似度大于所述预设关键词相似度时，则所述第二文本信息中存在所述第一产品信息对应的第二产品。Exemplarily, the first judgment module 202 is further configured to: extract a plurality of first keywords from the first text information; determine the plurality of first keywords from a preset mapping table according to the plurality of first keywords. first product information; obtain a plurality of associated keywords associated with the first product information, wherein each associated word corresponds to a second product information; extract a plurality of second keywords from the second text information; respectively Calculate the keyword similarity of each second keyword and each associated keyword to obtain a plurality of keyword similarities corresponding to each log keyword; and according to a plurality of keyword similarities corresponding to each second keyword and The preset keyword similarity determines whether there is second product information corresponding to the first product information in the second text information, wherein, when at least one keyword similarity among the plurality of keyword similarities is greater than the When the similarity degree of the preset keyword is higher, the second product corresponding to the first product information exists in the second text information.

示例性的，所述第一发送模块204，还用于：根据预先配置的多个属性分类模型，获取所述第二产品信息对应的多个属性和各个属性对应的预测属性值；根据所述多个属性和所述多个预测属性值，提取所述第二产品信息的属性预测向量；及根据所述属性预测向量，从数据库中获取与所述属性预测向量相似度最高的目标向量，并根据所述目标向量确定所述目标产品信息。Exemplarily, the first sending module 204 is further configured to: acquire multiple attributes corresponding to the second product information and predicted attribute values corresponding to each attribute according to multiple pre-configured attribute classification models; a plurality of attributes and the plurality of predicted attribute values, extracting an attribute prediction vector of the second product information; and according to the attribute prediction vector, obtaining a target vector with the highest similarity with the attribute prediction vector from a database, and The target product information is determined according to the target vector.

示例性的，所述第二发送模块208，还用于：提取所述异议信息对应的异议向量，并根据所述异议向量从所述数据库中获取与所述异议向量相识度最高的目标异议向量；及根据所述目标异议向量确定目标异议，并根据所述目标异议从数据库中获取所述异议推送信息。Exemplarily, the second sending module 208 is further configured to: extract an objection vector corresponding to the objection information, and obtain a target objection vector with the highest degree of acquaintance with the objection vector from the database according to the objection vector ; and determining the target objection according to the target objection vector, and obtaining the push information of the objection from the database according to the target objection.

示例性的，所述基于神经网络的信息推送***20还包括，第三发送模块，所述第三发送模块，用于如果所述第二对话信息中不存在异议信息，则将所述第二对话信息输入到预先训练好的环节分类模型，以通过所述环节分类模型输出所述第二对话信息的当前对话环节类型；及根据所述当前对话环节类型和所述第二对话信息从数据库中获取所述第二对话信息对应的对话推送信息，并将所述对话推送信息发送到所述用户终端。Exemplarily, the neural network-based information push system 20 also includes a third sending module, and the third sending module is configured to send the second dialog information if there is no objection information in the second dialog information The dialogue information is input into the pre-trained link classification model, so as to output the current dialogue link type of the second dialogue information through the link classification model; and according to the current dialogue link type and the second dialogue information from the database Obtain dialog push information corresponding to the second dialog information, and send the dialog push information to the user terminal.

示例性的，所述基于神经网络的信息推送***20还包括，上传模块，所述上传模块，用于将所述目标推送信息或所述异议推送信息上传到区块链。Exemplarily, the neural network-based information push system 20 further includes an upload module, configured to upload the target push information or the objection push information to a block chain.

实施例三Embodiment three

参阅图3，是本申请实施例三之计算机设备的硬件架构示意图。本实施例中，计算机设备2是一种能够按照事先设定或者存储的指令，自动进行数值计算和/或信息处理的设备。该计算机设备2可以是机架式服务器、刀片式服务器、塔式服务器或机柜式服务器(包括独立的服务器，或者多个服务器所组成的服务器集群)等。如图所示，计算机设备2至少包括，但不限于，可通过***总线相互通信连接存储器21、处理器22、网络接口23、以及基于神经网络的信息推送***20。Referring to FIG. 3 , it is a schematic diagram of a hardware architecture of a computer device according to Embodiment 3 of the present application. In this embodiment, the computer device 2 is a device capable of automatically performing numerical calculation and/or information processing according to preset or stored instructions. The computer device 2 may be a rack server, a blade server, a tower server or a cabinet server (including an independent server, or a server cluster composed of multiple servers) and the like. As shown in the figure, the computer device 2 at least includes, but is not limited to, a memory 21 , a processor 22 , a network interface 23 , and an information push system 20 based on a neural network that can communicate with each other through a system bus.

本实施例中，存储器21至少包括一种类型的计算机可读存储介质，所述可读存储介质包括闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘等。在一些实施例中，存储器21可以是计算机设备2的内部存储单元，例如该计算机设备2的硬盘或内存。在另一些实施例中，存储器21也可以是计算机设备2的外部存储设备，例如该计算机 设备2上配备的插接式硬盘，智能存储卡(Smart Media Card，SMC)，安全数字(Secure Digital，SD)卡，闪存卡(Flash Card)等。当然，存储器21还可以既包括计算机设备2的内部存储单元也包括其外部存储设备。本实施例中，存储器21通常用于存储安装于计算机设备2的操作***和各类应用软件，例如实施例二的基于神经网络的信息推送***20的程序代码等。此外，存储器21还可以用于暂时地存储已经输出或者将要输出的各类数据。In this embodiment, the memory 21 includes at least one type of computer-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, magnetic disk, optical disk, etc. In some embodiments, the memory 21 may be an internal storage unit of the computer device 2 , such as a hard disk or memory of the computer device 2 . In other embodiments, the memory 21 can also be an external storage device of the computer device 2, such as a plug-in hard disk equipped on the computer device 2, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, flash memory card (Flash Card), etc. Of course, the storage 21 may also include both the internal storage unit of the computer device 2 and its external storage device. In this embodiment, the memory 21 is usually used to store the operating system and various application software installed in the computer device 2, such as the program code of the neural network-based information push system 20 in the second embodiment. In addition, the memory 21 can also be used to temporarily store various types of data that have been output or will be output.

处理器22在一些实施例中可以是中央处理器(Central Processing Unit，CPU)、控制器、微控制器、微处理器、或其他数据处理芯片。该处理器22通常用于控制计算机设备2的总体操作。本实施例中，处理器22用于运行存储器21中存储的程序代码或者处理数据，例如运行基于神经网络的信息推送***20，以实现实施例一的基于神经网络的信息推送方法。In some embodiments, the processor 22 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 22 is generally used to control the overall operation of the computer device 2 . In this embodiment, the processor 22 is used to run the program code stored in the memory 21 or process data, for example, run the neural network-based information push system 20, so as to implement the neural network-based information push method in Embodiment 1.

所述网络接口23可包括无线网络接口或有线网络接口，该网络接口23通常用于在计算机设备2与其他电子装置之间建立通信连接。例如，所述网络接口23用于通过网络将计算机设备2与外部终端相连，在计算机设备2与外部终端之间的建立数据传输通道和通信连接等。所述网络可以是企业内部网(Intranet)、互联网(Internet)、全球移动通讯***(Global System of Mobile communicatI/On，GSM)、宽带码分多址(Wideband Code DivisI/On Multiple Access，WCDMA)、4G网络、5G网络、蓝牙(Bluetooth)、Wi-Fi等无线或有线网络。The network interface 23 may include a wireless network interface or a wired network interface, and the network interface 23 is generally used to establish a communication connection between the computer device 2 and other electronic devices. For example, the network interface 23 is used to connect the computer device 2 with an external terminal through a network, and establish a data transmission channel and a communication connection between the computer device 2 and the external terminal. Described network can be intranet (Intranet), Internet (Internet), Global System for Mobile Communications (Global System of Mobile communicatI/On, GSM), Wideband Code Division Multiple Access (Wideband Code DivisI/On Multiple Access, WCDMA), 4G network, 5G network, Bluetooth (Bluetooth), Wi-Fi and other wireless or wired networks.

需要指出的是，图3仅示出了具有部件20-23的计算机设备2，但是应理解的是，并不要求实施所有示出的部件，可以替代的实施更多或者更少的部件。It should be noted that FIG. 3 only shows computer device 2 having components 20-23, but it should be understood that implementation of all of the illustrated components is not required and that more or fewer components may instead be implemented.

在本实施例中，存储于存储器21中的基于神经网络的信息推送***20还可以被分割为一个或者多个程序模块，所述一个或者多个程序模块被存储于存储器21中，并由一个或多个处理器(本实施例为处理器22)所执行，以完成本申请。In this embodiment, the neural network-based information push system 20 stored in the memory 21 can also be divided into one or more program modules, and the one or more program modules are stored in the memory 21 and controlled by a or multiple processors (the processor 22 in this embodiment) to complete the application.

例如，图2示出了本申请实施例二之所述实现基于神经网络的信息推送***20的程序模块示意图，该实施例中，所述基于神经网络的信息推送***20可以被划分为信息获取模块200、第一判断模块202、第一发送模块204、第二判断模块206和第二发送模块208。其中，本申请所称的程序模块是指能够完成特定功能的一系列计算机程序指令段，比程序更适合于描述所述基于神经网络的信息推送***20在计算机设备2中的执行过程。所述程序模块200-208的具体功能在实施例二中已有详细描述，在此不再赘述。For example, FIG. 2 shows a schematic diagram of program modules for realizing the neural network-based information push system 20 in Embodiment 2 of the present application. In this embodiment, the neural network-based information push system 20 can be divided into information acquisition Module 200 , first judging module 202 , first sending module 204 , second judging module 206 and second sending module 208 . Wherein, the program module referred to in this application refers to a series of computer program instruction segments capable of accomplishing specific functions, which are more suitable than programs for describing the execution process of the neural network-based information push system 20 in the computer device 2 . The specific functions of the program modules 200-208 have been described in detail in the second embodiment, and will not be repeated here.

实施例四Embodiment four

本实施例还提供一种计算机可读存储介质，所述计算机可读存储介质可以是非易失性，也可以是易失性，如闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘、服务器、App应用商城等等，其上存储有计算机可读指令，计算机可读指令执行时实现相应功能。本实施例的计算机可读存储介质基于神经网络的信息推送***20，被处理器 执行如下步骤：This embodiment also provides a computer-readable storage medium, which can be non-volatile or volatile, such as 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, CD, server, App application store, etc., on which computer-readable instructions are stored, and the corresponding functions are realized when the computer-readable instructions are executed. The computer-readable storage medium of the present embodiment is based on the information push system 20 of the neural network, and the processor performs the following steps:

获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.

上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。The serial numbers of the above embodiments of the present application are for description only, and do not represent the advantages and disadvantages of the embodiments.

通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is better implementation.

以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structures or equivalent process transformations made by using the description of the application and the accompanying drawings are directly or indirectly used in other related technical fields. , are all included in the patent protection scope of the present application in the same way.

Claims (20)

1. 一种基于神经网络的信息推送方法，其中，所述方法包括：A neural network-based information push method, wherein the method includes:

   获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

   根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

   如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

   如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

   如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.
2. 如权利要求1所述的基于神经网络的信息推送方法，其中，所述根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息的步骤，包括：The method for pushing information based on neural network according to claim 1, wherein said first product information is acquired according to said first dialogue information, and it is judged according to said first product information whether there is The step of the second product information corresponding to the first product information includes:

   判断所述第一对话信息和所述第二对话信息是否为文本信息；judging whether the first dialogue information and the second dialogue information are text information;

   如果所述第一对话信息和所述第二对话信息不是文本信息，则分别对所述第一对话信息和所述第二对话信息进行转写操作，以得到第一文本信息和第二文本信息；及If the first dialogue information and the second dialogue information are not text information, performing a transcribing operation on the first dialogue information and the second dialogue information respectively to obtain first text information and second text information ;and

   根据所述第一文本信息获取第一产品信息，并根据所述第一产品信息判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息。Acquiring first product information according to the first text information, and judging whether there is second product information corresponding to the first product information in the second text information according to the first product information.
3. 如权利要求2所述的基于神经网络的信息推送方法，其中，所述根据所述第一文本信息获取第一产品信息，并根据所述第一产品信息判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息的步骤，包括：The neural network-based information push method according to claim 2, wherein said first product information is acquired according to said first text information, and it is judged according to said first product information whether there is The step of the second product information corresponding to the first product information includes:

   从所述第一文本信息中提取多个第一关键词；extracting a plurality of first keywords from the first text information;

   根据所述多个第一关键词从预设映射表中确定所述第一产品信息；determining the first product information from a preset mapping table according to the plurality of first keywords;

   获取所述第一产品信息关联的多个关联关键词，其中，每个关联词对应的一个第二产品信息；Obtaining a plurality of associated keywords associated with the first product information, wherein each associated word corresponds to a piece of second product information;

   从所述第二文本信息中提取多个第二关键词；extracting a plurality of second keywords from the second text information;

   分别计算每个第二关键词与各个关联关键词的关键词相似度，得到每个日志关键词对应的多个关键词相似度；及Calculating the keyword similarity between each second keyword and each associated keyword respectively to obtain a plurality of keyword similarities corresponding to each log keyword; and

   根据各个第二关键词所对应的多个关键词相似度和预设关键词相似度判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息，其中，当所述多个关键词相似度中存在至少一个关键词相似度大于所述预设关键词相似度时，则所述第二文本信息中存 在所述第一产品信息对应的第二产品信息。Judging whether there is second product information corresponding to the first product information in the second text information according to the multiple keyword similarities corresponding to each second keyword and the preset keyword similarity, wherein, when the When at least one keyword similarity among the plurality of keyword similarities is greater than the preset keyword similarity, the second product information corresponding to the first product information exists in the second text information.
4. 如权利要求3所述的基于神经网络的信息推送方法，其中，所述如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端的步骤，包括：The neural network-based information push method according to claim 3, wherein, if the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send The step of sending the target push information corresponding to the target product information to the user terminal associated with the target user includes:

   根据预先配置的多个属性分类模型，获取所述第二产品信息对应的多个属性和各个属性对应的预测属性值；Acquiring multiple attributes corresponding to the second product information and predicted attribute values corresponding to each attribute according to a plurality of pre-configured attribute classification models;

   根据所述多个属性和所述多个预测属性值，提取所述第二产品信息的属性预测向量；及extracting an attribute prediction vector of the second product information according to the plurality of attributes and the plurality of predicted attribute values; and

   根据所述属性预测向量，从数据库中获取与所述属性预测向量相似度最高的目标向量，并根据所述目标向量确定所述目标产品信息。According to the attribute prediction vector, the target vector with the highest similarity to the attribute prediction vector is obtained from the database, and the target product information is determined according to the target vector.
5. 如权利要求1所述的基于神经网络的信息推送方法，其中，如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端的步骤，包括：The neural network-based information push method according to claim 1, wherein, if the second dialog information is objection information, the step of sending objection push information corresponding to the second dialog information to the user terminal, include:

   提取所述异议信息对应的异议向量，并根据所述异议向量从所述数据库中获取与所述异议向量相识度最高的目标异议向量；及extracting the objection vector corresponding to the objection information, and obtaining a target objection vector with the highest degree of acquaintance with the objection vector from the database according to the objection vector; and

   根据所述目标异议向量确定目标异议，并根据所述目标异议从数据库中获取所述异议推送信息。The target objection is determined according to the target objection vector, and the push information of the objection is acquired from a database according to the target objection.
6. 如权利要求1所述的基于神经网络的信息推送方法，其中，还包括：The method for pushing information based on neural network as claimed in claim 1, further comprising:

   如果所述第二对话信息中不存在异议信息，则将所述第二对话信息输入到预先训练好的环节分类模型，以通过所述环节分类模型输出所述第二对话信息的当前对话环节类型；及If there is no objection information in the second dialogue information, then input the second dialogue information into a pre-trained link classification model, so as to output the current dialogue link type of the second dialogue information through the link classification model ;and

   根据所述当前对话环节类型和所述第二对话信息从数据库中获取所述第二对话信息对应的对话推送信息，并将所述对话推送信息发送到所述用户终端。Obtain dialog push information corresponding to the second dialog information from a database according to the current dialog link type and the second dialog information, and send the dialog push information to the user terminal.
7. 如权利要求1至6中任一项所述的基于神经网络的信息推送方法，其中，还包括：将所述目标推送信息或所述异议推送信息上传到区块链。The neural network-based information push method according to any one of claims 1 to 6, further comprising: uploading the target push information or the objection push information to a block chain.
8. 一种基于神经网络的信息推送***，其中，包括：A neural network-based information push system, including:

   信息获取模块，用于获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；An information acquisition module, configured to acquire the first dialogue information of the target user and the second dialogue information replied by the target object according to the first dialogue information;

   第一判断模块，用于根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；A first judging module, configured to acquire first product information according to the first dialog information, and judge whether there is second product information corresponding to the first product information in the second dialog information according to the first product information ;

   第一发送模块，用于如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；The first sending module is configured to determine target product information according to the second product information if the second product information exists in the second dialog information, and send target push information corresponding to the target product information to a user terminal associated with said target user;

   第二判断模块，用于如果所述第二对话信息中不存在所述第二产品信息，则判断所述 第二对话信息是否为异议信息；及A second judging module, configured to judge whether the second dialog information is objection information if the second product information does not exist in the second dialog information; and

   第二发送模块，用于如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。The second sending module is configured to send objection push information corresponding to the second dialog information to the user terminal if the second dialog information is objection information.
9. 一种计算机设备，所述计算机设备包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机可读指令，其中，所述计算机可读指令被处理器执行以下步骤：A computer device comprising a memory, a processor, and computer-readable instructions stored on the memory and operable on the processor, wherein the computer-readable instructions are executed by the processor in the following steps :

   获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

   根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

   如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

   如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

   如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.
10. 如权利要求9所述的计算机设备，其中，所述计算机可读指令被处理器执行时还实现以下步骤：The computer device of claim 9, wherein the computer-readable instructions further implement the following steps when executed by the processor:

    判断所述第一对话信息和所述第二对话信息是否为文本信息；judging whether the first dialogue information and the second dialogue information are text information;

    如果所述第一对话信息和所述第二对话信息不是文本信息，则分别对所述第一对话信息和所述第二对话信息进行转写操作，以得到第一文本信息和第二文本信息；及If the first dialogue information and the second dialogue information are not text information, performing a transcribing operation on the first dialogue information and the second dialogue information respectively to obtain first text information and second text information ;and

    根据所述第一文本信息获取第一产品信息，并根据所述第一产品信息判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息。Acquiring first product information according to the first text information, and judging whether there is second product information corresponding to the first product information in the second text information according to the first product information.
11. 如权利要求10所述的计算机设备，其中，所述计算机可读指令被处理器执行时还实现以下步骤：The computer device of claim 10, wherein the computer readable instructions further implement the following steps when executed by the processor:

    从所述第一文本信息中提取多个第一关键词；extracting a plurality of first keywords from the first text information;

    根据所述多个第一关键词从预设映射表中确定所述第一产品信息；determining the first product information from a preset mapping table according to the plurality of first keywords;

    获取所述第一产品信息关联的多个关联关键词，其中，每个关联词对应的一个第二产品信息；Obtaining a plurality of associated keywords associated with the first product information, wherein each associated word corresponds to a piece of second product information;

    从所述第二文本信息中提取多个第二关键词；extracting a plurality of second keywords from the second text information;

    分别计算每个第二关键词与各个关联关键词的关键词相似度，得到每个日志关键词对应的多个关键词相似度；及Calculating the keyword similarity between each second keyword and each associated keyword respectively to obtain a plurality of keyword similarities corresponding to each log keyword; and

    根据各个第二关键词所对应的多个关键词相似度和预设关键词相似度判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息，其中，当所述多个关键词相似 度中存在至少一个关键词相似度大于所述预设关键词相似度时，则所述第二文本信息中存在所述第一产品信息对应的第二产品信息。Judging whether there is second product information corresponding to the first product information in the second text information according to the multiple keyword similarities corresponding to each second keyword and the preset keyword similarity, wherein, when the When at least one keyword similarity among the plurality of keyword similarities is greater than the preset keyword similarity, the second product information corresponding to the first product information exists in the second text information.
12. 如权利要求11所述的计算机设备，其中，所述计算机可读指令被处理器执行时还实现以下步骤：The computer device of claim 11, wherein the computer readable instructions further implement the following steps when executed by the processor:

    根据预先配置的多个属性分类模型，获取所述第二产品信息对应的多个属性和各个属性对应的预测属性值；Acquiring multiple attributes corresponding to the second product information and predicted attribute values corresponding to each attribute according to a plurality of pre-configured attribute classification models;

    根据所述多个属性和所述多个预测属性值，提取所述第二产品信息的属性预测向量；及extracting an attribute prediction vector of the second product information according to the plurality of attributes and the plurality of predicted attribute values; and

    根据所述属性预测向量，从数据库中获取与所述属性预测向量相似度最高的目标向量，并根据所述目标向量确定所述目标产品信息。According to the attribute prediction vector, the target vector with the highest similarity to the attribute prediction vector is obtained from the database, and the target product information is determined according to the target vector.
13. 如权利要求9所述的计算机设备，其中，所述计算机可读指令被处理器执行时还实现以下步骤：The computer device of claim 9, wherein the computer-readable instructions further implement the following steps when executed by the processor:

    提取所述异议信息对应的异议向量，并根据所述异议向量从所述数据库中获取与所述异议向量相识度最高的目标异议向量；及extracting the objection vector corresponding to the objection information, and obtaining a target objection vector with the highest degree of acquaintance with the objection vector from the database according to the objection vector; and

    根据所述目标异议向量确定目标异议，并根据所述目标异议从数据库中获取所述异议推送信息。The target objection is determined according to the target objection vector, and the push information of the objection is acquired from a database according to the target objection.
14. 如权利要求9所述的计算机设备，其中，所述计算机可读指令被处理器执行时还实现以下步骤：The computer device of claim 9, wherein the computer-readable instructions further implement the following steps when executed by the processor:

    如果所述第二对话信息中不存在异议信息，则将所述第二对话信息输入到预先训练好的环节分类模型，以通过所述环节分类模型输出所述第二对话信息的当前对话环节类型；及If there is no objection information in the second dialogue information, then input the second dialogue information into a pre-trained link classification model, so as to output the current dialogue link type of the second dialogue information through the link classification model ;and

    根据所述当前对话环节类型和所述第二对话信息从数据库中获取所述第二对话信息对应的对话推送信息，并将所述对话推送信息发送到所述用户终端。Obtain dialog push information corresponding to the second dialog information from a database according to the current dialog link type and the second dialog information, and send the dialog push information to the user terminal.
15. 一种计算机可读存储介质，其中，所述计算机可读存储介质内存储有计算机可读指令，所述计算机可读指令可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：A computer-readable storage medium, wherein computer-readable instructions are stored in the computer-readable storage medium, and the computer-readable instructions can be executed by at least one processor, so that the at least one processor performs the following step:

    获取目标用户的第一对话信息和目标对象根据所述第一对话信息回复的第二对话信息；Obtaining the first dialog information of the target user and the second dialog information replied by the target object according to the first dialog information;

    根据所述第一对话信息获取第一产品信息，并根据所述第一产品信息判断所述第二对话信息中是否存在所述第一产品信息对应的第二产品信息；Acquiring first product information according to the first dialogue information, and judging whether there is second product information corresponding to the first product information in the second dialogue information according to the first product information;

    如果所述第二对话信息中存在所述第二产品信息，则根据所述第二产品信息确定目标产品信息，并将与所述目标产品信息对应的目标推送信息发送到与所述目标用户关联的用户终端；If the second product information exists in the second dialogue information, then determine the target product information according to the second product information, and send the target push information corresponding to the target product information to the user associated with the target user. user terminal;

    如果所述第二对话信息中不存在所述第二产品信息，则判断所述第二对话信息是否为异议信息；及If the second product information does not exist in the second dialogue information, then determine whether the second dialogue information is objection information; and

    如果所述第二对话信息为异议信息，则将所述第二对话信息对应的异议推送信息发送到所述用户终端。If the second dialog information is objection information, sending objection push information corresponding to the second dialog information to the user terminal.
16. 如权利要求15所述的计算机可读存储介质，其中，所述计算机可读指令还可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：The computer-readable storage medium of claim 15, wherein the computer-readable instructions are further executable by at least one processor, so that the at least one processor performs the following steps:

    判断所述第一对话信息和所述第二对话信息是否为文本信息；judging whether the first dialogue information and the second dialogue information are text information;

    如果所述第一对话信息和所述第二对话信息不是文本信息，则分别对所述第一对话信息和所述第二对话信息进行转写操作，以得到第一文本信息和第二文本信息；及If the first dialogue information and the second dialogue information are not text information, performing a transcribing operation on the first dialogue information and the second dialogue information respectively to obtain first text information and second text information ;and

    根据所述第一文本信息获取第一产品信息，并根据所述第一产品信息判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息。Acquiring first product information according to the first text information, and judging whether there is second product information corresponding to the first product information in the second text information according to the first product information.
17. 如权利要求16所述的计算机可读存储介质，其中，所述计算机可读指令还可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：The computer-readable storage medium of claim 16, wherein the computer-readable instructions are further executable by at least one processor, so that the at least one processor performs the following steps:

    从所述第一文本信息中提取多个第一关键词；extracting a plurality of first keywords from the first text information;

    根据所述多个第一关键词从预设映射表中确定所述第一产品信息；determining the first product information from a preset mapping table according to the plurality of first keywords;

    获取所述第一产品信息关联的多个关联关键词，其中，每个关联词对应的一个第二产品信息；Obtaining a plurality of associated keywords associated with the first product information, wherein each associated word corresponds to a piece of second product information;

    从所述第二文本信息中提取多个第二关键词；extracting a plurality of second keywords from the second text information;

    分别计算每个第二关键词与各个关联关键词的关键词相似度，得到每个日志关键词对应的多个关键词相似度；及Calculating the keyword similarity between each second keyword and each associated keyword respectively to obtain a plurality of keyword similarities corresponding to each log keyword; and

    根据各个第二关键词所对应的多个关键词相似度和预设关键词相似度判断所述第二文本信息中是否存在所述第一产品信息对应的第二产品信息，其中，当所述多个关键词相似度中存在至少一个关键词相似度大于所述预设关键词相似度时，则所述第二文本信息中存在所述第一产品信息对应的第二产品信息。Judging whether there is second product information corresponding to the first product information in the second text information according to the multiple keyword similarities corresponding to each second keyword and the preset keyword similarity, wherein, when the When at least one keyword similarity among the plurality of keyword similarities is greater than the preset keyword similarity, the second product information corresponding to the first product information exists in the second text information.
18. 如权利要求17所述的计算机可读存储介质，其中，所述计算机可读指令还可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：The computer-readable storage medium of claim 17, wherein the computer-readable instructions are further executable by at least one processor, so that the at least one processor performs the following steps:

    根据预先配置的多个属性分类模型，获取所述第二产品信息对应的多个属性和各个属性对应的预测属性值；Acquiring multiple attributes corresponding to the second product information and predicted attribute values corresponding to each attribute according to a plurality of pre-configured attribute classification models;

    根据所述多个属性和所述多个预测属性值，提取所述第二产品信息的属性预测向量；及extracting an attribute prediction vector of the second product information according to the plurality of attributes and the plurality of predicted attribute values; and

    根据所述属性预测向量，从数据库中获取与所述属性预测向量相似度最高的目标向量，并根据所述目标向量确定所述目标产品信息。According to the attribute prediction vector, the target vector with the highest similarity to the attribute prediction vector is obtained from the database, and the target product information is determined according to the target vector.
19. 如权利要求15所述的计算机可读存储介质，其中，所述计算机可读指令还可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：The computer-readable storage medium of claim 15, wherein the computer-readable instructions are further executable by at least one processor, so that the at least one processor performs the following steps:

    提取所述异议信息对应的异议向量，并根据所述异议向量从所述数据库中获取与所述异议向量相识度最高的目标异议向量；及extracting the objection vector corresponding to the objection information, and obtaining a target objection vector with the highest degree of acquaintance with the objection vector from the database according to the objection vector; and

    根据所述目标异议向量确定目标异议，并根据所述目标异议从数据库中获取所述异议推送信息。The target objection is determined according to the target objection vector, and the push information of the objection is acquired from a database according to the target objection.
20. 如权利要求15所述的计算机可读存储介质，其中，所述计算机可读指令还可被至少一个处理器所执行，以使所述至少一个处理器执行如下步骤：The computer-readable storage medium of claim 15, wherein the computer-readable instructions are further executable by at least one processor, so that the at least one processor performs the following steps:

    如果所述第二对话信息中不存在异议信息，则将所述第二对话信息输入到预先训练好的环节分类模型，以通过所述环节分类模型输出所述第二对话信息的当前对话环节类型；及If there is no objection information in the second dialogue information, then input the second dialogue information into a pre-trained link classification model, so as to output the current dialogue link type of the second dialogue information through the link classification model ;and

    根据所述当前对话环节类型和所述第二对话信息从数据库中获取所述第二对话信息对应的对话推送信息，并将所述对话推送信息发送到所述用户终端。Obtain dialog push information corresponding to the second dialog information from a database according to the current dialog link type and the second dialog information, and send the dialog push information to the user terminal.

Images